How HP-UX Works: Concepts

for the System Administrator

HP 9000 Computers

F//]m HEWLETT
�/:a PACKARD

HP Part No. 82355-90029

Printed in USA August, 1 992

Third Edition
E0892
 Legal Notices
The information in this document is subject to change without notice.
Hewlett-Packard makes no warranty of any kind with regard to this manual,
including, but not limited to, the implied warranties of merchantability and
fitness for a particular purpose. Hewlett-Packard shall not be held liable for
errors contained herein or direct, indirect, special, incidental or consequential
damages in connection with the furnishing, performance, or use of this
material.
Warranty. A copy of the specific warranty terms applicable to your
Hewlett-Packard product and replacement parts can be obtained from your
local Sales and Service Office.

@copyright 1983-91 Hewlett-Packard Company

This document contains information which is protected by copyright. All rights
are reserved. Reproduction, adaptation, or translation without prior written
· permission is prohibited, except as allowed under the copyright laws.
Restricted Rights Legend. Use, duplication or disclosure by the U.S.
Government is subject to restrictions as set forth in subparagraph (c) ( 1 ) (ii)
of the Rights in Technical Data and Computer Software clause at DFARS
252.227- 7013 for DOD agencies, and subparagraphs (c) ( 1 ) and (c) (2 ) of the
Commercial Computer Software Restricted Rights clause at FAR 52.227- 19 for
other agencies.
HEWLETT-PACKARD COMPANY
3000 Hanover Street
Palo Alto, California 94304 U.S.A.

Use of this manual and media supplied for this release are restricted
to this product only. Additional copies of the programs may be made for
security and back-up purposes only. Resale of the programs in their present
form or with alterations is expressly prohibited.
 @copyright 1980, 1984, 1986 AT&T Technologies, Inc.
UNIX is a registered trademark of Unix System Laboratories Inc. in the USA
and other countries.
@copyright 1979, 1980, 1983, 1985- 1990 Regents of the University of California.
@copyright 1979 Regents of the University of Colorado, A Body Corporate.
@copyright 1986- 1988 Sun Microsystems, Inc.
@copyright 1986 Digital Equipment Corporation.
.
@copyright 1985-86, 1988 Massachusetts Institute of Technology.
X Window System is a trademark of the Massachusetts Institute of Technology.
MS-DOS and Microsoft are U.S. registered trademarks of Microsoft
Corporation.
OSF /Motif is a trademark of the Open Software Foundation, Inc. in the
U.S. and other countries. Certification for conformance with OSF /Motif user
environment pending.
All rights reserved.
Printing History
This edition documents the 9 .0 release of HP-UX, and is intended as a
companion manual to the System Administration Tasks manual.
The printing date and part number indicate the manual's current edition. The
printing date changes with each new edition. Minor changes may be made at
reprint without changing the printing date. The manual part number changes
when extensive editorial changes are made.
Manual updates may be issued between editions to correct errors or document
product changes. To ensure that you receive the updated or new editions, you
should subscribe to the appropriate product support service. See your HP sales
representative for details.
• September 1989-First edition, as HP- UX System Administration Concepts
• January 1991-Second edition, as How HP- UX Works: Concepts for the
System Administrator
• June 1991-Third edition
• August 1992-Fourth edition
This Edition documents the 9.0 release of HP-UX, and is intended as a
companion manual to the System Administration Tasks manual.

iv
Contents

1. Introduction
Typographical Conventions . . . . . . . . 1-2
Chapter Summaries . . . . . . . . . . . 1-3
The System Administration Manager ( SAM ) 1-4
Related System-Administration Manuals 1-5

2. System Startup
Boot ROM Startup Sequence Overview 2-2
Secondary Loader . . . . . . . . . 2-3
Boot ROM Startup Sequence ( Series 300/400 ) 2-4
Boot ROM Finds a System Console . . . . 2-4
Serial Interface Line Control Switch Pack Settings . 2-5
Boot ROM Tests Hardware . . . . . . . . . . 2-5
Boot ROM Finds and Loads an Operating System 2-6
Unattended Mode . . . . . . 2-6
Boot ROM Search Sequence . . . . . . . 2-7
Boot Search Criteria . . . . . . . . . . 2-8
Series 400 Unattended-Mode Boot Variation 2-9
Attended Mode . . . . . . . 2-9
Boot ROM I/O Configuration 2- 1 1
HP-UX Takes Control . . . . 2- 12
Boot ROM Startup Sequence ( Series 700 ) 2- 13
Attended Mode . . . . . . . . . . . 2- 14
Boot Administration Mode . . . . . 2- 15
Boot ROM Startup Sequence ( Series 800 ) 2- 16
System Boot Overview . . . . . . . 2- 17
Boot ROM Search Sequence ( Series 800 ) 2- 17
Autoboot 2- 17
Manual Boot 2- 18
Autosearch . 2- 19

Contents- 1
 HP-UX Startup Sequence .. . . . 2-20
HP-UX Finds the Root File System 2-20
HP-UX Starts the init Process . . 2-20
System Initialization File-/etc/inittab 2-21
Default Run Levels-initdefault . . . . 2-22
Boot Check Run Command-/etc/bcheckrc 2-23
Shared Libraries Check-/etc/recoversl 2-24
Additional Boot Tasks-/etc/brc . . 2-24
Initial Customization Script-/etc/rc . 2-24
Powerfail Routines-/etc/powerfail .. 2-27
Terminal Processes Startup-/etc/getty 2-27
For More Information . . . . .. .. 2-29

3. System Shutdown
Halting vs Rebooting 3- 1
Shutdown vs Reboot 3-2
When to Use shutdown 3-2
When to Use reboot . 3-3
Examples . . . .. . . 3-4

4. Login
Overview of Login . . . .. . . . . . . .
. 4-2
The login Process.. . .. . . . . . . . . 4-3
The Shell Environment Initialization Sequence 4-5
Correcting Typing Mistakes during Login 4-7
The /etc/passwd File . . . . 4-8
Syntax of Entries .. . . . . . ... 4-8
Sample /etc/passwd Entries . . ... 4- 11
Login Tracking Files (/etc/btmp and /etc/wtmp) 4- 12
/etc/btmp . ... . 4- 12
/etc/wtmp. .. . .... . . . . 4- 12
Environment Variables ... . . . . . 4- 13
The Command Search Path (PATH) 4- 15
PATH Variable Format . . ... 4- 15
Commonly Used PATH Directories 4- 15
Specification of Terminal Characteristics (TERM) . 4- 16
Selecting a Value for the TERM Variable 4- 16
Setting TERM with the tset Command . . . . 4- 17

Contents-2
 Using stty to Modify and Display Terminal Characteristics. 4- 18
Specifying the Working Environment with System Login Scripts 4- 19
/etc/profile ...... 4-20
/etc/csh.login ..... 4-22
Default Local Login Scripts 4-24
/etc/cl.profile 4-24
/etc/cl.login ...... 4-26
/etc/d
.cshrc ...... 4-28
Sample Bourne Shell.profile Script 4-30
Sample C Shell.login Script 4-32
Key Shell-keysh ....... . 4-34

5. Process Management
What Is a Process? . 5-2
Process Relationships .... . 5-2
Process and Parent Process IDs . 5-3
User and Group IDs (Real and Effective) 5-4
Audit ID (Trusted Systems Only) 5-5
Jobs.... 5-6
Job Control 5-6
Process Groups 5-6
Group Access Lists 5- 7
Sessions ..... 5-7
Processes and Terminal Affiliation 5-8
Attempted read by Background Process Group 5-8
Attempted Write by Background Process Group 5-8
Process Creation ..... 5-9
The fork System Call .. 5-9
The vfork System Call. 5- 10
The exec System Call 5- 1 1
Open Files....... 5- 1 1
Process Termination.... 5- 12
Process Management Commands 5- 13
Understanding Process Status-ps 5- 13
Understanding Process Termination-kill 5- 14
Understanding Relative Process Priority-nice and renice 5- 15
Tools for Monitoring Process Management Performance 5- 16
Process Management and the Kernel ......... . 5- 17

Contents-3
 Process Modes . 5- 17
Process Priorities 5- 18
Run Queues 5-19
Process Context 5-21
Process States and Transitions 5-22
Interrupts . 5- 24
Signals . . . . . . . . . . 5-25
Multiprocessing . . . . . . . . 5-26
How Multiprocessing Compares to Uniprocessing 5-26
Symmetry . . . . . . . . . . . . . . . . 5-26
Multiprocessor Boot-Up . . . . . . . . . . 5-26
Scheduling Processes using Semaphores and Spinlocks 5-27
Processor Affinity . . . . . . 5-30
Uniprocessor Emulation Issues . . . . . . . . . . 5-30

6. Run-Levels
What Is a Run-Level? 6- 1
Predefined Run-Levels 6-2
Run-Level 2 6-2
Run-Level 0 6-2
Run-Level s 6-2
The init Process 6-2
/ etc/inittab 6-3
Setting the Default Run-Level 6-4
User-Defined Run-Levels 6-5
Changing Run-Levels . . . . 6-5

7. Memory Management
Physical Memory . . . . . . . . . . 7-2
Power and Data Permanence . . . . 7-2
Transactions between RAM and Disk 7-3
Available Memory 7-3
Lockable Memory . . . . . . . . . 7-5
Secondary Storage . . . . . . . . 7-6
HP-UX Demand-Paged Virtual Memory 7-9
Hardware Perspective on Memory Transactions 7-9
Process-to-Page Mapping 7- 12
Pages . . . . . . . . . . . . . . . . .
. 7- 13

Contents-4
 Virtual Address Space....... 7- 13
Configuring Virtual Address Space 7- 14
Per-Process Regions (pregions) ... 7- 14
Per-Process Region Layouts 7- 15
Placement of Segments in Address Space ( Series 300/400) 7- 16
Placement of Segments in Address Space (Series 700/800) 7- 18
Regions ............... 7-21
Virtual Nodes ( vnodes) ....... 7-21
Memory-Mapped Files (Series 300/400/ 700) 7-22
Limitations to Memory Mapping .... 7-23
How the Kernel Executes Processes using Demand Paging 7-24
copy-on-write................... 7-24
Maintaining Page Availability-vhand and swapper Daemons. 7-25
Thrashing .... ................... 7-27
How the Memory-Management System Handles Executable Code 7-28
Standard Executable Code (EXEC_MAGIC) 7-30
Shared Code ( SHARE_MAGIC) ........... 7-30
Demand-Loaded Code (DEMAND_MAGIC) ...... 7-31
Benefits and Shortcomings of Shared and Demand-Loaded
Code...................... 7-33
Comparison of Shared and Demand-Loaded Code... 7-33
Shared Code in an HP-UX Cluster ( Series 300/400/700 only). 7-34
Shared Libraries .............. 7-34
How Shared Libraries are Dynamically Loaded 7-36
Shared Libraries vs . Archived Libraries 7-38
Administering Shared Libraries . .. 7-39
HP-UX Memory-Management Features 7-40
Swap Space Management 7-41
Swap Parameters . 7-42
Device Swap Space .. 7-42
File-System Swap... 7-45
Guidelines for Adding Swap Space 7-45
Sample /etc/checklist Entry for Device Swap 7-46
Sample /etc/checklist Entry for File-system Swap 7-46
Comparing Device and File-System Swap 7-48
Swap Space Priorities . ... 7-48
The swapon Command 7-48
Evaluating Swap-Space Needs 7-51

Contents-5
 -

Swapping in an HP-UX Cluster ( Series 300/400/700 only) 7-52

Pseudo-Swap Reservation (Series 800 only) . . . . . . . . 7-53

8. HFS File System

Understanding File-System Creation 8-2
Disk Layout . ....... 8-8
Series 300/400 Disk Layout 8-9
Series 700 Disk Layout 8- 10
Series 800 Disk Layout 8- 1 1
Understanding Disk Partitioning ( Series 800 only) . 8- 11
HP-UX Disk Sections 8- 12
Logical Volumes .... 8- 13
File System Size . . . . . 8- 14
Disk and File System Tools 8- 14
Disk Geometry Database-/etc/disktab 8- 14
Disk Characteristics Command-/etc/diskinfo 8- 16
Free Disk Blocks Command-bdf 8- 18
Disk Usage Command-du . . . . . . . . 8- 19
Boot Area . .............. 8-20
Series 300/400 Boot Area Implementation 8-20
Series 700 Boot Area Implementation 8-22
Series 800 Boot Area Implementation 8-22
Primary Swap Area 8-22
File System Layout .. 8-24
The Superblock . .. 8-25
The Cylinder Group . 8-26
!nodes ...... 8-29
Data Blocks . .. 8-30
How a File is Accessed from !node to Data Blocks 8-32
Minimum Free Space . . 8-36
How Disk Space is Allocated . 8-37
Allocation Policies .... 8-38
The File-System Buffer Cache 8-39
Dynamic Buffer Cache (Series 300, 400, and 700 only) 8-42
How the HFS File System Modifies Files . . . 8-43
Immediate Reporting . . . . . . . . . . 8-45
Synchronous vs. Asynchronous Disk Writes 8-46
Minimizing File-System Corruption . . . . . 8-47

Contents-&
 System Shutdown and Startup Guidelines . . . . . . . . 8-47
The /lost+found Directory . . . . . . . . . . . . . . . 8-48
Understanding Use of fsck to Detect and Correct File-System
Corruption . . . . . 8-48
Superblock Consistency 8-50
File System Size 8-50
Free-Block Checking . 8-50
Inode Checking . 8-5 1
Inode Consistency . 8-51
Format and Type 8-52
Link Count 8-52
Duplicate Blocks 8-52
B ad Blocks . 8-53
Inode Size . . . 8-53
Block Count . . 8-54
File-System Connectivity 8!55
Uncorrectable File System Corruption 8-55
Transferring Files between HP-UX and Other Systems . 8-55
File Protection . . . . . . . . . . . . . . . . . . 8-58
Protecting Directories . . . . . . . . . . . . . . 8-60
Setting Effective User and Group ID Bits (suid, sgid) 8-61
Access Control Lists . . 8-63
File Sharing and Locking 8-64
Advisory Locks . . 8-64
Enforcement Mode 8-65
Locking Activities . 8-66

9. Logical Volume Manager

The LVM Paradigm . . . . . . . . . . . . . . . . . . . 9-2
Understanding Migration from Sections to Logical Volumes 9-4
LVM Terminology . . . . . . . . . . . . . . . 9-5
Using LVM Device Files . . . . . . . . . . . . 9-8
Major and Minor Numbers of LVM Data Structures 9- 10
Understanding Volume Groups . . . . . . . . . 9- 12
HP-IB Limitations . . . . . . . . . . . . . 9- 12
Consider newfs Limitations when Organizing Volume Groups 9- 14
Consult /etc/diskinfo for Disk Attributes 9- 15
The Root Volume Group . . . . . . . . . . . . . . . . 9- 16

Contents-7
 Contiguous vs. Non- Contiguous Allocation of LVM Disk Space 9- 18
Understanding Logical Volumes . . . . . . . 9-21
Allocating Disk Space for Logical Volumes . . 9-23
How LVM Maps Extents to Logical Volumes 9-24
LVM Disk Space Allocation Commands . . 9-24
Understanding Physical Volumes (LVM Disks) 9-25
Sector Size of Physical Volumes (LVM Disks) 9-26
Characteristics and Layout of LVM Disks 9-29
Boot Data Reserved Area (BDRA) . . . 9-30
LIF Information . . . . . . . . . . 9-31
Physical Volume Reserved Area (PVRA) . 9-32
Volume Group Reserved Area (VGRA) 9-33
Volume Group Descriptor Area (VGDA) 9-33
Volume Group Status Area (VGSA) 9-34
Mirror Consistency Record (MCR) 9-34
User Data Area . . . . . . 9-35
Bad Block Relocation Policy . . . . 9-35
LVM Overhead . . . . . . . . . . 9-36
Understanding LVM Configuration Maintenance 9-37
Guidelines for Configuring the Root Volume Group 9-37
The /etc/lvmtab File . . . . . . . . . . . . . 9-37
Dealing with Removable Media and Volume Groups 9-39
LVM Maintenance Mode 9-39
Mirroring . . . . . . . . . . . . . . . . 9-40
I/ 0 Channel Separation . . . . . . . . . 9-43
How Mirrored Logical Volumes are Written 9-47
Allocation Policies for Mirroring . . . . 9-4 7
Scheduling Policies for Disk Writes 9-4 7
Synchronization Policies for Recovering Mirrored Data . 9-48
Manual Synchronization 9-49
Backing Up Mirrored Data . . 9-49
Useful Mirroring Commands . 9-49
Internal Representation of LVM . 9-51

Contents-a
10. System Architectures
Architecture of the Series 300 Bus . 10-2
Model 375-a DI0-11 Implementation 10-3
Processor-I/ 0 Board . . . . 10-5
Graphics Interface . . . . . 10-6
Architecture of the Series 400 Bus . 10-7
Addressing on a Series 300 / 400 . 10-8
Series 300 / 400 Select Codes 10-8
Series 300 / 400 Bus Addresses . . 10- 10
Architecture of the Series 700 Bus . 10- 1 1
Series 700 Modules . . . . . . 10- 13
Addressing on a Series 700 . . . 10- 14
Interpreting Series 700 Hardware Paths using ioscan . 10- 14
Architecture of the Series 800 Bus . . . . . . 10- 15
Hardware Modules Connecting to the Buses 10- 1 7
Addressing on Series 800 . . . . . . 10- 19
HP-PB Implementations-Models 8x2 10-20
Addressing on HP-PB Models 10-21
A Model 8x7 Implementation . . . . 10-23
Addressing Models 8x7 . . . . . 10-26
A Mid-Bus/CIO Implementation-Model 845SE 10-27
Addressing on Mid-Bus/CIO Models 10-27
Factory-Floor Communication . . . 10-28
An SMB Implementation-Model 850S 10-30
Addressing on SMB Models . . . . 10-30
Processor-Memory Bus (PMB) Implementation-Model 890 10-33
Physical Layout of the Model 890 10-33
Addressing on a Model 890 . . . . . . . . . . . 10-36
Model 890 Guidelines . . . . . . . . . . . . . 10-37
Finding Out About the Current System Configuration 10-38
Viewing Hardware Configuration Using /etc/dmesg 10-38
Interpreting Hardware Paths on a Series 800 . . 10-41
Viewing Hardware Configuration Using ioscan 10-41

Contents-9
 --

11. System Configuration

What Does Configuring Mean? 1 1- 1
Understanding the Role of the Kernel 1 1-2
Kernel Configuration Files . . . . . 1 1-3
dfile-Series 300/400/700 Kernel Configuration File 1 1-4
S 800 file-Series 800 Kernel Configuration File . . 1 1-7
The io Statement of the S 800 File . . . . . . . 1 1-9
The io Statement of a Mid-Bus/CIO Configuration 11-10
The io Statement of an HP-PB Configuration . 1 1- 12
When Do You Reconfigure the Kernel? 1 1-13
Statements to Include Device Drivers 1 1- 14
Kernel Devices . . . . . . . 1 1- 14
Kernel Dump Device . . . . . . 1 1-15
Operating-System Parameters 1 1-16
What Happens when you Regenerate the Kernel? 1 1- 17
Kernel Regeneration-the Series 800 Case 1 1- 17
Device Drivers . . . . . . . 1 1-20
The /etc/master File . . . 1 1-22
Device Driver Configuration 1 1-23
Pseudo-Drivers . . . . . 1 1-25
Device Files . . . . . . . . 1 1-26
Characteristics of Device Files 1 1-27
Organization of Device File Directories 1 1-27
Device File Directories in a Cluster (Series 300/400/700
only) . . . . . . . . . . . 1 1-27
Block versus Character Device Files . . 1 1-28
Device File Listing Format . . . . . . 1 1-28
Overview of Major and Minor Numbers 1 1-30
Major Numbers . . . . . . 1 1-31
Major Numbers and lsdev . . . . 1 1-31
Minor Nuinbers . . . . . . . . . . 1 1-34
Minor Numbers on Series 300/400 . 1 1-35
Minor Numbers on the Series 700 . 1 1-38
Decoding a Minor Number for a Series 700 SCSI Device 1 1-39
Decoding a Minor Number for a Series 700 EISA S C SI
Device . . . . . . . . 1 1-40
Minor Numbers on Series 800 . 1 1-42
Logical Unit Number . . . 1 1-43

Contents-1 o
 Sources for Minor Number and other Device Information . 1 1-44
Device File Name Conventions . . . . . 1 1-45
Naming Disk Devices . . . . . . . . . 1 1-45
Naming Magneto- Optical Disk Devices 1 1-47
Naming Nine-Track Magnetic Tape Drives 1 1-47
Naming DDS-Format Tape Drives . 1 1-48
Naming QIC-Format Tape Drives . . . 1 1-48
Naming Cartridge Tape Drives . . . . 1 1-49
Naming HP-IB Devices ( Series 800 only) 1 1-49
Naming Modem Device Files . . . . . 1 1-50

12. HP-UX Peripherals

Magnetic Tape . . . . . . . . . . . . 12-2
Magnetic Tape Physical Characteristics 12-2
Tape Density . . . . . . . . . . . 12-2
Logical Organization of Nine-Track Magnetic Tape 12-3
File Marks . . . . . . . 12-3
End of Tape Markers . . 12-3
Magnetic Tape Capabilities 12-4
Records and Record Sizes 12-4
Cyclic Redundancy Check 12-5
Coding . . . . . . 12-5
Write/Read Errors 12-6
Preventive Maintenance 12-6
Tape Streaming . . . . 12-6
Immediate Response Mode 12-7
DDS-Format Tape Drive . . . 12-9
Comparison of DDS-Format and Nine-Track Magnetic Tape 12-9
Logical Organization of DDS-Format Tape 12-9
For More Information . . . 12- 1 1
Cartridge Tape Drives . . . . 12-12
CD-ROM File System (CDFS) 12-13
Overview of Implementation 12- 13
CD-ROM Standard Documents 12- 1 3
System Call Changes 12- 14
Command Changes . . . . . . 12- 1 6
Enabling CDFS . . . . . . . . 12-16
Series 300 / 400 / 700 Configuration for CD FS 12- 17

Contents- 1 1
 Series 800 Configuration for CDFS 12- 17
HP-UX Usage 12- 17
Optical Technology . . . . . . 12-18
Rewritable Optical . . . . . 12-18
Why Use Rewritable Optical? 12-19
Hewlett-Packard's Rewritable Optical Products 12-20
HP Series 6300 Model 650/ A - Optical Disk Drive . 12-20
Optical Disk Drive Guidelines . . . . . . . . . 12-20
HP Series 6300 Model 20GB /A - Optical Disk Library System 12-21
Optical Disk Library System Guidelines 12-23
Terminals and Modems . . . . . . . 12-25
Terminal Configuration . . . . . . 12-25
Special Considerations for Terminals 12-26
/etc/inittab Entries for Terminals 12-26
HP-IB Guidelines . . . . . . 12-28
HP-IB Electrical Guidelines 12-29
S C SI Device Guidelines 12-30

13. Networking
Network Architecture and the OSI Model 13-2
HP-UX Networking . . . . 13-3
Networking Documentation . . . . . . 13-4

14. System Accounting

What Is in This Chapter? . . . . . . 14-2
Accounting System Planning and Billing 14-3
Grouping Users for Billing . . . . . 14-3
Criteria for Billing Computer Resources 14-3
Time Considerations . . . . . . . 14-4
Prime vs Non-prime Connect Time 14-4
Updating the Holidays File . 14-4
Collecting Data Strategically . 14-6
Overview of System Accounting . 14-7
Gathering Data . . . . . . . 14-8
Automating Data Collection 14-8
Generating Reports . . . . 14-9
Total Accounting Records 14-9
Summary Reports . 14-9

Contents· 1 2
System Accounting Set-up Guidelines . . . . . . . . 14- 10
Administering System Accounting in a Cluster (Series
300/400/700 only) . . . . . . . . 14- 10
How System Accounting is Turned On . 14- 1 1
How System Accounting i s Turned O ff . 14- 12
Logging In for System Accounting . 14- 12
System Accounting Files . . . . . . . . 14- 13
Files in the /usr/adm Directory . . . . 14- 14
Files in the /usr/adm/acct/nite Directory 14- 15
Files in the /usr/adm/acct/sum Directory 14- 16
Files in the /usr/adm/acct/fi.scal Directory . 14- 16
Disk-space Usage Accounting . . . . . . . . 14- 17
Reporting Disk Space Usage . . . . . . . 14- 17
Computing Disk Resource Consumption by Login-acctdusg 14- 18
Generating Disk Accounting Data by User ID-diskusg 14- 18
Comparing acctdusg and diskusg . 14- 19
Creating Total Accounting Records 14-21
Other Disk-related Commands . . 14-22
A Security Note . . . . . . . 14-22
Accounting for Disk-space Usage in a Cluster (Series
300/400/700 only) . . . . . . . . . . . . . . 14-22
Connect-session Accounting . . . . . . . . . . . . . 14-23
The Key Connect-session Accounting File-/etc/wtmp 14-23
Connect-session Accounting Commands . . . 14-24
Writing Records to wtmp - acctwtmp . . . 14-24
Displaying Connect Session Records - fwtmp 14-25
Fixing wtmp Errors - wtmpfi.x
. . . . . . 14-26
Recording Connect-session Statistics . . . . 14-27
Tabulating Output for Connect-session Accounting-
acctconl . . . . . . . . . . . . . . . . . . 14-28
Heading Columns of acctconl Output-prctmp . . . 14-30
Creating Connect-session Total Accounting Records-
acctcon2 . . . . . . . . 14-30
Process Accounting . . . . . . . 14-31
Turning On Process Accounting 14-32
Turning Off Process Accounting 14-33
Managing the Size of the Process Accounting File pacct 14-34
Checking the Size of pacct-ckpacct . . . . . . . . 14-34

Contents- 1 3
 Changing Process Accounting Files-turnacct switch 14-35
Displaying Process Accounting Records - acctcom 14-35
Sample Report using acctcom . . . . . . . . 14-36
Definitions of Information Produced by acctcom 14-36
Additional Capabilities of acctcom 14-37
Formating Options of acctcom . . . . . . . . 14-39
Record Selection Options of acctcom . . . . . 14-41
Process-accounting Report of Commands - acctcms 14-42
Producing a Readable Report - the -a option 14-42
Sample Report by acctcms . 14-42
Header Descriptions . . . . . . . . . 14-43
Additional acctcms Options . . . . . 14-45
Creating Total Process-accounting Records 14-46
User Fees - chargefee . . . . . . . . . . 14-48
Accounting Summaries and Reports . . . . 14-49
Displaying Total Accounting Records - prtacct 14-49
Sample Output by prtacct . . . . 14-50
Report Format of prtacct . . . . . . . 14-50
Origin of Columnar Data by prtacct . . . 14-51
Merging Total Accounting Files - acctmerg . 14-51
Command Options for acctmerg . . . . 14-52
Creating Daily Accounting Information - runacct 14-53
Files Processed by runacct 14-54
Safeguards of runacct . . . . . 14-55
States of runacct . . . . . . . 14-55
Recovering from runacct Failure 14-57
Daily and Monthly Reports of runacct 14-57
Daily Line Usage Report of runacct 14-57
Daily Resource Usage Report . . . 14-58
Daily and Monthly Command Summary 14-59
Last Login . . . . . . . . . . . . . 14-59
Creating Daily Summary Reports - prdaily 14-60
Creating Monthly Accounting Reports - monacct 14-61
Disk Quotas . . . . . . . . . . . . . . . . . 14-62
Planning to Set up Disk Quotas . . . . . . . 14-62
Deciding Which File Systems Require Disk Quotas 14-62
Setting Quotas and Limits 14-63
How Disk Quotas Work . . . . . . . . . . . . . 14-63

Contents- 1 4
 Disk-Quotas Data File-quotas . 14-64
How Disk Quotas Appear to Users 14-65
Disk Quota Commands . . . . . 14-66
Enabling and Suspending Disk Quotas-quotaon and
quotaoff . . . . . . . . . . . . . . . . . 14-66
Maintaining File-system Consistency-quotacheck . 14-67
Setting and Editing User Quotas-edquota . 14-67
Editing Disk-Quota Time Limits . . . 14-68
Displaying Users' Disk Quotas-quota . . . 14-69
Enforcing Disk Quotas Remotely-rquotad . 14-70
Manipulating Disk Quotas Programmatically-quotactl 14-70
Summarizing Disk Usage and Quotas-repquota 14-70

Glossary

Index

Contents-15
Figures

2- 1 . The Boot Program Searches for an Operating System 2-3

4- 1 . Sample /etc/passwd Entry for User terry 4-2
4-2. The login Process Control Flow 4-4
5- 1 . Process Tree Graph . . . . . . . . . . 5-3
5-2. Run Queue, showing Priorities . . . . . 5- 19
5-3. System Switching Context from One Process to Another 5-21
5-4. Process States and Transitions . . . . . . . . . . . 5-23
5-5. Multiprocessors Booting Up . . . . . . . . . . . . 5-27
5-6. Processes Lock Kernel Resources to Synchronize Execution . 5-28
5- 7. Kernel Allocates Resources to Three Concurrent Processes on
Two Processors Using Semaphores and Spinlocks 5-29
7- 1 . Physical Memory Available to Processes . . . . . . . . . 7-4
7-2. Relationship between Main Memory and Secondary Storage 7-7
7-3. Virtual Address Translation . . . . . . . . . . . . . . 7- 10
7-4. Role and Magnitude of the Address-Translation Components 7- 11
7-5. Simplification of an Active Process Executing in Main Memory. 7- 12
7-6. Series 300 / 400 User Process Virtual Address Space 7- 17
7- 7. Series 700 User Process Virtual Address Space with
EXEC_MAGIC implemented. . . . . . . 7- 19
7-8. Series 800 User Process Virtual Address Space 7-20
7-9. The Pageout Daemon vhand . . . . . . . . 7-25
7- 10. Maintaining Memory Availability . . . .. . . 7-26
7- 1 1 . Alignment of Page Boundaries Simplifies Transfer of Data and
Code to Memory . . . . . . . . . . . . 7-32
7- 12. Shared Libraries Use Memory More Efficiently . . 7-35
7- 13. The Dynamic Loader at Work . . . . . . . . . 7-37
7- 14. Series 300/400/700 and 800 Swap Space Compared 7-43
8- 1 . File System, Viewed Physically and Logically . . . 8-4
8-2. newfs Builds the Disk Infrastructure for a File System 8-6

Contents- 1 6
 8-3. File System /dev/dsk/ lsO Mounted to Root File System at
/users . . . . . . . . . . . . 8- 7
8-4. Disk Sections and Relative Sizes . . . . . . 8- 12
8-5. Examples of Valid Sectioning Schemes . . . . 8- 13
8-6. Track, Cylinder, and Cylinder Group on a Disk 8-27
8- 7. Mapping from Inode to File Data Blocks . . . 8-33
8-8. Inode Addressing Example . . . . . . . . . 8-35
8-9. Buffer Cache Holds the a.out Header of Executing Programs 8-39
8- 10. File Permission Bits . . . . . . . . . . . . . . . 8-59
8- 1 1 . File Permission Bits of rwxr-xr-- . . . . . . . . 8-60
8- 12. Permission Bits of an suid/sgid file set to rwsr-sr-- 8-62
9- 1 . Disk Space Apportioned into Logical Volumes 9-3
9-2. Device File for an LVM Disk (physical volume) . . . 9- 10
9-3. Volume Group , Showing Minor Numbers (shaded) . . 9- 10
9-4. Hexadecimal format of LVM Major and Minor Numbers 9- 1 1
9-5. Sample Verbose Output of vgdisplay . . . . . . . . 9- 13
9-6. Root Volume Group . . . . . . . . . . . . . . . . 9- 17
9- 7. Contiguous vs. Non- Contiguous Mapping of Logical to Physical
Extents . . . . . . . . . . . . . . . . . . . . 9- 19
9-8. Sample Output of lvdisplay . . . . . . . . . . . . 9-2 1
9-9. Verbose Output of lvdisplay, Showing Logical Extents 9-22
9- 10. Sample Output of pvdisplay . . . . . . . . . . . . 9-25
9- 1 1 . Verbose Output of pvdisplay , showing Mapping of Physical
and Logical Extents . . . . . . . . . . . . . . . . 9-26
9- 12. Relationship Between Disk Sector Size and DEV _B SIZE. 9-27
9- 13. Bootable and Non-Bootable Physical Volume Layouts, Showing
Organization of Data Structures . . . . . . 9-29
9- 14. Strict Mirroring of Logical Volume lvol2 . . . . 9-41
9- 15. Three LVM Disks with Mirrored Logical Volumes 9-42
9- 16. Three LVM Disks Mirror a Single Logical Volume 9-43
9- 17. I/O Channel Separation via Separate Interface Cards 9-44
9- 18. I/O Channel Separation via Separate Buses, and using Physical
Volume Groups . . . . . . . . . 9-45
9- 19. Architecture of the LVM Subsystem . . 9-51
10- 1 . Backplane of a Model 375 Workstation. 10-3
10-2. Model 375 Workstation Functions . . . 10-4
10-3. HP Apollo 9000 Series 400 Workstations Block Diagram 10-7
10-4. Series 300/400 Addressing . . . . . . . . . . . . . 10-8

Contents- 1 7
 10-5. Models 720/730 and 750/7�0 Rear Panels, showing Ports 10- 1 1
10-6. HP Apollo 9000 Series 700 Bus Architecture. 10- 12
10-7. S ample Model 8x2S Card- Cage Layout 10-20
10-8. Model 8x2S bus architecture . . . . . . . . 10-21
10-9. A typical backplane of a high-end Model 8x7S 10-23
10- 10. Model 8x7S Bus Architecture . . . . . . 10-24
10- 1 1 . Addresssing the Modules of a Model 8x7S 10-25
10- 12. Sample Model 845 card-cage layout . . 10-28
10-13. Model 845 Bus Architecture . . . . . 10-29
10- 14. Sample Model 850S Card- Cage Layout . 10-31
10-15. Model 850 bus architecture . . . . 10-32
10-16. Model 890 Card Cage . . . . . . . . 10-34
10- 17. Bus Architecture of the Model 890 10-35
11-1. Kernel in Relation to the User Environment and I/O Devices. 1 1-2
1 1-2. Kernel Configuration Files. . . . . 1 1-3
1 1-3. System Initialization by the Kernel . . . . . . 1 1-18
1 1-4. Series 800 I/O Auto-configuration . . . . . . . 1 1-19
1 1-5. Device Files Contain Major and Minor Numbers. 1 1-30
1 1-6. Series 700 Minor Number Format . 1 1-38
12- 1 . Magnetic Tape Format . . . . . 12-4
12-2. Organization of DDS-Format Tape 12- 10
12-3. Direct Access Secondary Storage . 12-19
12-4. Optical Disk Library System Concept 12-22
14- 1 . System Accounting Overview . . . . 14-7
14-2. How startup Enables System Accounting . 14- 1 1
14-3. System Accounting Directory Structure 14- 13
14-4. Disk-space Usage Accounting 14- 17
14-5. Connect-session Accounting 14-23
14-6. Process Accounting . . . . 14-32
14-7. Charging Fees . . . . . . 14-48
14-8. Displaying Total Accounting Records 14-49
14-9. How runacct Generates Summary Reports 14-53
14-10. Disk Quotas Flow Chart . . . . . . . . 14-64
14- 1 1 . How repquota Generates Disk-quota Reports 14-71

Contents- 1 8
Tables

2- 1 . /etc/re functions . . . . . . . . . . . 2-26

2-2. Additional Startup Information . . . . . 2-29
3- 1 . Examples of using shutdown and reboot . 3-4
4- 1 . Password Aging Encryption Characters 4-9
4-2. Shell Environment Variables . . . . . . 4- 14
4-3. Possible Directories to Include in PATH . 4- 15
4-4. Settings for the TERM Environment Variable 4- 17
7- 1 . Characteristic Types of Executable Code . 7-29
8- 1 . Series 300/400 Root Disk Layout 8-9
8-2. Series 700 Disk Layout . . . . 8- 10
8-3. Series 300/400 Boot Area Layout 8-21
8-4. Cylinder Group Layout . . . . 8-28
8-5. Sample Free Block Bitmap in an 8KB/ 1KB File System 8-37
8-6. Utilities and Services for File Transfer . . . . 8-56
8-7. Explanation of File Permission Bits rwsr-sr- . . . 8-63
9- 1 . LVM Device File Names (defaults) . . . . . . . 9-8
9-2. When to Use Block vs. Raw Special Files for LVM 9-9
10- 1 . Typical Series 300/400 Select Codes . . 10-9
10-2. Buses Supported on Series 800 Models . 10-15
10-3. Sample Hardware Paths . . . . . . . 10-41
1 1- 1 . Device-Driver Capabilities (Series 800) . 1 1-24
1 1-2. Summary of Series 300/400 Minor Number Formats 1 1-36
1 1-3. Decimal, Binary, and Hexadecimal Equivalents 1 1-39
1 1-4. Summary of Series 700 Minor Number Formats 1 1-41
12- 1 . HP-IB Connectivity . . . . . 12-28
13- 1 . O SI Model and Layers . . . . 13-2
13-2. HP 9000 Networking Products 13-3
13-3. Available Networking Services 13-5

Contents- 1 9
 1

1
Introduction

This manual supplements the System Administration Tasks manual to

describe the essential concepts of system administration. Whereas the System
Administration Tasks manual describes how to administer your HP-UX system,
this manual discusses the underlying principles.
Use the System Administration Tasks manual to perform specific
system administration tasks, such as adding a new user or reconfiguring
operating-system parameters. Use this manual to learn about what happens
to the operating system when you administer it ( for example, how does the
system get to a multi-user state during system startup; how does the virtual
memory subsystem apportion user address space on each HP-UX platform ) .
This manual is especially aimed at users who simply have a need to know.
Some users are satisfied with procedural explanations of how to do things;
others require more background conceptual information. How HP- UX Works
attempts to address the concerns of these users.
This introductory chapter contains the following:
• Typographical conventions used in this manual.
• Table of chapter contents.
• Description of the System Administration Manager ( SAM ) .
• List of related system-administration manuals.

Introduction 1·1
1

Typographical Conventions
Throughout this manual, the following typographical conventions are used:

If You See This ... It Means . . .

term Boldface text indicates a term being introduced for the first
time in a chapter .
command typewriter text denotes commands, file names, and
information displayed by the computer.
emphasis italic text is used for emphasis (for example: never remove
this file . . . ).
refpage (l) is a reference to a page in the HP-UX Reference . This
particular example says that the page re/page is found in
section 1 of the HP-UX Reference .
what you type underlined text is used in examples to distinguish what you
type from what the computer displays; for example:
Y. ls
report s personal proj ects
When the user typed ls to the HP-UX prompt (Y.), the
system displayed three file names : reports , personal ,
pro j ect s .

1 -2 Introduction
 1

Chapter Summaries

Chap Title Summary

No.
2 System Startup What happens from the time you restart your system to the
time you get a login : prompt .
3 System Different methods of shutting down a system and what
Shutdown happens when you do so.
4 Login What happens from the time you respond to the login :
prompt to the time you get a shell prompt.
5 Process What processes are, process states, commands for managing
Management processes, and a discussion of multi-processing.
6 Run-Levels What run-levels are and how they are set .
7 Memory How HP-UX implements demand-paged virtual memory,
Management basics of physical memory and swap-space management.
8 HFS File Overview of the file system characteristics, relationship to
System physical disks, file-system creation, storage, modification,
and protection.
9 Logical Volume Conceptual discussion of a Series 800 subsystem for
Manager managing disk space.
( LVM )
10 System Survey of the various HP platforms for HP-UX-Series 300,
Architectures 400 , 700, and 800 platforms.
1 1 System Descriptions of the HP-UX kernel, device drivers, and device
Configuration files, and their combined role in system configuration.
12 HP-UX Information on tape and disk drives, CD-ROM file system,
Peripherals terminals , and modems.
13 Networking Pointers to networking documentation.
14 System Discussions of various approaches to account for computer
Accounting usage.
A Glossary Definitions of all pertinent terms used in the manual.

Introduction 1 -3
1

The System Administration Manager (SAM)

The System Administration Manager ( SAM ) is a user interface for performing
most routine administrative tasks without using the underlying HP- UX
commands. For maximum flexibility, the System Administration Tasks manual
documents both SAM and commands methods of performing these tasks .
SAM enables you to manage:
• Peripheral devices
• File Systems and disk space
• Routine tasks, such as backup , recovery, and user accounts
• Processes , including control, and monitoring
• Kernel and cluster configuration
• Network connectivity
• Remote administration
• Auditing and security
SAM in HP-UX 9.0 works on a new, object-oriented paradigm: You list or view
information in a functional area and choose an action to perform your task.
To run SAM, type:
/usr/bin/s am
Activating the � button from a dialog or message box gives you information
about the attributes and tasks you can perform from the currently displayed
window.
Choosing an item from the Help menu within a functional area gives you
information about :
• the current functional area
• keyboard navigation within SAM
• using the SAM help system
• displaying the version of SAM you are currently running
Pressing the @) key gives you context-sensitive information for the object field
at the location of the cursor.
See Chapter 1, "Introduction to System Administration", in the System
Administration Tasks manual for detailed information about using SAM.

1 -4 Introduction
 1

Related System-Administration Manuals

How HP- UX Works supplements information in the following procedural and
reference manuals:
• System Administration Tasks manual.
Provides detailed procedures for most system administration tasks .
• HP- UX Reference.
Contains specifications of all HP-UX commands, system calls, library
routines, and system files.
• Installing and Updating HP- UX .
Describes how to install (or reinstall ) HP-UX, and how to update HP-UX
from a previous release.
• Installing Peripherals.
Discusses how to install and configure peripheral devices on your HP-UX
system.
• Solving HP- UX Problems .
Provides guidance for troubleshooting problems in key areas of HP-UX.
• Managing Clusters of HP 9000 Computers (Series 300/400/700 only).
Details how to set up , configure, and maintain computers belonging to an
HP-UX cluster.
The following books will help you acquire basic HP-UX skills:
• Using HP- UX, HP part number B2910-9000 1 .
• HP Visual User Environment User's Guide, HP part number B l 171-90061 .
• The Ultimate Guide to the vi and e x Text Editors, ISBN 0-8053-4460-8.
For pointers to networking documentation, refer to Chapter 13, "Networking
Overview" of this manual.

Introduction 1 -5
 2

2
System Startup

From the time you boot your system to the time you get a login : prompt ,
the system performs several important tasks automatically. The system tests
computer hardware, loads and initializes HP-UX, communicates messages
to users, and runs scheduled routines . When these system startup tasks are
finished, HP-UX is in the initdefault run-level. As shipped, this is run-level 2
or multi-user HP-UX.

Note In case of trouble when starting up Series 300, 400, or 700

systems , you can boot from a recovery system ( a minimal
system kept on DDS media or cartridge tape ) . Be sure to
make a recovery system as soon as you verify that the system
is installed correctly. Refer to Chapter 2, "Constructing an
HP-UX System" of the System Administration Tasks manual
for procedure and mkrs(lM) of the HP- UX Reference for
explanation of command syntax.

System startup occurs in two phases:

• The Boot ROM startup sequence.
• The HP-UX startup sequence.
As these phases occur, a series of messages appears quickly on the console
screen. You can review the HP-UX messages once you have a login prompt by
invoking the command I etc/ dmesg with superuser privilege.
The Boot ROM startup sequence is different for each architecture, but the
HP-UX startup sequence is virtually identical.
This chapter describes the Boot ROM startup sequence for Series 300/400, 700,
and 800 computers, and the HP-UX startup sequence. By understanding them,
you can tailor system startup to your unique needs.

System Startup 2-1

2

Boot ROM Startup Sequence

Overview
Conceptually, the boot program performs the same basic functions regardless of
architecture. When the system is powered up, the boot program initializes and
tests hardware to bring the system into a usable state by the operating system.
This includes:
•test LEDs
• determine processor type
•initialize and test timers
•find system console and initialize video circuitry
•load configuration data from EEPROM
•reset and initialize 1/0, including audio and HP-HIL
•display copyright and Boot ROM banners, CPU type, and EEPROM status
•test utility chip
• test RAM, report amount of memory found and any failures
•test DMA
•find and report built-in interface circuits
•test SCSI and LAN interfaces
•display boot choices
After checking the viability of the system, the boot program searches for
a copy of the operating system (/hp-ux). It searches a list of potential
sources , including disks and network (LAN) , as shown by Figure 2-1. (LIF
implementation differs, depending on architecture, but the basic concept
remains true.)

2-2 System Startup

 2

Boot Program

LAN

LG200211_013

Figure 2· 1 . The Boot Program Searches for an Operating System

From the first available source it finds, the boot program also locates another
program, called a secondary loader, and loads it into memory. The secondary
loader then loads /hp-we into memory and starts it up to enable you to use
your system.

Secondary Loader
The boot program contains code to work with a specific hardware architecture.
The secondary loader is a larger program, whose additional code provides
the flexibility to deal with the changes to the booting process from
operating-system release to release. (From the perspective of the boot program,
the terms operating system and secondary loader are synonymous.)
Further, the secondary loader is written in LIF (Logical Interchange Format) ,
code that i s understood by HP 1000, 3000, and 9000 systems.
The boot program finds the secondary loader on the boot media (typically in
the LIF volume of a mass storage media) , loads it into memory, and starts it
running.

System Startup 2·3

2

Boot ROM Startup Sequence

(Series 300/400)
The Boot ROM is a small machine-language program that resides in your
computer's ROM (Read-Only Memory) on the SPU (System Processing Unit) .
When you boot (or power-up) the system, the computer starts the Boot ROM
program, which causes the CPU to execute instructions that get the system
running.
On Series 300/400 computers, the Boot ROM performs several steps during
system startup:
• Locate and write to a system console.
• Search and test hardware internal to the SPU.
• Find, load, and start an operating system (in this case, HP-UX) .
• Optionally, configure internal interfaces in the SPU.

Boot ROM Finds a System Console

The system console is the terminal (keyboard and display) to which the Boot
ROM and HP-UX communicate with the system administrator. Boot ROM
error messages, HP-UX system error messages, and certain system status
messages appear on the system console screen. Sometimes (such as during the
administrative run-level) , the system administrator can communicate with
HP- UX only through the system console.
The Boot ROM searches for a system console until it finds:
1. An RS-232C serial interface with its remote bit set .
Ifthe Boot ROM finds more than one serial interface card with its remote
bit set , it selects the card with the lowest select code. In the case of the HP
98642A (four-channel multiplexer) , port 1 is used.
2. An internal bit-mapped display. There can be only one internal bit-mapped
display. Note that an HP 98700H display station can have its display
interface card (HP 98287A or HP 98720) configured for internal or external
control; if configured for external control, it is never chosen.
3. An external bit-mapped display with select code 132-255. They are selected
in order of increasing select code.
4. An HP 98546A compatibility video interface.

2-4 System Startup

 2
5 . An RS-232C serial interface without its remote bit set . If more than one is
present , the one with the lowest select code is used. In the case of the HP
98642A (four-channel multiplexer) , port 1 is used. The Boot ROM does not
recognize the HP 98628 serial interface card as console when this condition
is met; however, HP-UX does.
HP-UX communicates with the system console through the special device
file, /dev/console . Although the boot ROM can function without a system
console, HP-UX cannot . Therefore you must have a system console in one of
the locations specified below to boot HP-UX.

Serial Interface Line Control Switch Pack Settings

When a serial interface card is selected as the system console, the boot ROM
requires that the line control switch pack settings on the card be set to the
same settings as the connected remote terminal. These settings provide the
communications protocol for the console. Assuming an RS 232-C console,
HP-UX resets the following values :
Stop Bits 1
Baud Rate 9600 bps
Parity /Data O 's/7
Bits
Enq/Ack NO
Pace XON/XOFF
(Handshake)
If the serial interface is part of the built-in Human Interface card, these values
are set appropriately and you need not change them. Other configurations
might require different line control switch pack settings on the interface card;
see the documentation that comes with the card or the console for information.

Boot ROM Tests Hardware

The Boot ROM searches for and tests hardware internal to the SPU, including
interface cards , RAM, and internal peripherals and chips. It displays names of
installed peripherals and internal hardware. If hardware fails a test , the Boot
ROM displays an error message. For example, two interfaces set to the same
select code will cause a boot ROM error. For details on the tests performed,
refer to your system's hardware Installation Reference.

System Startup 2-5

 -

Boot ROM Finds and Loads an Operating System

As explained earlier, much of the code in the Boot ROM consists of the
secondary loader, which executes during system boot and is able to load
the HP- UX system kernel, despite variations from release to release. Using
the secondary loader, the Boot ROM searches for an operating system on
connected mass storage media, or on cluster servers on your computer's local
area network (LAN ) .
The Boot ROM has two modes for selecting an operating system: unattended
and attended. In unattended mode, the Boot ROM automatically boots the
first operating system found. Unattended mode is the default method for
loading an operating system. In attended mode, you select the operating
system to boot from all operating systems found.
If, for any reason, the Boot ROM is unable to find and load an operating
system, it displays messages to that effect on the system console. These
messages are described in the "Boot ROM Error Messages" section of your
computer's Installation Reference.

Unattended Mode

Use the unattended mode of booting if you have only one bootable operating
system online, or if you know the operating system you wish to boot is the first
operating system the Boot ROM will find.
The Boot ROM searches for an operating system by a prioritized list of select
codes. This includes LAN cards, disks, and other peripheral devices . The first
operating system found on one of these devices is loaded. If no operating
system is found, the list will be searched again until a system is found. This
means that disks not found at power-up will be found after their initialization
is complete.
To load HP-UX using unattended mode:
• Make sure that HP-UX is the first system found, following the prioritized list
below.
• Make sure the device holding the HP- UX operating system is fully powered
up and has achieved a ready state before booting the computer. If the disk
drive has not completed its power-up sequence ( which may require several
minutes on some disk drives ) , the Boot ROM will not be able to access the

2-6 System Startup

 2
disk and load the system. If you are booting a cluster client , make sure the
cluster server has completed the bootup sequence, including clustering and
boot daemons .
Boot ROM Search Sequence.
The Boot ROM searches for an operating system in the following sequence:
1 . S C SI disks at select codes 0-3 1 , beginning from highest priority address
( that is, 7), then searching addresses in decreasing numerical order ( that is,
6, 5, and so on ) .
2. HP-IB disks at select codes 0-3 1 , bus address O , unit O , volumes 0 through
7. ( Note, however, that the volume scan differs on older Series 300
systems. )
3 . Shared Resource Managers ( S RM ) at Select Code 2 1 , Volume 8.
4. LAN interface at Select Code 2 1 .
If there are multiple remote servers at the same select code ( as i n an
HP-UX cluster ) , the first server to respond is the first system listed on the
menu ( if attended mode ) or the remote system selected ( if unattended
mode ) . This can change each time you boot your system.
5. HP 98259 ( Bubble Memory ) at Select Code 30.
6. HP 98255 ( EPROM card ) on Unit 0.
7. ROM systems ( for ROM-based operating systems ) .
8. Remaining S C SI devices at Select Codes 0 through 3 1 bus addresses
( device ID ) 4-0, units 0- 16. ( note: the search order on S C SI addresses is
high to low. )
9. Remaining HP-IB devices at select codes 0�3 1 , bus addresses 0-7, units
0- 16, volumes 0- 7.
10. Remaining SRMs at Select Codes 0-31 ( not at Select Code 21, Volume 8).
11. Remaining LANs ( not at select code 21) .

12. Remaining HP 98259 ( Bubble Memory ) on Select Code 0-29 and 3 1 .

1 3 . Remaining HP 98255 ( EPROM ) units.
Each category in the search sequence might have multiple units at the same
select code and bus address, all of which are searched before moving to the

System Startup 2-7

2
next ascending select code or bus address. Thus, an HP-UX system on the
root mass storage device at select code 14, bus address O , unit 0, is found and
loaded before any operating systems at select code 14, bus address O, unit 1 .
Boot Search Criteria.
The boot search sequence given is accurate for Revision 2.0 Boot ROM .
Different revisions of the Boot ROM might vary slightly. Regardless of revision,
however, the boot search sequence was developed on the following guidelines:
1 . High-to-low priority addresses
2. Operating systems external to the SPU before ROM- or EEPROM­
embedded operating systems
3. Local systems (dedicated devices) before remote systems (shared devices,
such as a LAN server)
4. Removable media before non-removable media
5. Internal systems before external systems (Internal refers to operating
systems on integrated devices)
6 . Small-capacity devices before large-capacity devices
Other points worth noting:
• You must have Boot ROM rev. C or later to boot from a SCSI disk drive.
At rev. C , S CSI disk drives are treated the same as any other disk drive.
• You must have Boot ROM rev. B or later to boot across a local area network
(LAN ) , as in the HP-UX clustered environment .

2-8 System Startup

 2
Series 400 Unattended-Mode Boot variation.
Series 400 systems provide additional flexibility during unattended mode. If
the Series 400 has never been booted, a Configuration Control menu appears,
which allows you to:
• Boot automatically from the first operating system found, or
• Specify an operating system by using Auto System Selection.

Configuration Control
Keys Mode Class
1 I/O Configuration
2 Auto System Selection
3 Boot Mode Selection

A Abort without changes

Type [key] RETURN ?
The I/O Configuration key is explained later in this chapter, since it applies to
some Series 300 systems, as well as all Series 400 systems.
You can use the Auto System Selection key to override the Boot ROM search
priority and specify which system to boot. Auto System Selection resembles
Attended mode booting, but here, the Boot ROM will remember your choice
for future use. You can change your system selection by re-entering this menu
at a later time.
The Boot Mode Selection key allows you to choose either HP-UX or DOMAIN
operating system. ( DOMAIN is available on the Series 400 only. )

Attended Mode

If you want to choose which specific version of the operating system to be

used by the Boot ROM , you must enter the attended mode of selection. For
example, when you are installing HP-UX, you need to boot from the install
media, so you would use attended mode.
To enter attended mode, type a space, (Retijffi), or any letter or number during
the time:

System Startup 2·9

2
• After the keyboard has been initialized ( the keyboard is initialized when the
word Keyboard appears on the system console )
• Before the Boot ROM finds a default system.
The character used to enter attended mode is used as part of the string to
select the operating system.
The best way to enter the attended mode is to hold down the space bar until
the word Keyboard appears in the installed interfaces list on the left side of the
screen.
Once you enter attended mode, the Boot ROM displays a list of one or more
loadable operating systems , along with their mass storage device name and
address. The following HP- UX operating systems are possible:
SYSHPUX contains a pointer to the file, /hp-ux , which is the kernel, or
compiled core program of the HP-UX operating system.
SYSDEBUG used only for device driver writing.
SYSBCKUP contains a pointer to the file /SYSBCKUP, a backup kernel.
In unattended mode, the Boot ROM loads SYSHPUX as the HP-UX operating
system.
Ifyou are booting a cluster client , the cluster server ( or servers ) will look like
this:
: LAN , 2 1 , cluster_server_ name
To select one of the operating systems listed, enter the two characters to the
left of the operating system name and press ( Return ) . Do not use the (Shift )
key as the characters you type are automatically shifted to upper case. For
example, if the Boot ROM displays this list :
: LAN , 2 1 , your_system_name
1H SYSHPUX
1D SYSDEBUG
1B SYSBCKUP
to select SYSHPUX, you type C!Xill , and 1H appears at the lower-right corner
of the screen.
For details on using attended mode, see the System Administration Tasks
Manual and appendices of Installing Peripherals Manual.

2-1 O System Startup

 2
Boot ROM 1/0 Configuration

The Boot ROM of Models 345, 375, and all Series 400 systems allow you to
change the characteristics of some built-in I/O interfaces .
To select the I/O Interface Configuration menu, enter the key sequence
( Ctrl- C I Return ) , instead of using the space bar, to enter attended mode.
The I/ 0 Interface Configuration menu displays a list of configurable interfaces
( depending on which interfaces are installed in your SPU ) . When you make
a selection, a menu specific to the chosen interface appears with appropriate
configuration choices.
Depending on interface, configuration choices include the following kinds of
settings:
Select Code Set value for most interfaces . ( Select code for
standard HP-IB cannot be reconfigured. )
Interrupt Level Set value for most interfaces.
Remote/Local Convert a serial terminal to system console.
Fast/Normal Set data transfer or handshake speed.
Modem Enable modem handshaking for serial ports.
Parity Check parity of data for SCSI transfers .
Bus Address Set value for S C S I controller bus address.
HP-IB System Controller Enable or disable.
DMA Mode Set the width of data transfers.
Once made, the configuration values are stored in EEPROM, reprogrammable
Electrically Eraseable Read-Only Memory. Like the Boot ROM, the contents of
EEPROM are retained even if power is removed from the SPU.
Note that not all settings apply to every interface. For details on reconfiguring
your internal interfaces , consult BootROM Configuration Mode User's Manual
and Installing Peripherals Manual.

System Startup 2-1 1

2
HP-UX Takes Control

Once the operating system is loaded, the Boot ROM passes program control to
the operating system. The operating system then controls the system until you
re-boot the system. A later section in this chapter, entitled "HP-UX Startup
Sequence," describes what HP-UX does between the time it takes control and
the time you see the login : prompt.

2· 1 2 System Startup
 2

Boot ROM Startup Sequence

(Series 700)
The automatic boot process on the Series 700 resembles the Series 800 moreso
than the Series 300 or 400, due to similarities in the Series 700 and 800
processor-dependent code (PDC) .
Switching on the system or pressing the system transfer-of-control (Toe) button
causes PDC and I / 0-dependent code (IODC) firmware to be executed, to
verify hardware and system integrity. The PDC runs self-tests and locates the
console, using IODC and the paths stored in Stable Storage. The PDC displays
on the console screen copyright information, PDC and IODC ROM revisions,
and amount of memory configured. ( See the manual page for pdc(lM) in the
HP- UX Reference.)
Once checking i s complete, one of two things happen:
• If AUTOSEARCH is off, PDC searches for a potential boot device, as described
in the section, "Attended Mode," which follows.
• If AUTOSEARCH is on, PDC boots from Stable Storage, or searches for an
appropriate path for console and boot , if any paths in Stable Storage fail.
PDC displays the following to the console:
Selecting a system to boot .
To stop selection process , pres s and hold the Escape key .
If the � key is not pressed, PDC loads the initial system loader ( ISL ) from
the primary path in Stable Storage ( for example, a SCSI disk with address
s c s i . 6 . 0 ) and transfers control to it.
The ISL , scheduled by the autoboot sequence, finds the autoexecute
file and executes the command specified in it ( usually hpux or hpux boot
disc ( ; 0) /hp-ux ). By default, the autoexecute arguments load /hp-ux into
memory and execute.
While loading, the secondary loader displays information about the device and
file being booted, the text , data, and B S S size of the kernel and the kernel's
startup address.
If necessary, you have 10 seconds to override the default autoboot sequence by
pressing the (ESCap;) key and enter Attended Mode.

System Startup 2- 1 3
 --

2
For details on the Series 700 boot capabilities, including boot sequences from a
variety of devices , autoboot , and restore, see the manual page hpux_ 700 ( 1M )
in the HP- UX Reference . Also, see the Owner's Guide for your Series 700
system for information on secondary loader enhancements.

Attended Mode
Each time the computer is powered on, you have the opportunity to interact
with it by entering Attended Mode. For example, you might need to interrupt
the boot sequence for either of the following reasons:
• To redirect the boot sequence.
• To perform a boot administration function provided by the Boot Console
User Interface.
Pressing the (ESCaaj key halts the automatic boot sequence and puts you into
Attended Mode. The system searches the S CSI, LAN , and EISA interfaces for
all potential boot devices-devices for which boot I / O code ( IOD C ) exists.
The system then displays a table, such as the following:

Device Select ion Device Path Device Type

PO scsi . 6 . 0 CD-ROM HEWLETT-PACKARD

P1 lan . 123456-789abc hpwgux42
P2 eis a . 1 . 0 . 0 . 0 . 0 . 0 SCSI Disk Drive
At this point , the Boot Console User Interface main menu offers the following
options :
b) Boot from specif ied device
s) Search for bootable devices
a) Enter Boot Administration mode
x) Exit and cont inue boot se quence
?) Help
Using the b) option, you can direct hpux to boot from a specific device. For
example, to direct hpux to boot from the S CSI disk drive, you would type the
key sequence (EE .
Using the s ) option of the Boot Console User Interface main menu, you direct
the system to search through the list of potential boot devices for only those

2-14 System Startup

 2
that have a LIF volume and initial program loader (IPL). The system then
displays a table listing only those devices from which you can boot .
For example, a system might return the following selection:

Device Select ion Device Path Device Type & Ut ilit ies

PO scsi . 6 . 0 Quantum PD2 10S

IPL
P1 scsi . 5 . 0 Quantum PD2 10S
IPL
P2 scs i . 1 . 0 HP 22 13A
IPL

Boot Administration Mode

Using the a) option of the Boot Console User Interface main menu, you can
enter Boot Administration mode, from which you can alter default behaviors
exhibited at boot-up or obtain useful information about the hardware as
configured. The following commands are available at the Boot Administration
Mode menu:
AUTO Display stat e of Aut oboot /Aut osearch flags
AUTO SE.ARCH Set state of Aut osearch flag
AUTOBOOT Set stat of Aut oboot £lag
BOOT Boot from Primary/Alternat e path or Specified Device
DATE Read/Set the Real-Time Clock
EXIT Return t o previous menu
FASTSIZE Display/Set FASTSIZE memory parameter
HELP <it em> Display Help informat ion for < item>
IIFO Display boot/revis ion informat ion
Lil_ADDR Display LAI Stat ion Address
OS Display/Select Operat ing System
PATH Display/Kodify Path In£ormat ion
Plll_IRFO Display Processor Internal Kemory In£ormat ion
RESET Reset the System
SEARCH Search for boot device
SECURE Display/set secure boot mode
SHOV Display the results of the previous search

For detailed information on using the Boot Administration mode, see the
Owner's Guide for HP- UX Users for your Series 700 system.

System Startup 2- 1 5
2

Boot ROM Startup Sequence

(Series 800)
When more than one operating system is present on the system's mass storage
devices , both the location of the operating systems and the type of media
on which they are stored determine which operating system is loaded. The
primary boot path in Stable Storage determines the default boot path. Stable
Storage is the memory in Series 800 computers reserved for maintaining critical
configuration parameters used during system boot . For example, the primary
and alternate boot paths, console path, and autoboot settings are stored in
Stable Storage.
The Boot ROM contains general-purpose software developed to support both
present and future Hewlett-Packard operating systems . This section describes
how the ROM boots the HP-UX operating system.
When you turn the computer on, the Boot ROM goes through the following
sequence:
1. Performs System Processing Unit ( SPU ) self-test .
2. Reads the console path from Stable Storage, tests it , and assigns a display
terminal for use as a system console.
3. Reads the boot device path from Stable Storage, searches for the paths on
Model 825, 835, and 845 computers using a capability called "autosearch" ,
or lets you enter the path from the system console and tests the boot device
path.
4. Loads the Initial System Loader ( ISL ) into memory.
The media on which your bootable system resides has a boot area and a root
partition. The boot area contains the bootstrap program and other files needed
for bringing up the system. See isl ( lM ) in the HP- UX Reference for more
information on these files. The disk section where the boot area resides is
/dev/dsk/c0d0s6 . The root partition contains a file, called hp-ux, which is the
operating system. The typical disk section for the root partition on a system
using an HP-IB interface card is /dev/dsk/cOdOsO.

2- 1 6 System Startup
 2

System Boot Overview

At powerup or reset:
1 . The processor selftest is executed from PDC ( Processor Dependent Code ) .
2. The PDC locates the system console and tests the console path using
I / 0-Dependent Code. The boot and console paths are kept in Stable
Storage. Stable Storage memory is static RAM.
3. The PDC checks to see if the autoboot flag is set. If not , the system asks
for a boot path. If set , the system prompts the console with the following
message:
Boot ing from def ault lpath
To interrupt , pres s any key within 10 s econds
If you press a key the system asks if booting will be done from the primary
autoboot path, the secondary autoboot path, or ask for an alternate path.
4. The PDC loads the ISL ( Initial System Load ) . The ISL loads an operating
system ( HP-UX in this case ) .
Once the operating system is loaded, the Boot ROM passes program control to
the operating system. The operating system then controls the system until
you re-boot the system. The section in this chapter calle d "HP-UX Startup
Sequence" describes what HP-UX does between the time it takes control and
the time you see the login : prompt.

Boot ROM Search Sequence (Series 800)

The Boot ROM initializes the primary boot path, loads ISL , and allows you
to select either the manual or autoboot mode. On Model 825, 835, and 845
computers, an additional mode called autosearch is available when autoboot is
enabled. In manual mode, you can select the boot device from all the available
peripheral devices. In autoboot mode, the Boot ROM automatically boots the
operating system from the primary boot path defined in Stable Storage.

Autoboot

You should use autoboot except for first-time installation and operating system
reconfiguration. The ISL autoboot on command enables autoboot . No reboot

System Startup 2-17

2
or automatic reboot on panic is possible. Panic is a condition when the system
becomes inoperative due to an abnormal condition detected by the kernel.
Before using autoboot , make sure the boot device is fully powered up and
ready for operation before you turn on your computer. Autoboot is selected if
you let the 10 second override period expire.

Manual Boot

Manual boot can be entered by pressing any key during the 10-second override
period in the beginning of the autoboot sequence. When manual mode is
activated, the Boot ROM prompts for the path to be used. The primary boot
path is not altered or disabled.
If you do not want to boot automatically from the primary boot path during
the boot process, disable the autoboot flag in Stable Storage. You can do this
using the ISL auto boot off command. However, this is not normally done
because the autoboot feature makes your system administration tasks more
efficient . Disabling autoboot requires your intervention each time the system is
rebooted. To re-enable autoboot , use the ISL autoboot on command.

2· 1 8 System Startup
 2
Autosearch

Many Series 800 computers have a feature called autosearch. If the system
cannot locate the console using the console path from Stable Storage, it
searches for a console device. Then, if the system cannot locate the boot device
using the primary or alternate boot paths and the autosearch flag is set, the
system continues to search for a boot device.
The autosearch flag is much like the autoboot flag. It is in Stable Storage and
can be enabled or disabled. Use the ISL autosearch on command to enable
autosearch and the ISL autosearch off command to disable the feature.
When autosearch is invoked ( if the two paths specified in Stable Storage fail ) ,
the following messages appear on the console:
Autosearch for boot device enabled .
To override , press any key within 10 seconds .
If you press a key, the system responds:
Do you want to cont inue an interact ive search? (Y or N) ?>
If you answer "no," autosearch halts at that point and proceeds to manual
boot . If you answer "yes," the system searches for a boot path, states it,
and asks you if you want to boot from it . If you respond "yes," the system
uses that path. Otherwise, it presents other logical paths until you respond
positively or until it finds no other paths. It then proceeds to manual boot.

System Startup 2-19

2

HP-UX Startup Sequence

Once HP-UX takes control from the Boot ROM, it performs two tasks:
• Find the root file system.
• Start the init process and bring the system to run-level 2, which is
multi-user HP-UX.

HP-UX Finds the Root File System

Once HP-UX starts, it searches for the root file system. The root file system is
the portion of the file system that forms the base of the file system hierarchy­
that is, the portion of the file system on which other file systems can be
mounted. The root file system contains the critical files (such as the kernel
in /hp-we) . Generally, the root file system is found on the disk from which
HP-UX booted. For cluster clients, the root file system is on the cluster server.
(Note, clusters are not supported on Series 800 as of HP-UX 9.0.)
After finding the root file system, the operating system starts a shell to read
commands from / etc/pre_ init _rc . Among these commands is fsck( lM),
which checks the root file system (currently in a read-only state) . (fsck is
discussed in Chapter 8, "HFS File System" of this manual and in Chapter 6,
"File System Problems" , in Solving HP- UX Problems.) Any problems fsck
encounters that it is unable to fix are caught later by I etc/bcheckrc . After
fsck exits, the operating system remounts the file system in a read-write state.

Caution Do not modify the /etc/pre_ init_rc script ; it might cause

the system to be unbootable.

HP-UX Starts the init Process

Then, HP-UX starts its first process , /etc /init . The init process has process
ID one ( 1 ) and no parent process.
The init process reads the /etc/ inittab initialization file to define the
environment for normal working conditions .

2-20 System Startup

 2
System Initialization File -/etcfinittab

The ini t process reads the I etc/ ini ttab file one line at a time, each line
containing an entry that describes an action to take.
The syntax of ini ttab entries is:

id: run-levels : action : process

id A one- or two-character ID that uniquely identifies the entry.
run-levels Defines the run-levels in which the entry can be processed. You can
specify multiple run-levels. If this field is empty, all run-levels are
assumed.
Typically, entries tell init to run a process at specific run-levels . If
no run-levels are specified, the process can execute in any run-level.
For example, the following entry tells ini t to run the I etc/ getty
process in all run-levels:
cons : : respawn : / etc/getty console H
action Identifies what action to take for this entry. The actions are as
follows:
sys ini t Performs system initialization on devices needed
by init for obtaining run-level information at
the console, such as tty characteristics . sys ini t
entries must finish executing before / etc/ initt ab
continues.
ini tdef aul t Causes the initial ( default ) run-level to be the value
of the run-levels field. If more than one run-level
is specified in run-levels , ini t uses the highest
specified run-level.
boot Run the command specified in the process field
at boot-time only. Do not wait for process to die
before reading the next entry.
Before enabling other users to access the system,
init executes all /etc/inittab entries marked
boot or bootwai t . These processes are known as
boot processes.

System Startup 2-2 1

2
bootwait Run the command specified in the process field
at boot-time only. Wait for process to die before
reading the next entry.
The following entries tell ini t to execute
the recoversl program and brc shell script ,
respectively. ini t must wait for each process to die
before moving to the next entry.
slib : : bootvait : /et c/recoversl </dev/console >/dev/console 2>t1 #shared libs
brc2 : : bootvait : /etc/brc >/dev/console 2>t1 #boottime commands

wait On entering the run-level that matches the

run-levels field of this entry, run process and wait
for it to die before reading the next entry.
respawn On entering the run-level that matches the
run-levels field of this entry, run process if it is not
already running. Do not wait for process to die
before reading the next entry. When process dies ,
run it again.
process This is a shell command to be run, if the entry's run-levels matches
the run-level and/or the action field indicates such action.
HP-UX recognizes any text following the # symbol as commentary.
Each system architecture has its own /etc/ inittab file, with some unique
architecture-specific /etc/ inittab characteristics . For example, on the
Series 800 , first /etc/ inittab calls io init , which initializes kernel 1/0 data
structures using information from /etc/ ioconf ig and then call s insf to assign
logical unit numbers and create special files for all new devices on the system.
The /etc/ inittab entry is as follows:
io : : sys init : /etc/io init - i >/dev/ console 2>&1
Series 300/400/700 systems do not require this initialization of kernel 1/0 data
structures.

Default Run Levels-initdefault

The / etc/ inittab file sets up system run levels . The entry marked
initdef ault sets the default run-level to 2:

2-22 System Startup

 2
init : 2 : initdefault :
( See Chapter 6, "Run-Levels" for information about run levels. )
Among the /etc/ inittab entries , are programs that ensure system integrity
and set up essential processes. These programs are discussed in the next
subheads.

Boot Check Run Command -/etc/bcheckrc

If your system is using disk mirroring ( an optional feature of the Series

800 only ) , the /etc/bcheckrc ( Boot Check Run Command ) program calls
I etc/mirrorrc as its first operation. mirrorrc configures mirrored disks, runs
f s ck, sets up logging for the mirrored disks, and reimages the disks.
If you are implementing the Logical Volume Manager ( LYM ) , bcheckrc then
calls I etc/1 vmrc to activate LYM volume groups.
On the Series 700, if appropriate, bcheckrc runs eisa_conf ig.
On all systems , / etc/bcheckrc verifies that the system was properly shut
down. To determine this, bcheckrc calls the fsclean program, which checks
each file system of type hfs in / etc/checklist for consistency. fsclean looks
at a flag called the clean byte in the primary superblock of each file system.
When a file system is created, the clean byte flag is set to FS_CLEAN . When the
file system is mounted ( using the mount command ) , the clean byte flag is set to
FS_OK. During a normal shutdown ( that is, during execution of the reboot or
shutdown command ) , the clean byte is reset to FS_CLEAN. Thus, under normal
conditions, the file system can be unmounted and set to FS_CLEAN , or mounted
and set to FS_OK.
If it encounters the file system unmounted and its clean byte set to FS_OK,
fsclean perceives the file system to be in an inconsistent state ( due to a crash
or other incorrect shutdown ) . In this case, fsclean exits and bcheckrc runs
/etc/fsck - P , f s ck in preen mode. This will correct most errors found.
If fsck run from bcheckrc fails for any reason, bcheckrc starts a shell with
the prompt ( in bcheckrc) # , instructing you to run fsck interactively. If this
occurs, you must run f s ck to ensure the integrity of your file system.
Some file system problems must be fixed in this way to minimize risk of data
loss. After running fsck interactively, you may be instructed to reboot the
system. If so, reboot the system using reboot -n to ensure your system 's

System Startup 2-23

2
integrity. If f sck does not tell you to reboot , exit the shell by typing @ill).
This returns control to bcheckrc.
bcheckrc then ensures that /lib is not a mounted volume. Since HP-UX uses
shared libraries, /lib must be on the root partition, and thus available to all
commands that need them to boot the system.
Finally, bcheckrc mounts all type hf s file systems and exits.

Shared Libraries Check -/etc/recoversl

After bcheckrc exits , ini t call s recoversl to check for the existence of
shared libraries critical to the system. recoversl checks for ( and corrects, if
necessary ) the owner, group , and permissions of these shared libraries.
The shared libraries considered critical to the system are dld . sl ( the dynamic
loader ) and libc . sl. If any of these libraries are missing or damaged,
recoversl guides the system administrator through procedures to recover the
shared libraries from update media.

Additional Boot Tasks -/etcfbrc

The I etc/brc program does these tasks:

• Sets the system PATH variable, which defines the default command search
path for all users.
• Downloads the floating point microcode if your system is set up for the
floating point accelerator. ( Series 300 only )
• Removes the /etc/rcflag file ( used later as a check for startup condition ) .
• Looks for and removes the I etc/mntt ab file on a standalone or cluster server
system. The / etc/mnttab file, a list of mounted file systems , is recreated
later when I etc/re remounts the file systems.
• Creates a new /dev/crt file, if needed. ( Series 300 only. )

Initial Customization Script-/etc/rc

The following entry in /etc/ inittab invokes /etc/re:

rc : : wait : /etc/rc >/dev/ console 2>& 1 # system init ial izat ion

2·24 System Startup

 2
The I etc/re script consists of a main script program and several shell
functions (subroutines in a shell program).
The main program checks to see if it is boot time according to whether the
/etc/rcflag is present or absent . (If the boot process has proceeded properly,
I etc/brc will have removed I etc/ref lag.) I etc/re determines whether
to perform system initialization and start various daemon (background)
processes. The actual tasks performed by I etc/re depend on the configuration
of the machine (see Table 2- 1 ) . / etc/re determines if you are running as
a standalone system, a cluster server, or a cluster client . It then calls shell
functions applicable to your system.
When finished, I etc/re calls a shell function called localrc. This is the shell
function you should use to customize the /etc/re script . It should contain any
tasks you wish to perform that are not part of the standard I etc/re functions.
In the localrc shell function, you can add commands you wish to perform
every time the system is booted or whenever there is a change in run-level
which init does not handle.
For each system state (standalone, cluster server, or cluster client) , Table 2- 1
shows what functions /etc/re does. Note that some commands (such as
networking commands) may not be available or installed. I etc/re checks for
the existence of all such commands before attempting to run them.
The ini t process waits until I etc/re dies before processing the next entry in
·

/etc/ inittab .

System Startup 2-25

2
Table 2- 1 . /etc/rc functions

Function Stand- Root Disldess

alone Server Cnode 1
Mount all hf s volumes listed in the /et c/checklist x x x
file .
Initialization: set TZ and other variables; set device x x x
files used by I et c/rbootd, set up vt gateway.
Set host name in variable SYSTEM_NAME . x x x
Set the date . x x
Save a core image of a previously crashed system. x x x
Start swapping to all swap devices in I et c/ checklist x x x
Start the syncer . x x
Start the lp scheduler. x x
Clean up editor and uucp files. x x
Start networking. x x x
Start cluster server processes (/et c/esp) . x x
Start the remote boot daemon (/et c/rbootd) . x x
Start cron. x x x
Start pty allocation daemon . x x x
Start vtdaemon. x x x
List files found in /tmp and /usr/tmp directories. x x x
Clean up logging (/usr/adm/$LOG) files. x x x
Start diagnostic logging for the 1/0 subsystem (Series x x x
700/800 only) .
Start logging system messages (Series 700/800 only) . x x x
Start auditing processes. x x
Set local initialization functions to re. x x x

1 HP-UX clusters are not supported on Series 800 as of HP-UX 9.0.

2-26 System Startup

 2
Powerfail Routines -/etc/powerfail

A local powerfail is a power failure that halts the computer by affe cting its
central bus. The Series 800 HP-UX operating system provides a mechanism for
recovery from a local powerfail, to ensure that any program running on the
system at the time of failure can resume executing when power to the bus is
restored.
Powerfail routines are invoked from I etc/ ini ttab to ensure that power is
maintained to the system in case of emergency.
p£ : : povervait : /etc/poverfail >/dev/console 2>t1 #power fail rout ines

If need be, powerfail can be disabled by reconfiguring an operating-system

parameter as follows:
pf ail_ enabled O ;
Be sure to follow guidelines for correct shutdown and start-up of a system
necessitated by powerfail. These guidelines are given in Chapter 3, "Starting
and Stopping HP-UX" of the System Administration Tasks manual.
Note, although powerfail appears in all /etc/ inittab files , the entry is only
used on systems that support powerfail.

Terminal Processes Startup -/etc/getty

Once / etc/re has :finished its run-level 2 execution, control returns to init ,
which runs the commands from the process field of all run-level 2 entries in
/etc/ inittab . Typically, /etc/ inittab 's run-level 2 command field entries
consist of I etc/ getty commands, one for each terminal on which users log in.
( When you add a new terminal with the SAM utility, it automatically adds an
appropriate I etc/ getty entry to I etc/ ini ttab.) The I etc/ getty command
runs the login process on the specified terminal, allowing users to login on the
terminal.
For example, the following / etc/ inittab entry runs a getty at the system
console:
cons : 01 23466 : respavn : /et c/getty -h console console # system console

The respawn action field tells ini t to restart the getty process after it dies .
This means that each time you log off the system console, a new login :
prompt is displayed, so you can log in the next time. The 0123456 run-levels

System Startup 2·27

2
field indicates that init runs getty in run-levels 0 through 6. However, when
getty respawns other terminal login processes, it is often set up to run only in
run-levels 2 and 3 or only 2.
As shipped, / etc/ inittab invokes /etc/getty only for the system console. If
your system has additional terminals on which you wish to support logins, you
must add the appropriate getty entries to / etc/ inittab . ( Note, however, that
SAM automatically creates these entries when you use it to add terminals) .
The I etc/ getty command is the first command executed for each
login terminal. It specifies the location of the terminal and its default
communication protocol, as defined in the /etc/gettydef s file. It prints
the /etc/issue file ( if present) and it causes the first login : prompt to be
displayed. Eventually, the getty process is replaced by your shell's process ( see
Chapter 4, "Login" ) .
When you logout , the / etc/ init process is signaled and takes control again.
The init process then checks /etc/ initt ab to see if the process that signaled
it is flagged as continuous ( denoted by respawn) . If the process is continually
respawned, ini t again invokes the command in the command field of the
appropriate ini ttab entry as described above ( that is, the getty runs and a
new login : prompt appears) . If the process is not flagged as continuous, it is
not restarted.
At the end of /etc/ inittab is an entry that runs powerfail recovery
procedures in case of power failure.

Note Do not add /etc/getty entries to / etc/inittab for

unconfigured terminals, unless action is "off" .
If the system finds itself having to respawn entries too rapidly,
it assumes that a problem exists and goes to sleep for five
minutes before trying to respawn again. If the problem involves
the getty for the system console, the system might not be
bootable without repair.

Users can log in at all terminals for which getty processes have been executed.

2-28 System Startup

 2
For More Information

The sections in this chapter describe the default operation of the system as
shipped to you. However, by altering certain configuration or system files, any
of the procedures can change. If, for example, you write your own /etc/re
script , the paragraphs which follow may no longer apply.
Table 2-2 shows where to look for additional information.

Table 2-2. Additional Startup Information

To Learn More about . . . Refer to . . .

Processes Chapter 5, "Process Management"
Run-Levels Chapter 6, "Run Levels"
/etc/init Chapter 6, "Run Levels"
init ( lM) in the HP-UX Reference
/etc/inittab Chapter 6, "Run Levels"
inittab ( 4) in the HP-UX Reference
/et c/brc the I et c/brc file
brc(lM) in the HP-UX Reference
/etc/bcheckrc the I et c/bcheckrc file
/et c/re the I etc/re file
/etc/getty getty ( lM) in the HP-UX Referen ce
/et c/recoversl recoversl( lM) in the HP-UX Reference

System Startup 2-29

 3
3

System Shutdown

System shutdown is the process of taking your system from a running state
( either run-level 2 or run-level s ) to a halted state in which no processes are
running. This chapter discusses:
• Halting the system versus rebooting the system.
• Shutdown with the shutdown command versus the reboot command.
• What happens during shutdown.

Halting vs Rebooting
There are two levels of shutdown: halting and rebooting. Halting brings the
system to a complete stop ; in this state, the only way to restart the system
is to cycle power or reset the hardware. Rebooting brings the system to a
complete stop , but restarts the system as if you had booted it.
Whether to halt or reboot your system depends on why you want to shutdown
in the first place: If you want to shut the computer off for an extended period
of time ( for example, to add new hardware or to leave the system off for a long
weekend, etc ) , then halting is appropriate. If you want to shutdown only to
boot the system again ( for example, to use a newly reconfigured kernel ) , then
rebooting is appropriate.

System Shutdown 3-1

 Shutdown vs Reboot
After deciding whether to halt or reboot the system, you must decide which
command to use: shutdown or reboot . Which command you use depends on:
3
• Whether users are logged in
• How quickly you need to shut the system down.

When to Use shutdown

The shutdown command shuts down more slowly than reboot , but more
gracefully. It uses kill - 14 to kill running processes, which lets processes
terminate naturally within a grace period. This is the safest way to shutdown
and it helps ensure file system integrity.
It also displays messages directing users to log off within a specified grace
period. You can specify the grace period when you invoke shutdown. It is
typically used when:
• The system is in a multi-user state (for example, run-level 2) .
• The system administrator is not the only person logged in and using the
system.
In addition to shutting down HP-UX, shutdown can also bring the system to
run-level s ( single-user administration level ) .
For details on shutting down HP-UX using shutdown, see the System
Administration Tasks manual and shutdown(lM) in the HP- UX Reference.

3-2 System Shutdown

When to Use reboot
The reboot command shuts down all processes very quickly. It uses kill -9
to kill any running processes. This can be dangerous (for example, cause loss
of data) because reboot will shutdown processes immediately without letting 3
them terminate normally. It is typically used when:
• The system is in run-level s ( single-user administration level) . ( For a
description of run-level s and how to bring the system to run-level s, see
Chapter 6 and the System Administration Tasks manual.)
• You need to bring the system down very quickly.
Like shutdown, reboot allows you to specify a grace period. But unlike
shutdown, when the grace period is over, reboot still kills processes quickly
and indiscriminately ( with kill -9).
For details on shutting down HP-UX using reboot , see the System
Administration Tasks manual. For details on what reboot does, see reboot(lM)
in the HP- UX Reference.

System Shutdown 3-3

 Examples

Table 3- 1 . Examples of using shutdown and reboot

3 Example Description
shutdown -h Shutdown and h alt the system; give users
a ( default ) 60-second grace period in
which to logoff.
shutdown -h 360 Shutdown and halt the system; give users
a 6-minute ( 360-second ) grace period in
which to logoff.
shutdown -r Shutdown and reboot the system; give
users the default 1-minute ( 60-second )
grace period in which to logoff.
reboot -h Shutdown and halt the system
immediately.
reboot Shutdown and reboot immediately.
reboot -t +6 Shutdown and reboot in five ( 5 ) minutes.
reboot -t 22 : 06 -m "please logoff" Shutdown and reboot at 10:05 PM. As
10:05PM approaches, display the message
"please logoft'' at more and more frequent
intervals.

3-4 System Shutdown

 4
Login

Loginis the means by which users gain access to HP-UX. Beginner's Guide
to HP- UX and Using HP- UX describe how to login. This chapter describes 4
what HP-UX does during login-that is, what happens between the time you
type your usemame and the time you get a shell prompt. Understanding this
procedure, you can modify it for your unique needs. For example, you can:
• Set up logins so that a user gets an application instead of a shell.
• Change all users' default login environments.
• Change all users' default login shells.
Such tasks can be done manually (for example, by manipulating system
files and running various commands) or through the SAM program, which
automatically updates system files and calls appropriate commands for
you. For details on doing various tasks related to login, see the System
Administration Tasks manual.
Note This chapter does not explore the intricacies of logging into
a workstation implementing HP Visual User Environment
(VUE). VUE has its own display and other environmental
requirements, and thus reads its own configuration files. For
information on VUE, consult the HP Visual User Environment
(VUE) System Administration Manual.

Login 4· 1
 Overview of Login
Login begins with the I etc/ getty process spawned b y the ini t program.
The getty program prompts the user for a username (login : ) and password,
information which corresponds to the user's entry in the file / etc/passwd.
getty passes the username and password to login, which verifies the user's
validity on the system. login then allows access to the valid user and starts a
shell.
Figure 4-1 shows a sample entry in I etc/passwd for a user named terry whose
4 user ID is 265, group ID is 20, home directory is /users/terry , and login shell
is /bin/ksh. You might wish to refer to this figure when reading the following
section. For details on the / etc/passwd file, see "The / etc / passwd File" later
in this chapter.

/ u sers/terry : / b i n / k s h
L D)f 1 p
terry : Z M P PAvH rX TDfM : 2 6 5 : 2 0 : Terry K e l log :

l u se r me I l u se r I g u ID I home d i rectory l og i n shell

ID stri n

Figure 4- 1 . Sample /etc/passwd Entry for User terry

4-2 Login
The login Process
The login process, described next, is illustrated in Figure 4-2.
1 . login searches the I etc/passwd file for the username.
a. If the username exists, login goes to step 2.
b . If the username does not exist , login:
i. Prompts the user for a password (Password : ) .
ii. Displays the message Login incorrect .
iii. Updates the I etc/btmp file ( if it exists ) , which tracks invalid login
attempts . 4
iv. If the user attempts three consecutive invalid logins , login exits;
otherwise, login prompts the user (login : ) for a username and
repeats step 1 .
2. login checks t o see if the username's password field i s set in I etc/passwd.
a. If so, it prompts the user for a password (Password : ) and goes to step 3.
b . If not , the user need not enter a password; go to step 4.
3. login compares the password to the username's encrypted password in
I etc/passwd.
a. If the password matches, login goes to step 4.
b. If the password does not match, login displays the message "Login
incorrect . " If the user attempts three consecutive invalid logins , login
terminates; otherwise, login prompts the user for a username (login : )
and control passes back to step 1 .
4 . login sets the username's user ID , group ID , and home directory from the
corresponding fields in / etc/pas swd ( see Figure 4- 1 ) . login also updates
the I etc/wtmp file, which tracks valid logins.
5 . login executes the command specified in the command field of / etc/pas swd
( see Figure 4- 1 ) . Typically, the path name of the user's preferred shell. If
the command field is empty, login starts a Bourne shell by default.

Login 4·3
 prompt for
-......:.::..:..:-.,.. ond i n p u t
password

ye s

4
es

get u ser ID,

g ro u p I D , ond
home d i rectory
from /etc/p a s s w d

exec command

exec /bin/sh

END

Figure 4-2. The login Process Control Flow

4-4 Login
The Shell Environment Initialization Sequence
A shell initializes its environment before becoming available to a user:
1. The shell runs the appropriate system login script, which initializes the
user's environment :

If the Shell Is . . . The System Login Script

Is . . .
Korn (/bin/ksh) /et c/prof ile
Bourne (/bin/sh) /etc/prof ile 4

Restricted (/bin/rsh , /bin/krsh) /et c/prof ile

C (/bin/csh) /et c/csh . login
Key (/usr/bin/keysh) /et c/prof ile
PAM (/bin/pam) none

As shipped, these scripts define and export for shell use the environment
variables PATH, TZ , and TERM. Since these scripts are run for all users at
login, the system administrator can modify these files to set global defaults
for all users.
2. The shell runs the user's local login script if it exists in the user's home
(login) directory:

If the Shell Is . . . The Local Login Script Is . . .

C (/bin/csh) . login
Korn (/bin/ksh) . prof ile
Bourne (/bin/sh) . prof ile
Restricted (/bin/rsh, /bin/krsh) . prof ile
Key (/usr/bin/keysh) . prof ile
PAM (/bin/pam) . environ

Typically, the system administrator initially creates a local login script for
each user. If used to add a user, SAM automatically creates a default local
login script . Users can customize their environment by modifying these files

Login 4-5
 -

to suit their needs. The Korn and C shells may have additional local login
scripts :
a. Korn shell-If the ENV environment variable is defined, the Korn shell
runs the file defined by ENV ( typically, . kshrc) whenever a new Korn
shell is started. Many programs ( for example, vi and mailx) allow users
to start a shell from within the program; this is called a shell escape.
The ENV file is re-run for a shell escape, whereas . prof ile is run only at
login.

4
b. C shell-Runs the . cshrc file whenever a new C shell is started. This is
similar to how the Korn shell ENV file works.
c. Key shell-Runs the . keyshrc file whenever a new Key shell is started.
If . keyshrc does not exist , /usr/keysh/C/keyshrc is used.

3. Once all initialization is complete, the shell displays a prompt and waits for
input from the user.

4-6 Login
Correcting Typing Mistakes during Login
During login, HP-UX does not know what kind of terminal the user is logging
in on. Therefore, HP-UX has a primitive set of keys for editing. These key
assignments are:
Does a backspace over the character just typed.
Kills the line (backspace to start of line).
Do not use these characters in usernames or passwords since they cannot be
��. 4

Once the user logs in, these key assignments are usually changed in the
system or local login script via the stty command (see stty(l) in the HP- UX
Reference), as follows:
stty erase -H kill -u

This sets the erase (erase previous character) key to (Backspace ) and sets the kill
(kill to start of line) key to @Ofil.

Login 4-7
 The /etc/passwd File
The I etc/passwd file contains essential information required during login.
It contains entries (one per line) for all valid users of the system. Typically,
the system administrator would modify or change these fields when adding
new users to the system. As an easier alternative, the system administrator
could use the SAM program to add users, which adds/changes the fields
automatically (see the System Administration Tasks manual for details) .

4 Syntax of Entries
The general form of I etc/passwd entries is:

username : password[ ,pw_age ] : userid: groupid: idstring : homedir: command

A colon. Fields are separated by colons .
username The username for which a user can login. login assigns this value
to the environment variable LOGNAME. The logname command
displays this value (see logname(l) in the HP- UX Reference).
password The encrypted form of the password the user must supply during
login for this username .
If this field is empty ( : : ) , the user has no password and can login
without typing one. This form of password is not recommended
since it could lead to security problems.
Users can change their password without superuser privilege
by using the passwd command (see passwd(l) in the HP- UX
Reference).
Password aging, an additional security feature, can b e added too,
as described next under ,pw_ age.

Note On trusted systems , the password field contains only a * , and

passwords are stored in a separate file, I . secure/ etc/passwd.
For details on trusted system, see HP- UX System Security.

, pw_age An optional aging field. Can be used to ensure that a user is forced
to change his or her password every n-week interval, where n is

4-8 Login
defined by the system administrator. This field is comprised of
these parts:

, max min wks

There should be no spaces between the max, min, and wks
parameters. They are shown separated here only to improve
readability.
The comma sets off the age fields from the encrypted
password which precedes it . 4
max A single character representing the maximum number of
weeks the user can use the password before being forced to
use a new password. This character is encrypted as shown in
Table 4- 1 :

Table 4- 1 .
Password Aging Encryption Characters

Character Weeks
0
I 1
0 through 9 2 through 1 1
A through Z 12 through 37
a through z 38 through 63

mm A single character representing the minimum number of

weeks the user must use the current password before being
allowed to change it. Like the max character, the min
character is encrypted, as shown in Table 4- 1 .
wks A string of encrypted characters defining the number of
weeks ( counted from the beginning of 1970) when the
password was last changed. The login program uses and
updates this field to enforce password aging. A null value is
equivalent to specifying zero weeks.
If max and min are zero ( e.g., ,pw_age is • . . ) , the user is forced
to enter a password at the next login attempt and this field

Login 4·9
 is removed. If min is greater than max, then only the system
administrator can change the password.
See the password example for terry in the next section, "Sample
/ etc / passwd Entries" .
use rid A numeric user ID . Must be unique for each user. The id
command displays this value ( see id(l) in the HP- UX Reference).
groupid Group I D of the group t o which the user belongs. Must b e a valid
group ID defined in /etc/group . The id command displays this
4 value ( see id(l) in the HP- UX Reference).
idstring A character string, typically holds the user's full name.
homedir Directory to use as the home direct ory that is, the directory in
-

which the user is initially placed after logi n. login assigns this
value to the HOME environment variable.
command login runs command using the exec system call ( see exec(2) in the
HP- UX Reference). Any command can be placed in this field, but
typically it contains the path name of a shell: /bin/ksh, /bin/ sh,
/bin/ csh, or /bin/pam.
login assigns this value to the SHELL environment variable.
Users can change their default shell without superuser privelege
via the chsh ( change shell ) command ( see chsh ( l ) in the HP- UX
Reference).
For more details o n the format of the /et c/passwd file, ( see passwd(4 ) i n the
HP- UX Reference).

4- 1 0 Login
Sample /etc/passwd Entries
root : xE5/0 qrnYf8Hg : 0 : 1 : System Administrator : / : /bin/sh
Defines the root user (superuser) with an encrypted password, user ID
0, group ID 1 , ID string "System Administrator" , home directory I
(root) , and who uses the Bourne shell.
michael : , . . : 125 : 10 : Michael Moos e : /users/michael : /bin/ksh
Defines a user named michael who when he next tries to login, the
system will force him to enter a new password (indicated by " , . . " in 4
the password field) . In addition, he has user ID 125, group ID 10, id
string "Michael Moose" , home directory /users/michael , and he uses
the Korn shell.
terry : , 9/ : 265 : 20 : Terry Kellog : /us ers/terry : /bin/ksh
Defines a user named terry who has password aging enabled. login
will force terry to change her password every 1 1 weeks-indicated by
the 9 as the max element of the pw_age field ( , 9 / ) . In addition, terry
cannot change her password more than once a week-indicated by the I
as the min component of pw_age ( , 9/ ) .
guest : /bin/rsh
Defines a guest user who has a restricted shell.
who : : 90 : 1 : : / : /bin/who
Defines a user named who with no password, user id 90, group id 1 ,
home directory I , and command /bin/who . If a user attempts t o login
using username who, the system simply runs the who command and
displays its output to the screen.
date : : 9 1 : 1 : : / : /bin/date
This is similar to the who entry. When the user types date to the
login : prompt, the system runs the date command.

Login 4- 1 1
 Login Tracking Files
(Jetc/btmp and /etc/Wtmp)
Two files keep a log of logins : /etc/btmp keeps track of bad ( failed ) logins and
/ etc/wtmp keeps track of successful logins . In HP-UX diskless clusters, these
files are context-dependent files ( CDFs ) .

/etcfbtmp
4 If/etc/btmp exists, the login process updates it automatically whenever a
bad login attempt occurs. To create this file, issue the command:
touch / etc/btmp
Note, you must be root to write to the /etc directory.
Read the I etc/btmp file using the lastb command, to determine whether
unauthorized users are attempting to login ( see last(l) in the HP- UX
Reference).

Note For security purposes, be sure I etc/btmp is not readable by

users.

/etc/wtmp
The login process automatically updates I etc/wtmp whenever a user
successfully logs in. Read this file using the last command ( see last(l) in the
HP- UX Reference).

4· 1 2 Login
Environment Variables
Environment variables define various characteristics of the environment in
which your shell runs. Environment variables consist of a name and a value
(NAME=value ) . Table 4-2 lists the environment variables used by the Korn,
Bourne, and C shells. The assigned values of environment variables are used by
your shell and passed to each process created during a session. For example, if
you type the cd command without any options, it returns you to the directory
contained in the HOME environment variable.
To see a list of all environment variables and their values defined in your shell 4

environment , use the env command (see env(l) in the HP- UX Reference).

Note VUE reads different customization files to set environment

variables than the standard HP-UX login scripts. Refer to the
HP Visual User Environment {VUE} System Administration
Manual for information on setting environment variables in
VUE.

Login 4- 1 3
 Table 4-2. Shell Environment variables

Variable Description Default Value

LOGNAME Contains the user's username . Set by login. from I etc/pas swd
HOME Defines the user's home directory. Set by from I etc/pas swd
login.
PATH Lists the directories the system searches to find : /bin : /usr/bin
executable commands. Set by login.
4 SHELL Contains the default shell. Set by login from /bin/sh
command field of I et c/passwd. If there is no
password field, the default is used.
TERM Specifies the kind of terminal the user is logged hp
in on. Set by the local login script .
TZ Provides the current time zone and difference MST7MDT
from Greenwich Mean Time . Set to Mountain
Standard Time by your shell's default login
script ( the system administrator should change
the value if you are in another time zone ) . TZ
must be set identically in five different places:
/et c/re , /etc/csh . login, /etc/prof ile ,
/et c/powerfail , and ( Series 300/400/700
only ) /etc/mkrs .
MAIL Determines where the system looks for mail. /usr/mail/$LOGNAME
Set by login, based on the username (for
example, /usr/mail/terry ) .

For details on setting and using environment variables , see Using HP- UX. The
remainder of this section discusses two important environment variables: PATH
and TERM.

4· 1 4 Login
The Command Search Path (PATH)
When you type a command, HP-UX searches all the directories specified by
the PATH variable until it finds the command. If the command is not in a
directory specified in PATH , the system displays:
command_name : Command not found .
If you or users of the system use commands in a particular directory frequently,
you may want to change PATH to include this directory.
4
PATH variable Format

The PATH variable contains a list of directories to search, separated by colons.

There should be no spaces surrounding the colons. For example, suppose you
use the echo command to determine the value of PATH, as follows:
$ echo $PATH
/bin : /usr/bin
This means that when you type a command, the shell first searches for the
command in the /b in directory, then in the /usr/bin directory.

Commonly Used PATH Directories

Table 4-3 shows the path names of the most frequently used directories. You
might want to add some ( or all ) of these directories to PATH.

Table 4·3. Possible Directories to Include in PATH

Directory What It Contains

/bin Frequently used HP-UX commands.
/et c Commands the system administrator uses.
/usr/bin Additional HP-UX commands .
/usr/contrib/bin Contributed programs not supported by Hewlett-Packard.
/usr/local/bin Programs and commands written locally ( at your location ) .
/usr/bin/X1 1 X l l programs .
$HOME/bin A directory you might create for your shell scripts and programs.
Commands and programs stored in your current directory. If . lS
in your PATH , it should be the last entry.

Login 4· 1 5
 Remember that directories in PATH are searched in the order in which they
appear ( left to right ) . In general, put the most frequently used directories first
in the path-unless two commands in the search path have the same name (for
example, /bin/rm and $HOME/bin/rm ) . In this example, if you want the shell
to find your version of rm first , put $HOME/bin/rm before /bin/rm in PATH.

Specification of Terminal Characteristics (TERM)

To communicate effectively with your terminal, HP-UX must know the type of
4
terminal or graphics display you're using. The TERM environment variable, tset
command, and stty command serve this purpose.
The default local login script prompts you to enter your terminal type as
follows:
TERM = (hp)
Pressing (Return ), sets the TERM environment variable to hp, the default value
( this value works with Hewlett-Packard terminals, but it may not let you
take full advantage of your terminal or graphics display features ) . Entering a
different value, sets the TERM environment variable to that value.

Selecting a value for the TER M variable

HP-UX supports many terminal types . The /usr/lib/terminfo database

tells HP-UX how to communicate with each terminal type. When you assign
a value to TERM, the value must equal a value in the t erm.info database ( see
terminfo(4) in the HP- UX Reference).
To illustrate, the files listed under usr/lib/terminfo/2 show all acceptable
TERM values that begin with 2 ( this is only a partial listing ) :

$ ls /usr/lib/terminfo/2
2382 2397a 262 1a 2623p 2626-x40 2640a
2392 2500 262 1k45 2624 2626A 2640b
2392A 262 1 262 1nl 2624a 2626P 2644
2392a 262 1 -48 262 1nt 2624p 2626a 2645
2393 262 1-ba 262 1p 2625 2626p 2647
2393A 262 1 -fl 262 1'Wl 2626 2627 2647F

4· 1 6 Login
Table 4-4 outlines the most common terminal and graphics display settings for
Hewlett-Packard equipment . When more than one choice is listed, all choices
are equivalent.

Note If you are using the HP Visual User Environment ( VUE) , HP

recommends that you keep TERM set to hp . VUE has its own
display requirements and thus looks in its own customization
files for environment variables . Refer to the HP Visual User
Environment System Administration Manual for information.
4

Table 4·4. Settings for the TER M Environment Variable

H You Are Using a . . . Set TERM to . . .

terminal the terminal's model number ( for
example, 2622 , hp2622 , 262x, 2392 )

Vectra 2392

medium resolution graphics display 3001 or hp3001

( 512x600 pixels)
high resolution graphics display 300h or hp300h
( 1024x768 pixels )
HP 98550 display station 98660 , hp98650 , 98650a, or hp98550a
( 1280x1024 pixels)
HP 98720 or HP 98721 display station The graphics interface card number
( 1280x1024 pixels) ( for example, 98720 , hp98720 ,
9873 1 , hp98736a)

Most users have a high-resolution terminal. The first setting you might try is
hp300h. If that doesn't work, try the number of your graphics interface card
( such as 987xx) . You can look in /usr/l ib/terminfo/9 for all the interface
cards beginning with number 9, or /usr/lib/terminfo/h for all interface cards
beginning with hp to find your exact number.

Setting TER M with the tset Command

Login 4- 1 7
 The tset command is a flexible command that sets the value of TERM and
initializes your terminal characteristics. If you always log in using the same
terminal type, you could change your local login script to eliminate the TERM
prompt. In the local script , this command displays the TERM prompt:
eval ' tset -s -Q m ' : ?hp ' '
-

To customize the above command, replace ?hp with the value of TERM. For
example, the following command initializes your terminal as a high-resolution
graphics display (300h) , but the TERM prompt itself does not display:
4 eval ' tset -s -Q -m ' : 300h ' '
If you use more than one type of terminal ( such as one at work and one at
home ) , you could modify your tset command to include multiple terminal
types. For an example, see "Sample C Shell .login Script" .

Using stty to Modify and Display Terminal Characteristics

The stty command can be used to change various terminal characteristics,

such as baud rate, erase and kill characters, etc. For example, the following
stty command resets the erase and kill characters to # and © , respectively ( the
defaults during login ) :
$ stty eras e # kill ©
To change them back to the ( Backspace ) (@fil)E)) and � keys,
respectively, use:
$ stty erase - H kill -u

It also displays terminal settings if invoked with the -a option; for example:
$ stty -a
speed 9600 baud ; 1ine = O ; susp= -@ ; dsusp
= -@
intr = -c ; quit
= -\ ; erase = -H ; ki11
= -u ; eof
= -n ; svtch
= -@ ; eo1
= -@
-parenb -parodd cs8 -cstopb hupc1 cread c1oca1 -1ob1k -crts
- ignbrk brkint ignpar -parmrk -inpck - istrip -in1cr - igncr icrn1 -iuc1c
ixon ixany ixoff ienqak
is ig icanon -xcase echo echoe echok -echon1 -nof1sh
opost -o1cuc on1cr -ocrn1 -onocr -on1ret -ofi11 -ofde1 -tost op

For details on the use and output of stty, see stty(l) in the HP- UX Reference .

4· 1 8 Login
Specifying the Worki ng Environment
with System Login Scripts
System login scripts contain commands to set a user's environment ; for
example, commands may check mail, set a user's terminal settings, and set
environment variables. System login scripts run before local login scripts. This
section describes the system login scripts /etc/prof ile and /etc/csh . login.
The system login scripts define a default environment . Local login scripts
can override or modify these defaults for each user. Thus , the local login
4
script provides a way for individual users to further modify their working
environment .

Note The /etc/prof ile and /etc/csh . login scripts shown here
may differ somewhat from those on your system. The main
point to be gained from looking at these examples is to
understand, generally, what these scripts do.

Login 4- 1 9
 /etc/profile
/ etc/prof ile is the default system login script for the Bourne, Korn, and Key
shells. Its features are described after this example:
# @ (#) $Revision : 66 . 6 $
# Defaul.t system-wide profile file ( /bin/sh initializat ion) .
# This shoul.d be kept to the bare minimum every user needs .
trap " " 1 2 3 # ignore HUP , IBT , QUIT now .

PATB=/bin : /usr/bin : /usr/contrib/bin : /usr/local/bin I defaul.t path .

JUJIPATB=/usr/man : /usr/contrib/man : /usr/local/man # defaul.t path .
4 if [ -r /etc/src . sh ]
then
/etc/src . sh I set the t imezone
unset SYSTEK_IAl!E
else
TZ=MST7MDT # change this for local t ime .
export TZ
fi
i£ [ " $TERll " = ] # if t erm is not set ,
then
TER!l=hp # defaul.t the t erminal type
fi
export PATH !IAIPATB TERll
stty 1 - # Set erase t o -1
trap "echo logout " 0
# This is to meet legal requirements . . .
cat /etc/copyright
if [ -r /etc/motd ]
then
cat /etc/motd # message of the day .
fi
if [ -f /bin/mail ] # not ify if mail .
then
if mail -e
then echo "You have mail . "
fi
fi
if [ -f /usr/bin/news ]
then news -n # not i£y if new news .
fi
if [ -r /tmp/changetape ] # might wish to delete this :
then echo "\007\nYou are the first t o log in since backup : "
echo "Please change the backup t ape . \n"
rm -f /tmp/changetape
fi
trap 1 2 3 # leave default s in user environment .

4-20 Login
 Task Explanation
Ignore signals The opening trap instruction prevents the login script from being
interrupted, by setting the following signals to be ignored: SIGHUP
( hangup , 1 ) , SIGINT ( interrupt, 2) , and SIGQUIT ( quit, 3) . Signals
are documented in sign al(5) of the HP-UX Reference .
Set PATH Your shell searches the default PATH through /bin, /usr/bin,
/usr/contrib/bin, and /usr/local/bin for the commands you
issue. Your current directory is not included. Likewise, MANPATH
identifies the locations of on-line specifications of the HP-UX
Reference Manual. 4
Set time The TZ variable is set to Mountain Standard Time; change the
zone ( TZ ) setting in /et c/prof ile for other time zones.
Set TERM If the TERM variable is not set , this script sets it to hp.
Export global The export command passes the values of shell variables to
variables subshells. This script exports PATH , MANPATH , TZ , and TERM.
Set stty Sets terminal port to recognize A H as the erase character .
Trap This trap captures the exit signal (0) generated when you log out .
logout When the shell receives this signal, the command "echo logout" is
executed.
Display The cat command displays the contents of /etc/copyright , a file
copyright file of copyright messages required by law.
Display The cat command displays the contents of /etc/motd. This file
motd file contains the message of the day.
Check If the /bin/mail file exists , mail -e checks for mail. If you have
for mail mail, "You have mail" is displayed.
List If the /usr /bin/news file exists, news -n is executed to list the
news items news files in /usr /news .
Check for If the file /tmp/ changetape exists and is readable, the message ,
backup tape "Please change the backup tape" is displayed, instructing you to
remove the tape from the tape drive.
Reset The final trap resets signals 1 , 2, and 3, after which, you can use
signals hangup , interrupt, and quit .

Login 4-2 1
 /etc/csh.login
I etc/ csh . login is maintained by the system administrator as the default
system login script for the C shell. This section presents a sample script , and
then explains its contents.
# @ (#) $Revision : 66 . 4 $
# Default ( example of) system-wide profile file (/bin/csh initializat ion) .
# This should be kept to the bare minim11111 every user needs .

# default path for all users .

4 set path= (/bin /usr/bin /usr/contrib/bin /usr/local/bin)
set prompt=" [\ ! ] % "
# default llAlllPATH
setenv !IAllPATH /usr/man : /usr/contrib/man : /usr/local/man

if ( -r /et c/src . csh ) then

source /et c/src . csh # set the TZ variable
else
set env TZ MSTltMDT # change this for local t ime .
endi£

if ( ! $?TERM ) then # if TERM is not set ,

setenv TERM hp # use the default
endi£

# This is to meet legal requirements . . .

cat /etc/copyright # copyright message .
# Miscellaneous shell-only act ions :
i£ ( -£ /etc/motd ) then
cat /etc/motd # message 0£ the day .
endif

if ( -f /bin/mail ) then
mail -e # not i£y i£ mail .
if ( $status == 0 ) echo "You have mail . "
endi£

i£ ( -£ /usr/bin/nevs ) then
nevs -n # not i£y i£ nev nevs .
endif

i£ ( -r /tmp/changetape ) then # might vish to delet e this :

echo
echo "You are the first to log in s ince backup : "
echo "Please change the backup t ape . \n"

rm -£ /tmp/changetape
endi£

4-22 Login
 Task Explanation
Set path The s et built-in command of the C shell sets path as a local shell
variable. The C shell automatically sets the environment variable
PATH to the shell variable path.
The default path tells your shell to search /bin, /usr/bin,
/usr/contrib/bin, and /usr/local/bin for source when you issue
a command. The current directory is not incJuded.
Set prompt The shell variable prompt is a local shell variable in the C shell.
This script sets prompt to " [n] Y. " , where n is the command event
number. Command event numbers are used with the C shell's 4
command history feature.
s et env MAJJPATH The s et env built-in command of the C shell sets the environment
variable MAJJPATH to search the paths listed for on-line specifications
of the HP-UX Reference Manual.
Set time The TZ variable is set by default to Mountain Standard Time, but
zone ( TZ ) the contents of I etc/src . csh override it .
Set TERM If the TERM variable is not set , this script sets it to hp.
Display The cat command displays the contents of /etc/copyright , a file
copyright file of copyright messages required by law.
Display The cat command displays the contents of I etc/motd. This file
motd file contains the message of the day.
Check for If thefile /bin/mail exists, mail -e checks for mail. If you have
mail mail, "You have mail" is displayed.
List news If the /usr/bin/news file exists, news -n displays the list of news
items files in /usr/news .
Check for If the file /tmp/changetape exists and is readable, the message,
backup tape "Please change the backup tape" is displayed, instructing you to
remove the tape from the tape drive.

Login 4-23
 Default Local Login Scripts
When adding a new user, the system administrator should make sure that the
new user has a local login script . HP-UX provides three default local login
scripts:
• /etc/d . prof ile for Korn and Bourne shell users.
• /etc/d . login and / etc/d . cshrc for C shell users.
When adding a new user manually, the administrator copies the appropriate
4 file into the user's directory. For example, if the new user's shell is ksh and
home directory is /users /newuser, the administrator copies as follows:
$ cp /etc/d . prof ile /users/newuser/ . prof ile
Ifthe administrator uses the SAM program to add a new user, SAM
automatically copies the appropriate default local login script into the user's
home directory.

/etc/d.profile
The default local login script resets the value of PATH, and sets terminal
characteristics, the shell environment , and shell variables. This section shows a
sample d . profile and describes its contents.
# @ (# ) $Revision : 66 . 1 $
# De£ault user . pro£ile £ile ( /bin/sh init ializat ion) .
# Set up the t erminal :
eval ' tset -s -Q -m • : ?hp ' '
stty hupcl ixon ixo££
# Set up the search paths :
PATH=$PATH : .
# Set up the shell environment :
set -u
trap "echo ' logout • " O
# Set up the shell variables :
EDITOR=vi
export EDITOR

4-24 Login
 Task Explanation
Set terminal The t s et command generates the TERM = (hp) prompt during
characteristics login. Based on the value you enter, the TERM environment
variable is set and the terminal's characteristics are defined. If you
press (Return), TERM is set to hp. Refer to tset ( l ) for more
information.
The stty command sets options for modem connections and
scrolling text on your screen. Refer to stty ( l ) for more information.
Set PATH Changes the value of PATH by adding the current directory to the
end of PATH. 4

Set shell The s et -u command prevents the shell from attempting to execute
environment command lines with undefined shell variables. This command is a
safety feature because, by default , the shell substitutes a null value
for variables that aren't defined. For example, if a script uses a
variable named WORK and you have not assigned a value to WORK , the
script will not execute.
The trap command captures the exit signal (0) that is generated
when you log out . When the shell receives this signal, the command
"echo logout" is executed. ( This trap command duplicates the
one set in /et c/prof ile.)
Set shell Shell variables such as EDITOR are set and exported as global
variables variables . EDITOR is a Korn shell variable that determines which
in-line editor the shell uses.

Login 4-25
 /etc/d.login
The default C shell local login script is similar to the Bourne shell / Korn shell
script . The default local login script resets the value of PATH, sets terminal
characteristics, and sets shell variables.
# @ (#) $Revision : 64 . 2 $

# Defaul.t user . login file ( /bin/csh init ialization )

# Set up the defaul.t search paths :

4 set path= (/bin /usr/bin /usr/contrib/bin /usr/local/bin . )

# Set up the t erminal type :

eval ' tset -s -Q -m • : ?hp ' '
stty erase 1 kill
11- 11u intr
11- c eof
11 11- 11 11-
n1 1
stty susp z hupcl ixon ixoff tost op
11- 11

t abs

# Set up shell enviroD111ent :

set noclobber
set history=20

4-26 Login
 Task Explanation
Set path The value of path, if set here, explicitly changes the default search
path. Note that for security purposes, the current directory should
be placed at the end of the search path.
Set terminal Because of the question mark in the string ' : ?hp ', tset queries you
characteristics for verification of TERM = (hp) during login. If you press (Return),
TERM is set to hp. If you enter a different value, the TERM
environment variable is set accordingly. See tset ( l ) for information
on the tset command and terminfo (4) on the terminal capabilities
database. 4
The stty command sets options for modem connections and
scrolling text on your screen. Refer to stty ( l ) for more information.
Set shell The local shell variables noclobber and history are set .
variables
The shell variable noclobber restricts output redirection , to prevent
accidental destruction of file contents.
The shell variable history creates your command history buffer and
sets its size . By default , history is set to record the last 20
commands executed. If this variable is not set , you have no
command history and cannot edit and reuse previously executed
commands.

The /etc/d . login ( used by C shell ) differs slightly in syntax and function
from /etc/d . prof ile ( used by Bourne and Korn shells ) . For information you
can use to customize users' shells, consult csh(l), ksh(l), and sh(l) in the
HP- UX Reference Manual

Login 4-27
 /etc/d.cshrc
The /etc/d . cshrc default script sets a restricted version of path for your
subshells, sets local shell variables, and defines command aliases .
# Defau1t user . cshrc file (/bin/csh init ializat ion) .

# Usage : Copy this file t o a user ' s home directory and edit it t o
# customize it t o t aste . It is run by csh each t ime it starts up .

# Set up defau1t command search path :

#
4 # (For security , this defau1t is a minimal set . )

set path= ( /bin /usr/bin

# Set up C shell environment :

if ( $?prompt ) then # shell is interact ive .

set history=20 # previous commands to remember .
set savehist=20 # number to save across sessions .
set syst em= 'hostname ' # name of this syst em .
set prompt = "$system \ ! : " # command prompt .

# Sample alias :

alias h history
end if

4-28 Login
 Task Explanation
Set PATH A restricted path is defined that includes only /bin and /usr/bin.
Since . cshrc executes before . login during the login process, this
restricted value of path is overridden by the path defined in
$HOME/ . login. When subshells are spawned, the restricted value of
path is used.
Set shell If the shell is interactive, various shell variables are set .
variables
The shell variable history creates your command history buffer and
sets its size. In this script, history is set to record the last 20
commands executed. If this variable is not set , your subshells have 4
no command history and you cannot edit and reuse previously
executed commands.
The shell variable savehist tells the shell to record the last 20
commands in the $HOME/ . history file. This file is used as the
history from which the next session starts.
A shell variable system is set to the value output by the hostname
command.
The shell variable prompt is set to the value in the system variable,
followed by the command event number and a colon.
Set command If the shell is interactive, command aliases are set . A command alias
aliases is an abbreviation for a long command line . This script sets an alias
of h for the history command.
The /etc/d . cshrc script also contains sample aliases that are
commented out by default . Because of space limitations , these
commands are not shown in the example. To use these aliases ,
remove the pound sign (#) at the beginning of the line, thus
changing the line from a comment to a command.

Login 4-29
 Sample Bourne Shell .profile Script
The Bourne shell is the simplest shell in which to customize your environment .
Only one login script is used, the $HOME/ . prof ile script. This section first
presents a sample prof ile script, and then explains the script commands.
.

In . prof ile, exported variables are passed to all subshells . Variables that are
not exported are used only by the login shell. The following script sets terminal
characteristics, global environment variables, and local shell variables used only
by the login shell. The script also sets up a logout script and displays some
4 system status information.
# Set terminal characterist ics :
eval ' tset -s -Q -m ' : hp ' '
tabs -T$TERM

# Set PATH :
PATH=$HOllE/mybin : /bin : /usr/bin : /usr/contrib/bin : /usr/local/bin :

# Export global variables :

export PATH

# Set local shell variables :

PS1= ' $LOGIAllE $ '
CDPATH= . : . . : $HOllE

# Set up logout script :

trap "echo logout ; $HOllE/ . logout " 0

# Display syst em status in£ormat ion :

dat e
echo Currently ' vho I vc -1 ' users logged in .

4-30 Login
 Task Explanation
Set terminal This tset command sets and exports the TERM environment variable
characteristics for a default HP terminal. The t s et command also initializes the
terminal's characteristics.
The tabs command with the -T sets the tabs to the default format
for your terminal.
Set PATH This example sets PATH explicitly to search $HOME/mybin, /bin,
/usr/bin , /usr/contrib/bin, /usr/local/bin, and the current
directory. A colon ( : ) separates each directory. The trailing colon
specifies the current directory. 4

Export global The export command allows shell variables to be passed to all
variables commands that you execute. In this script, PATH is exported.
Set local shell The Bourne shell has several pre-defined shell variables. This script
variables sets two: PS1 and CDPATH . (These variables could also be exported
to become global variables.)
PS1 is the primary shell prompt. Its default value is "$ " . In this
script, PS1 is changed to your username, followed by a dollar sign
and a space.
CDPATH sets the search path for the cd command. A colon ( : )
separates each directory. As set here, the cd command searches for
the named directory in the current directory ( ) , the parent
.

directory ( ) , and the home directory ( $HOME ) .

. .

Set up logout The Bourne shell does not check for a logout script automatically.
.

script The trap command captures the exit signal (0) and executes the
specified commands . In this example, the trap echoes "logout" and
executes a logout shell script when you log out .
Display The dat e command displays the current date and time . The ' who I
system status we - 1 ' pipe counts the number of users logged in and displays the
information results in the echoed string.

Login 4-3 1
 Sample C Shell .login Script
If the C shell is your login shell, HP-UX will look for a . cshrc script and a
. login script in your home directory. The . login script is read after the
. cshrc script when you log in. The . login script sets global environment
variables. This section first presents a sample C shell login script , and then
.

explains the script commands.

Like the Korn shell, the C shell distinguishes between environment variables
and shell variables. Environment variables are used by your login shell and are
4 passed to subshells. Environment variables are defined in the . login script ,
and shell variables are defined i n the . cshrc script .
The following script sets terminal characteristics, the global environment
variable PATH, and the local shell variable prompt . The script also displays a
login message.
% cat . login
# Set up the terminal type :
eval ' t set -s -Q -m • console : 300h ' -m ' =1 200 : pc ' '

# Set PATH :
set path= ( /bin /usr/bin /usr/local/lib usr/lib . $HOKE/mybin)

# Set local shell variables :

set prompt=" [\ ! ] "

# Provide login message :

echo ' grep $LOGIAKE /etc/passvd I cut -d : -£5 ' -- have a nice day .

4-32 Login
 Task Explanation
Set terminal The t s et command sets up your terminal when you first log in to
characteristics an HP-UX system. With the t s et command, the login script can
select between two or more terminal types. This example chooses
between a high-resolution graphics display station (300h) that
functions as the console or a personal computer (pc) logging in via a
1200 baud modem.
To modify this command for your own use, replace console and
=1200 with the descriptions you need. For example, you might set
one option to =9600 ( equals 9600 baud ) and the other to <9600 (less 4
than 9600 baud ) . Then replace the 300h and pc with the terminal
types you use. ( Refer to tset ( l ) for more information ) .
Set PATH The s et command sets local environment variables. PATH is a global
variable, and path is a local variable in the C shell. The C shell
automatically sets PATH equal to path.
This example explicitly sets the entire value of path. The dot ( )
.

represents the current directory.

Set local shell The s et prompt command redefines the default prompt (% ) to
variables show the event number being typed. This feature is useful with the
command history mechanism. As set here the prompt will be " [n]
" where n is the event number.
Provide login The echo commands write messages onto the terminal screen. This
messages message echos the user's name and the message "have a nice day."
The user's name is extracted from the I et c/pas swd file . The grep
command finds the entry that contains $LOGNAME. The cut
command extracts the fifth field which should be the user's name.

Login 4-33
 Key Shell -keysh
The Key shell (keysh) is a menu-based, context-sensitive softkey interface to
the Korn Shell (ksh) . With Key shell, you can perform most HP-UX user tasks
with softkeys, rather than command-line entries. Key shell also provides online
help for building softkey commands. Full information on using Key shell is
found in the following locations:
• keysh(l) in the HP- UX Reference Manual.
• Shells: User's Guide
4

4-34 Login
 5
Process Management

Understanding how HP-UX manages processes will help you better interpret
how your system carries out its computations. This chapter discusses:
• What a process is.
• How processes are created.
• How processes are killed.
• Commands for managing processes.
5
• How the kernel manages processes.
• HP-UX multiprocessing.

Process Management 5-1

 What I s a Process?
A process is a running program, managed by such system components as the
scheduler and the memory management subsystem. Although processes appear
to the user to run simultaneously, in fact a single processor is executing only
one process at any given moment .
Described most simply, a process consists of text ( the code that the process
runs ) , data ( used by the code ) , and stack ( a "place" in the kernel where the
parts of a program are stored as the process is running ) . Two stacks are
associated with a process-kernel stack and user stack. The process uses the
user stack when in user space and the kernel stack when in kernel space.
In addition, a process includes:
• The program's data structures ( variables, arrays, records ) .
5 • A process ID, parent process ID , and process group ID .
• The process's user and group IDs ( both real and effective IDs ) .
• A group access list.
• Information on the process's open files .
• The process 's current working directory.
• An audit ID ( on trusted systems only ) .

Process Relationships
Processes maintain hierarchical, parent-child relationships. Every process has
one parent , but a parent process can have many child processes . Processes can
create processes, which in turn, can create more processes . The relationships
among processes can be represented by a tree graph. For example, Figure 5-1
shows the relationships among processes P i , P2, P3, P4, PS, and P6 .

5-2 Process Management

 Figure 5- 1 . Process Tree Graph

In this graph, you can see the following relationships:

• P i created P2 and P4. Therefore, P i is the parent of P2 and P4; P2 and P4
are the children of P 1 .
5
• P 2 created P 3 . Therefore, P 2 i s the parent of P3 ; P 3 i s the child of P2.
• P4 created PS and P6. Therefore, P4 is the parent of PS and P6; PS and P6
are the children of P4.
• Pi is the grandparent of P3, PS, and P6; P3, PS, and P6 are the grandchildren
of P i .
• P 3 , P S , and P 6 have not yet created any child processes .
A child process inherits its parent's environment (including environment
variables, current working directory, open files) . Also, all processes except
system processes (such as ini t, pagedaemon, and swapper ) belong to process
groups. Process groups are explained later in this chapter.

Process and Parent Process IDs

When a process is created, HP-UX assigns the process a unique integer number
known as a process ID (PID ) . The HP-UX kernel identifies each process by its
process ID when executing commands and system calls.
A process also has a parent process ID (PPID ) , which is the PID of the parent
process.
Using the ps command, you can see the process and parent process IDs
of processes currently running on your system (see ps(l) in the HP- UX

Process Management 5-3

 Reference). For example, if your username were t erry, you might see
something like this when executing ps:
$ ps -f
UID PID PPID C STIME TTY TIME COMMAND
terry 3865 3699 2 13 : 35 : 43 ttyp3 0 : 00 ps -f -
t erry 3699 3698 0 12 : 58 : 2 1 ttyp3 0 : 00 ksh
This indicates that the user t erry has two processes running, the ps -f
command and ksh ( the Korn shell ) . Notice that the PPID of the ps -f
command is the same as the PID of ksh. This is because ps -f was spawned
from the ksh command line; thus, ksh is the parent of ps -f , and ps -f is the
child of ksh. ( For details on the other columns shown, see ps(l) in the HP- UX
Reference.

s User and Group IDs (Real and Effective)

In addition to the process ID , a process has other identification numbers:
• a real user ID
• a real group ID
• effective user ID
• effective group ID .
A real user ID is an integer value that identifies the owner of the process­
that is, the user ID of the user who invoked the process. A real group ID is
an integer value that identifies the group to which the user belongs. The real
group ID is shared by all users who belong to the group. It allows members of
the same group to share files, and to disallow access to users who do not belong
to the group .
The id(l) command displays both integers and names associated with real user
ID and real group ID . The / etc/passwd file assigns the real user ID and real
group ID to the user at login. ( For details, see Chapter 4, "Login" . )
% id
uid=513 (terry) gid=20 (users)
% grep 513 /etc/passvd
terry : EqqHevH : 513 : 20 : Terry Ho , [44KY] , 495-0601 , : /users/terry : /bin/csh

The effective user ID and effective group ID allow a process running a program
to act as the program's owner while the program executes. The effective IDs

5·4 Process Management

are usually identical to the user's real IDs. However, the effective user ID and
group ID can be set individually to protect a program, by making the effective
IDs of the program's processes equal to the real IDs of the program's owner.
The effective IDs are used to allow a user to access or modify a data file or to
execute a program in a limited manner. When the effective user ID is zero,
the user is allowed to execute system calls as the superuser ( described in the
following section ) .
For example, suppose the dean of a university keeps all students' records in
a file on the system. He wishes to allow an English professor to modify a
student 's record, but only for that professor's class. The dean first sets read
and write permissions on the file contruning the student 's records. He then
writes a program that takes as input the professor's modifications, verifies that
the professor, as user, may change the record, then modifies the record if it is
allowed. Finally, the dean protects the program by assigning the effe ctive IDs 5
of the user to the dean's real IDs when the program is executed. Thus, when
the program accesses the student record file, the system allows the program
to read from or write to the file because it believes that the dean is accessing
the file ( the effective user and group IDs are those of the dean ) . However,
the real user and group IDs of the process rem.Un the same as those of the
user invoking the program. The program can use these IDs to verify access
permission to the data ( that is, ensuring that an Engli sh professor is accessing
English class records ) .
The effective user and group IDs rem.Un set until:
• The process terminates.
• The effective IDs are reset by an overlaying process, if the setuid or setgid
bit is set . ( See exec(2) in the HP- UX Reference.
• The effective, real, and saved IDs are reset by the system calls setuid,
setgid, s etresuid, or s etresgid. ( See section 2 in the HP- UX Reference).

Audit ID (Trusted Systems Only)

If you have converted to a trusted system, each user has an audit ID . This
audit ID does not change, even when the user executes programs that use
a different effective user ID. Refer to HP- UX System Security for more
information.

Process Management 5·5

 Jobs
A job is any command or command pipeline invoked from an HP-UX shell.
When you execute a command, the operating system interprets the entire task
required by that command as a job .

Job Control
HP-UX supports job control for both the Korn shell and C shell. Job control
provides users with greater flexibility in managing and controlling jobs. For
example, you can:
• Temporarily stop (suspend) a foreground job , by pressing (CTRL}(D. This
can be customized using the stty command. see the HP- UX Reference.
• Bring a background job into the foreground, using the fg built-in shell
5 command.
• Move a foreground job into the background, using the bg built-in shell
command.
For more information, see stty(l), csh(l), glossary(9), and ksh(l) in the
HP- UX Reference. Job control is also discussed in Shells: User Guide and A
Beginner's Guide to HP- UX .

Process Groups
Every process (except system processes , such as init and swapper) , belongs
to a process group. When you create a job , the shell assigns all the processes
in the job to the same process group . Signals can propagate to all processes
in a process group; this is a principal advantage of job control. (Signals are
discussed later in this chapter.)
Each process group is uniquely identified by an integer calle d a process group
ID . Each process group also has a process group leader. The process group's
ID is the same as the process ID of the process group leader. Every process in
a process group has the same group ID .
A process group ID cannot be re-used by the system until its process group
lifetime ends. The process group lifetime is a period of time beginning when

5-6 Process Management

a process group is created and ending when the last remaining process in the
group leaves. A process leaves either:
• When another process executes a wait ( ) or waitpid ( ) function for an
inactive process, or
• When calling the setsid or s etpgid system calls (see setsid(2) and
setpgid(2) in the HP- UX Reference).

Group Access Lists

Each process has a group access list of up to NGROUPS_MAX groups to which the
process belongs. NGROUPS_MAX is defined in /usr/ include/limits . h, and is
typically 20. Programmatically, you can use sysconf ( ) to obtain the value.
A process is permitted to access the files of any group in this list as though
that group were the process's effective group ID. The access list is assigned at
5
login based on the group memberships specified in the file /etc/group (Series
300/400/700) or / etc/logingroup (Series 800).
You can control group access by using the chgrp command. (See chown(l) in
HP- UX Reference Manual.)

Sessions
Each process group is a member of a session. All processes started during a
single login session belong to a session. Or, think of a session as a login shell
with all the jobs the login shell spawns. On a system running windows or shell
layers, each window or layer is a session.
A process belongs to the same session as its creator (parent) . A process can
alter its session affili ation using the s et s id system call. The shell uses the
s etpgid system call to create jobs and ensure that all processes in a job
belong to the same process group . ( See setsid(2) and setpgid(2) in the HP- UX
Reference.)
A session leader i s the process that created the session via the setsid system
call (normally the login shell). Session lifetime is the period between when a
session is created and the end of the lifetime of all process groups that belong
to the session.

Process Management 5· 7
 A special instance of process groups is the orphaned process group. This is
a process group in which the parent process of every member is either itself
a member of the group or is not a member of the group 's session. Terminal
signals (such as ' Z) cannot stop an orphaned process group .

Processes and Terminal Affiliation

Every session has one controlling terminal. The session leader connected to the
controlling terminal is called the controlling process. The exceptions to this are
daemon processes (such as cron and inetd); they have no controlling terminal.
All processes belonging to the session use the controlling terminal for standard
input , standard output , and standard error.
At any one time, one process group in a session is the foreground process
group. Typically, the foreground process group is the job that the login
5 shell runs in the foreground, or the shell itself if no foreground job has been
spawned. The foreground process group has primary access to the controlling
terminal, beyond that of all background process groups. The foreground
process group can read from and write to the controlling terminal without
restrictions.

Attempted read by Background Process Group

If a process in a background process group attempts to read from its

controlling terminal, its process group is signaled with SIGTTIN, which
suspends the process to which it is sent , except in two instances:
• When the reading process ignores or blocks SIGTTIN
• When the reading process belongs to an orphaned process group .
In either case, the read returns - 1 and no signal is sent.

Attempted Write by Background Process Group

If a process in a background process group attempts to write to the controlling

terminal, its process group is signaled with SIGTTOU which, by default, stops
(suspends) the process to which it is sent . There are three cases where SIGTTOU
is not sent :
• If the TOSTOP terminal characteristic is not set , the process is allowed to
write.

5-8 Process Management

• If the TOSTOP terminal characteristic is set, and if the writing process ignores
or blocks SIGTTOU, the p rocess is allowed to write.
• If the TOSTOP characteristic is set, and the writing process belongs to an
orphaned process group, and the writing process does not ignore or block
SIGTTOU, the call to write returns - 1 .

Process Creation
A process can create another process to:
• concurrently execute another program.
• execute another program and wait for its completion.
At a shell prompt, you create a new process by typing the command's name 5
and pressing (RETURN ). The shell is the parent process; the process you create is
the child process. The child process can then create its own child processes, of
which it is the parent .
At the system programming level, a new process is created when a program
calls either the fork or vfork system call. The parent process is the calling
process; the child process is the created process. The system subroutine can
also spawn a child process; system's calling process waits for the child process
to terminate before continuing.
The following section discusses briefly key system calls initiated when you
run a program. For detailed information on the intricacies of processes and
programs, consult the manual Programming on HP- UX , and system(3S),
fork(2), vfork(2), and exec(2) i n the HP- UX Reference .

The fork System Call

The fork system call creates ( "forks" ) a new process .
In earlier implementations of HP-UX, the system copied the entire data
segment of a process every time fork was issued. fork time increased as the
size of the data and stack segments increased. Because processes can be larger
than available memory, copy time could increase dramatically, degrading
system performance.

Process Management 5-9

 HP-UX now implements two architecture-specific means by which the system
can create processes more quickly-"copy-on-write" (Series 300/400) and
"copy-on-access" (Series 700/800). These are described in Chapter 7, "Memory
Management" of this manual.
When the parent process is written to the child address space, it differs from
the parent process in its process ID and parent process ID); it has exact copies
of the parent 's code, data and current variable values. (See fork(2) in the
HP- UX Reference for a detailed list of similarities and differences.)
When the fork system call is executed, the system must have enough free
swap space to duplicate the parent process or fork fails. Once the child
process is created, both processes begin to execute from the program statement
immediately following the call to fork.
The fork system call returns the child's process ID (a non-zero value) to the
5 parent process, while the identical call in the child's copy of the code always
returns zero. Since the process IDs returned by fork are distinguishable, each
process can determine whether it is the parent process or the child process .
For example, suppose a process consists of a program that tests the life of car
batteries. The program has read 1000 data values from a voltmeter and is
ready to print and plot the data. The program could have been written to do
one task completely (such as printing the data) and then perform the second
task (plotting the data) . However, the programmer has included a fork system
call in his program at a location after the data has been read.
When the program completes the statement containing the fork system call,
two nearly identical processes exist . Each process examines the value returned
by its fork system call to determine whether it is the child process or the
parent process. Following the fork statement is a conditional branch statement
that states, "If the process is the child process, print the data. If the process
is the parent process , plot the data." The fork statement and the conditional
branch statement enable both printing and plotting to be done simultaneously.
And since each process has its own copy of the test data, each can modify the
data without affecting the other process.

The vfork System Call

The vfork system call is functionally identical to fork, and provided only for
backward compatibility.

5· 1 0 Process Management
The exec System Call
Often following the fork system call is an exec to another program. exec is
a system call that overlays separate code and data on top of already existing
process code and data. In this manner a parent process is able to create a new
process using fork, and subsequently execute an entirely different program via
exec.
For example, suppose you are editing a file and need to verify the name
of another file. You might use a shell escape to fork a shell and exec the
program ls.

Open Files
For a process to access files, it must first open them. A process inherits all
open files from the parent . Three files that are usually open are standard input 5
( stdin ) , standard output ( stdout ) , and standard error ( stderr ) . When a
process terminates , the system closes any files opened by the process.
Two configurable operating-system parameters govern the number of open
files permitted per process. The parameter maxf iles defines the soft limit ,
representing the maximum number of open files a process can have without
issuing the setrlimit ( 2 ) system call. By default , maxf iles is set to 60.
The parameter maxf iles _lim defines the hard limit of open files permitted per
process. By default, maxf iles_lim is set to 1024. Only a superuser process
can increase this parameter beyond its hard limit . maxf iles _lim is used
with setrlimit ( 2 ) to change number of open files permitted. maxf iles_lim
can be increased up to a maximum of 2048 or decreased to a minimum of 60.
( Programmatically, you can use sysconf ( ) to get the current value. )
A process's soft limit can be increased up to the hard limit or decreased to
any value greater than number of the highest file-descriptor allocated. An
absolute limit of 2048 is imposed on both hard and soft limits. If the value of
maxf iles_lim is 60, no one can increase their open-file (maxf iles ) limit .
When considering the appropriate setting of soft and hard limits, consider too
the operating-system parameters , nf ile and ninode , because they are affected
as well. ( See Appendix A, "System Parameters" in the System Administration
Tasks manual for descriptions of nf ile and ninode . )

Process Management 5-1 1

 Process Termination
At the system programming level, a process terminates ( dies ) when:
• The process successfully finishes running.
• The process intentionally terminates itself by calling the exit system call
( see exit(2) in the HP- UX Reference) .
• The process receives a signal whose default action i s fatal.
When a process dies, all its open files are closed and most of the resources
the process holds are deallocated. The process then enters an inactive state
in which it still holds some system resources. The remaining resources are
returned to the system, the last being process's entry in the proc table. The
process dies.
At the shell level, processes can be terminated via the kill command,
5 described in more detail in the next section.

5- 1 2 Process Management
Process Management Commands
From both command level (from a shell) or from SAM (Process Management
area) , you can manage and monitor processes. The most useful tools are
described in this section. HP-UX can also provide system accounting
information for terminated processes (see Chapter 14, "System Accounting" ) .

Understanding Process Status-ps

The ps command, run from a shell or from SAM, displays the following
information about processes currently running on the system:
• User ID of the user who spawned the process
• Process ID
• Parent process ID
• Command line used to spawn the process 5
• tty (terminal) line from which command was invoked
• Length of time the process has been running (real CPU time) .
ps also allows you to query the system about processes selectively. It is a very
versatile command which is quite useful to system administrators and general
users alike.
When invoked without options, ps displays the process ID , terminal ID (tty),
real CPU time usage, and name of all commands a user is running. If the -f
(l ull) option is specified, ps also displays login name, parent process ID , and
time the process was forked.
$ ps -:f
UID PID PPID C STillE TTY TillE COlllU.ID
mickey 3286 2016 9 16 : 19 : 03 ttyp1 0 : 00 ps -:f
mickey 25705 25649 0 08 : 47 : 58 ttyp1 0 : 02 -ksh /users/michael [ksh]
mickey 2016 25705 0 1 5 : 13 : 02 ttyp1 0 : 24 vi processes . t ag

Including the - e option causes ps to display information for all processes in the
system, not just those of the user who invoked it :
$ ps -e:f
UID PID PPID C STillE TTY TillE COlllU.ID
mickey 25737 25715 14 08 : 48 : 13 ttyp3 0 : 00 xcal1 /usr/local/b:in/xcal1
root 13322 1 0 Jun 6 ? 0 : 00 cron
mickey 24357 1 0 19 : 45 : 46 console 0 : 01 -ksh [ksh]
r\)ot 4 0 0 Jun 6 ? 7 : 15 net isr

Process Management 5- 1 3
 If the - 1 ( long ) option is specified, ps displays user ID , state (s), nice value
( NI ) , memory or disk address of the process ( ADDR ) , priority ( PRI ) , and size
( SZ ) in blocks of the process core image. ( State and priority are discussed in
some detail later in this chapter.)
$ p s -1
F S UID PID PPID C PRI !II I ADDR sz VCHAlll TTY TIJIE CDKD
1 R 513 11009 7793 5 179 20 d6e200 16 ttyu4 0 : 00 ps
1 s 0 7792 133 15 154 20 e06100 13 214fb0 ttyu4 0 : 00 rlogind
1 s 513 7793 7792 16 168 20 d:f5a80 52 7ffe6000 ttyu4 0 : 00 csh

The -1 option is useful for isolating problems on a sluggish system. For

example, a runaway process might show up with an excessively large entry in
the TIME column. You can trace the process back to its owner via the UID
column. Similarly, if a process is not responding or getting time to run, the S
(status) column might indicate whether the process is sleeping.
5 For details on using ps, see ps(l) in the HP- UX Reference.

Understanding Process Termination-kill

The kill command, run from a shell or from SAM, terminates a running
process when you specify the process 's PID . By default , kill terminates the
process gracefully via the SIGTERM termination signal. "Graceful" means
that the process first handles any additional signals pending. By specifying a
signal number, you can kill processes more abruptly if necessary (for example,
kill -9, where 9 specifies the SIGKILL signal) . See signal(5) in the HP- UX
Reference for complete description of the signals handled by the multiple signal
interfaces supported by HP-UX.
For example, suppose a user is running a program that has hung the terminal;
that is, the program does not respond to keyboard input from the user and the
user cannot exit the program. As system administrator, you can kill the user's
program from another terminal by doing the following:
1. Determine the process ID of the program, by running ps .
2. Kill the process via kill.
If the program was named list_ data and the user's user ID was 265, you
might see the following:
$ ps -f u 265-

UID PID PPID C STIME TTY TIME COMMAND

5-14 Process Management

 terry 3677 3638 0 12 : 27 : 29 ttyp 1 0 : 53 list_data
terry 3638 3632 0 12 : 25 : 02 ttyp 1 0 : 00 ksh
To kill the process gracefully, type:
$ kill 3677
This would, in most cases, kill the process and the user would again be able to
use the terminal. If this does not work, you might try:
$ kill -9 3677

Note This form of kill should be used cautiously because it bears a

higher risk of data loss.

Understanding Relative Process Priority-nice and renice 5

All processes have a priority, set when the process is invoked and based on
factors such as who is running the process ( user, system ) and whether the
process is created in a time-share or real-time environment .
The nice command can be used to set a process to run at a lower priority
than would be set by default . nice does not lower the priority of an already
running process . nice is useful for running programs whose execution time is
not critical.
For example, suppose you have a program, named numcrunch, that
manipulates large arrays of data, but the data is not critical to your work
at the moment. How long it takes for the program to manipulate the data
is unimportant; more critical programs should have greater access to CPU
resources. To run numcrunch as a low priority background process , type:
$ nice numcrunch &
Note that both Korn and C shells handle nice slightly differently: ksh
automatically lowers priority of background processes by four; this behavior
can be modified using the bgnice argument . If you specify nice from ksh, it
executes /usr/bin/nice and lowers priority by ten. If you specify nice from
csh, it executes its built-in command and lowers priority by four; however, if
you specify /usr/bin/nice , csh lowers priority by ten.

Process Management 5· 1 5
 The renice command allows you to alter the priority of running processes.
Running processes can also be altered from the Process Management area of
SAM.
For details, see nice(l), ksh(l), csh(l), renice(l), and nice(2) in the HP- UX
Reference.

Tools for Monitoring Process Management Performance

The following tools may provide additional information to help you examine
your system's operation. Although they are commands, some ( top , monitor,
and sar ) are accessible from the Process Management area of SAM:
top(l) Displays and updates information about the top processes
on the system. Summarizes the general state of the system
(load average) , quantifies amount of memory in use and free,
5
and reports on individual processes active on the system.
Whereas ps gives a single "snapshot" of the system, top
samples the system and updates its display at intervals. On
multi-processing systems, top reports on the state of each each
CPU.
monitor(lM) Displays a wide variety of system performance statistics,
in each major system area ( processes, memory, 1/0 , and
networking) .
sar(l) (Series 800 only.) Reports on cumulative system activity,
including CPU utilization, buffer activity, transfer of data to
and from devices, terminal activity, number of specific system
calls used, amount of swapping and switching activity, queue
lengths, and other kernel tables.
vmstat( l) Quantifies the use of virtual memory by processes on the
system; also reports on traps and CPU activity.
iostat(l) Reports 1/0 statistics for active disks, terminal, and processor.

5- 1 6 Process Management
Process Management and the Kernel
You can think of the process control system as containing the kernel's
scheduling subsystem, memory management subsystem, and interprocess
communication ( !P C ) subsystm.
The process control system interacts with the file system when reading files
into memory before executing them. Several processes may be instances of the
same program; for example, more than one person might be using vi at same
time; each invocation of vi is an instance of the same process .
A process reads and writes its own data and stack, but cannot write / read any
other processes'. ( Note, however, shared memory can be read and written by
several processes. )
Processes communicate with other processes via shared memory or system
calls. Communication between processes ( !P C ) includes asynchronous signaling 5
of events and synchronous transmission of messages between processes . System
calls are requests by a process for scme service from the kernel, such as I / O ,
process coordination, system status , and data exchange. All HP-UX system
call s are documented in section 2 of HP- UX Reference Manual.
Much valuable information about CPU usage can be obtained by running
monitor or top from the Process Management area of SAM.

Process Modes
An HP-UX process can execute in two modes-kernel or user mode-and
through its process lifetime, switches between them. Information about the
process ( such as variables , process addresses, buffer counts ) accumulates in a
"stack" for each mode, and it is through these stacks that the process executes
instructions and switches modes.
For Series 300/400, certain kinds of instructions trigger mode changes; for
example, when a program invokes a system call, an entry point in the system
call library is encoded in assembly language and contains "trap" instructions,
which, when executed, cause an interrupt that results in a switch to kernel
mode. When the process executes the instruction, it switches mode to the
kernel, executes kernel code, and uses the kernel stack.

Process Management 5-17

 For Series 700/800, when you make a system call, you go through system call
stub code, which passes the system call number through a gateway page that
adjusts privilege bits to put the process in kernel mode.
Every process has an entry in a kernel process table and a u_area structure,
which contains private data such as control and status information . The
context of a process is defined by all the unique elements identifying it-the
contents of its user and kernel stacks , values of its registers , data structures ,
and variables .
Although HP-UX gives the illusion of executing multiple processes
simultaneously, each CPU actually executes one process at a time, according
to its priority. ( In multiprocessing systems, several SPUs can be executing a
process simultaneously.)
The principle behind the distribution of CPU time is called time-slice-the
5 amount of time a process can run before the kernel checks to see if there is an
equal-priority process ready to run.

Process Priorities
Processes are chosen by the scheduler to execute a time-slice based on their
priority. Priorities range from 0 ( highest priority) to 256 ( lowest priority) , and
are classified by need. You can see at what priority a process is set to run at
by looking in the PRI column when you invoke ps or top . The following lists
the three categories of priority:
Real-time priority (0- 127) Reserved for processes started with rtprio ( )
system calls.
System priority ( 128- 1 77) Used by system processes.
User priority ( 1 78-256) Assigned to user processes.
The kernel can alter the priority of time-share priorities (128-256) but not
real-time priorities .

5- 1 8 Process Management
Run Queues
A process must be on a queue of runnable processes before the scheduler can
choose it to run. Figure 5-2 shows an abstraction of run queue organization.
Run queues are link-listed in decreasing priority. Each process is represented
by its header on the list of run queue headers; each entry in the list of run
queue headers points to the process table entry for its respective process.
Processes get linked into the run queue based on the process's priority, set in
the process table.

Run Queue Headers

0
Proc table

5
Real-lime

1 27
1 28

System
x
-Z. c=J -+ [J[J - - - - - + x

Time
Share 1 77
Priorities 1 78

User

25 6

LG20021 1 _025

Figure 5-2. Run Queue, showing Priorities

Process M anagement 5- 1 9
 The kernel maintains separate queues for system-mode and user-mode
execution. When a timeshare process is not running, the kernel improves
the process's priority (lowers its number). When a process is running, its
priority worsens. The kernel does not alter priorities on real-time processes.
Timeshared processes (both system and user) lose priority as they execute and
regain priority when they do not execute.
The scheduler chooses the process with the highest priority to run for a given
time-slice. system-mode priorities take precedence for CPU time. User-mode
priorities can be preempted-stopped and swapped out to secondary storage;
kernel-mode priorities cannot . Processes run until they have to wait for a
resource (such as data from a disk, for example) , until the kernel preempts
them when their run time exceeds a time-slice limit , until an interrupt occurs,
or until they exit . The scheduler then chooses a new eligible highest-priority
process to run; eventuaJ.ly, the original process will run again when it has the
5 highest priority of any runnable process.

5-20 Process Management

Process Context
When the kernel switches from executing one process to executing another
process, it switches context . That is, the kernel switches from executing in the
context of one process to the context of another process. When this happens,
the kernel saves information about the interrupted process to be able to later
resume execution. The kernel also restores the context of the process it is
about to start.

system mode
P rocess 1
user mod e
5

I I
I l
P rocess 2
system mod e I
user mod e
l

Time ----.

LG20021 1 _026

Figure 5-3. System Switching Context from One Process to Another

Context switching is shown in Figure 5-3. In this figure, the system switches
from executing in the context of process Pi to executing in the context of
process P2. The dotted vertical lines indicate the points of context switching.
Note that context switching can only take place when the process is executing
in system mode.

Process Management 5-2 1

 Process States and Transitions
Throughout the course of its lifetime, a process transits through several states.
Queues in mafa memory keep track of the process by its process ID . A process
resides on a queue according to its state; process states are defined in the
sys/proc . h include file. Events, such as receipt of a signal, cause the process
to transit from one state to another.
Processes can be described as being in one of the following states:
idle Process has either just been forked; an idle process can be
scheduled to run.
run Process is on a run queue, available to execute in either kernel
or user mode.
stopped Executing process is stopped by a signal or parent process.
sleep Process is not executing, while on a sleep queue (for example,
5
awaiting 1/0 to complete ) .
zombie Having exited, the process no longer exists, but leaves behind
for the parent process some record of its execution.
These states and their transitions are shown in Figure 5-4.

5-22 Process Management

 fork ( )

acquiring
system resources

shell
sends
process
5
gets resource
SIGCONT

process
gets
gets resource
(loses)
time
slice swapper
swaps
out
process

waits for

mode
switch process
ends

LG20021 1 _0 1 8

Figure 5-4. Process States and Transitions

Process Management 5-23

 When a program starts up a process , the kernel allocates a structure for it
from the process table; the process is now in idle state, waiting for system
resources. Once it acquires the resource, the process is linked onto a run queue
and made runnable. When the process acquires a time-slice, it runs , switching
as necessary between kernel mode and user mode. If a running process receives
a SIGSTOP signal ( as with control-Z in vi) or is being traced, it enters a stop
state. On receiving a SIGCONT signal, the process returns to a run queue
(in-core, runnable ) . If a running process must wait for a resource ( such as a
semaphore or completion of 1/0), the process goes on a sleep queue ( sleep
state ) until getting the resource, at which time the process wakes up and is put
on a run queue ( in-core, runnable ) . A sleeping process might also be swapped
out , in which case, when it receives its resource ( or wakeup signal ) the process
might be made runnable, but remain swapped out . The process is swapped in
and is put on a run queue. Once a process ends, it exits into a zombie state.
5
Interrupts

Here is an example of how interrupts work:

Process A might make a system call requiring disk 1/0 . While waiting for the
disk 1/0 to complete, Process A goes to sleep . Meanwhile Process B starts up.
The disk 1/0 completion generates an interrupt. In the course of completing
the disk 1/0, the driver controlling it also puts the sleeping Process A onto
a run queue, making it runnable at a high priority, because Process A now
holds a resource. Interrupt activity "belongs" to Process A, but happened
while Process B was running. The interrupt is asynchronous to the thread of
execution.

5·24 Process Management

Signals
Signals are transmissions between processes and are used for interprocess
communication (IPC). Signals are also viewed as events to which processes
respond.
An example of signal use is the kill command, which sends a SIGTERM signal
to terminate the specifies processes.
When a process receives a signal, the process might ignore the signal,
terminate, defer the signal, or execute a signal handler. HP-UX defines several
signal interfaces that allow a process to specify the action taken upon receipt of
a signal. ( See sigaction(2) signal(2), sigvector(2), bsdproc(2), and sigset(2V)
in the HP- UX Reference Manual for the various HP-UX signal interfaces. See
signal(5) for description of HP-UX signals. Acceptable signal values are defined
in /usr I include/ sys/ signal . h.)
5

Process Management 5·25

 Multiprocessing
Standard HP-UX executes in a uniprocessing system architecture consisting of
one central processing unit ( CPU ) , memory, and peripherals . The architecture
of Series 870 and 890 systems supports multiprocessing ( MP ) -two, three, or
four system processing units ( SPU s ) , memory shared among the processors ,
and peripherals . Multiprocessing is a limited form of parallel processing.
Header files that define MP include mp . h, sp inlock . h, and several
semaphore-related header files. MP header files can be read but they define
operating-system behavior that is not configurable.

How Multiprocessing Compares to Uniprocessing

Despite architectural differences, multiprocessing systems execute code
5 identically to standard HP-UX. Although multiple processes can run
concurrently on the several processors, all processors use the same kernel code.

Symmetry

Symmetry in MP has to do with use of the SPU. With some asymmetrical

MP systems , only a primary SPU can execute 1/0 and system code; other
SPUs execute only user code. With symmetrical MP, every SPU can execute
1/0 and system code. One copy of the kernel resides in memory and controls
all processors. Since any kernel task can execute on any processor in the MP
system, HP-UX MP is termed symmetrical.

Multiprocessor Boot-Up

Boot-up procedure in an MP system is almost identical to that of a

uniprocessing system. At boot-up , the processor-dependent code ( PDC )
arbitrates among the multiple processors to determine the "monarch
processor." The serf ( non-monarch, or remaining ) processors wait in
"rendezvous code," while the monarch processor alone executes the bootup
code. Upon reaching the MP portion of the configuration code ( near the end
of booting process ) , the monarch processor signals each serf processor in turn,
to bring it into the multiprocessing environment , at which time they go into
idle ( see Figure 5-5, ready to run processes. Once all serfs are brought in, the
monarch processor can go into idle also. From idle, processors do their work:
they look in run queues for processes that need CPU time, grab locks, execute

5-26 Process Management

processes, release locks , and thus perform their normal operations on behalf of
the kernel.

I I
real main main i d 1eo
MONARCH SS

Time -----+ 5

LG20021 1 _0 1 7

Figure 5·5. Multiprocessors Booting Up

Scheduling Processes using Semaphores and Spinlocks

When processes execute simultaneously, it becomes harder for the operating
system to maintain the integrity of data. To solve this problem, HP-UX
controls access to data by using semaphores.
Semaphores ( and spinlocks, a kind of semaphore ) are a locking mechanism for
ensuring that only one processor accesses critical kernel resources at a time.
Critical kernel resources include run queues, file and inode tables, linked lists,
any kind of table that processes might want to access or change.
A spinlock is held very briefly; other semaphores lock resources for longer
durations and allow the processor to perform another task ( such as switch
processes ) while the lock is held.
The MP system kernel uses semaphores to protect global data structures
while multiple processors are executing concurrently. Locks are also used in a
uniprocessing system to ensure data integrity.
Figure 5-6 shows in a very general way how semaphores and other locks allow
processes to synchronize execution.

Process Management 5·27

 � Locks
P1 I
i- I _l,,_!_
-
1 /0---+- ___
___
_

1
P2 ---�--1-
,
I
0-·------------ --

Time __,.

LG20021 1 _0 1 5

Figure 5-6. Processes Lock Kernel Resources to Synchronize Execution

Processes run concurrently, but each lock may only be held by one process at a
5
time. If another process attempts to acquire an already-held lock, the process
must wait for the lock to be released.
The example shown in Figure 5-7 demonstrates how the kernel allocates critical
regions to concurrently running processes. A process ( P l ) , executing in user
mode on SPU 1 , might take a trap , which interrupts tl;i.e process's execution
and sends it into kernel mode, where it starts executing system code. P l puts
itself on a run queue, which pulls another process ( P2 ) off the run queue. P2
starts to run. Meanwhile, a third process ( P3 ) , executing in user mode on SPU
2, takes a trap into kernel mode and wants to go on the run queue. Since P2
is executing in the critical region and controls the spinlock, P3 must wait ,
spinlocked, until P2 releases the lock; then P3 can continue to run.

5·28 Process Management

 SPU 1

I
Run Q

Lock

System Space
I
User Space

SPU 2
5

System Space

User Space

Time ----.
LG200211_018

Figure 5-7.
Kernel Allocates Resources to Three Concurrent Processes on Two Processors
Using Semaphores and Spinlocks

When multiple processes want to access critical resources at the same time,
there is potential for a race condition or failure. Race conditions might occur
any time a program takes an action dependent on data. Spinlocks enable the
kernel to arbitrate between competing processes so that each process gets to
access resources in a prioritized order.

Process Management 5-29

 For example, suppose there are two processes, executing on two SPUs, each
process accessing the same location in kernel memory simultaneously (in
lockstep) . Each process is counting objects, with the statement
nobj ects=nobj ects+ 1 ;
The value of nobj ects is the number of objects counted, and equals 10.
If the same code is running simultaneously without spinlocks, the two
processors independently compute and change the total twice to 1 1 . With
spinlocks, the first process must complete and release its lock before the second
process can compute. The total is thus incremented twice, first to 1 1 , then 12.
Each processor uses spinlocks to control access to data structures on behalf of
its executing process. Likewise, competing processors must wait to obtain a
spinlock held by another processor, until the lock is available.
5
Processor Affinity
Processor affinity is the tendency of a process to execute from beginning to end
on the same processor. The kernel is inclined to allocate processing time so
that this happens, because a process running on the same processor executes
faster.
The kernel, however, must also balance the workloads among the SPU s, and
will move a process from one SPU to another if necessary to achieve the best
overall system throughput .

Uniprocessor Emulation Issues

Underlying the design of HP-UX multiprocessing is the concept of uniprocessor
emulation. In all user interactions, HP-UX on MP behaves exactly as if
the operating system were running on a single processor. As noted earlier,
widespread use of spinlocks require that device drivers, in I/O operations
and in virtually all areas of program design, share kernel data structures
sequentially to ensure their integrity. Because of MP 's greater complexity,
programmers must take care to write code more carefully: timing hazards not
seen in uniprocessing can occur in an MP environment . Careful regression
testing on MP computers is essential to screen for timing hazards.

5-30 Process Management

A problem resulting from this sequential sharing of resources is characterized
by excessive idle time. This situation can be detected with tools, such as
top(lM), sar(lM), and the optional HP products LaserRX and Glance/UX.
These tools, however, will not reveal why the problem occurs; thorough review
of your application design will be necessary.

Process Management 5·3 1

 6
Run-Levels

This chapter discusses run-levels; specifically:

• The ini t process and the I etc/ ini ttab file
• Run-level 2 ( multi-user run-level )
• Run-level s ( administration run-level )
• User-defined run-levels.

What Is a Run-Level?
A run-level is a system state in which a specific set of processes are allowed to 6
run. This set of processes is defined in the / etc/ initt ab file for each run-level.
You can define up to six run-levels (1-6).
Run levels s and 2 are predefined: s is the administration run-level and 2 is
the normal operating run-level ( users must provide user name to gain access ) .
Most systems do not need to define additional run levels or use any run levels
other than s and 2.

Run-Levels 6-1
 Predefined Run-Levels
There are three pre-defined run-levels as shipped: run-level 2, run-level O, and
run-level s.

Run-Level 2
The run-level in which your system automatically boots is called the initdefault
run-level. As shipped, run-level 2 is the initdefault run-level. On multi-user
systems , it may be necessary to make certain, additions to run-level 2 ( such as
the addition of more getty entries to /etc/ inittab ). For more information,
refer to Chapter 2 and the System Administration Tasks manual.

Run-Level 0
Run-level 0 is a special run-level reserved for system installation. Do not run in
run-level 0.

Run-Level s
6 Run-level s is a special run-level reserved for system administration tasks.
Shutting down the system for system administration tasks brings you to
run-level s. Change to run-level s only by using the shutdown command ( i.e. ,
do not execute init s ) . For details on switching to run-level s, see the System
Administration Tasks manual.

The init Process

The init process ( /etc/init ) is the first process HP-UX starts during system
startup . The init process has process ID one ( 1 ) and has no parent .
The init process reads the /etc/ initt ab configuration file, which defines
what processes to start at various run-levels . When init is run at system
startup time, it typically brings the system to the initdefault run-level. An
entry in /etc/ inittab defines the initdefault run-level for your system. As
shipped, the initdefault run-level is run-level 2 ( multi-user HP-UX).

6-2 Run-Levels
/etc/inittab
The init process reads the /etc/ inittab file a line at a time. Each line
contains an entry describing an action to take. Typically, entries tell ini t to
run a process in a given run-level or run-levels . For example, the following
entry tells init to run the /etc/getty process in run-levels 0 and 2 through 6:
co : 023456 : respawn : /et c/getty console H
The syntax of inittab entries is:

id: run-levels : action : process

id A character id, from 1 to 4 characters long, which uniquely
identifies the entry.
run-levels Defines the run-level( s) in which the entry should be processed.
Multiple run-levels can be specified. If this field is empty, then all
run-levels are assumed.
action Identifies what action to for this entry. The actions we discuss here
are:
ini tdefaul t Causes the initial (default) run-level to be the value 6

of the run-levels field. If more than one run-level

is specified in run-levels , ini t uses the highest
specified run-level.
boot Run the command specified in the process field
at boot-time only. Do not wait for process to die
before reading the next entry.
bootwait Run the command specified in the process field
at boot-time only. Wait for process to die before
reading the next entry.
wait On entering the run-level that matches the
run-levels field of this entry, run process and wait
for it to die before reading the next entry.
re spawn On entering the run-level that matches the
run-levels field of this entry, run process if it is not
already running. Do not wait for process to die

Run-Levels 6·3
 before reading the next entry. When process dies,
run it again.
process This is a shell command to be run, if the entry's run-levels matches
the run-level and / or the action field indicates such action.
Comments of the form # comment can be appended to entries.

Setting the Default Run-Level

The following entry, which appears near the start of / etc/inittab , sets the
default run-level:
is : 2 : initdefault :
This line tells init that the initdefault run-level is 2.
Before bringing the system to run-level 2, init processes all entries whose
action field is boot or bootwait ; such processes called boot processes . Two
such entries are: /etc/bcheckrc and /etc/brc .
6
blboot wa i t : /et c/bcheck r c < /dev/syscon > /dev/syscon 2 > � 1
bcbootwait : /etc/brc 1>/dev/syscon 2>t 1
These entries tell init t o execute the bcheckrc and brc shell scripts,
respectively. init must wait for each process to die before moving to the next
entry. In addition, these processes are run only at boot-time.
After running the boot processes, ini t runs all processes at the initdefault
run-level. Typically, these include /et c/re and /etc/getty . For details on
these processes , see Chapter 2.

6-4 R un-Levels
User-Defined Run-Levels
Because of special requirements , some systems may need additional run-levels
(other than 2 and s). Adding new run-levels involves changing /etc/ initt ab .
For details on creating your own run-levels, see the System Administration
Tasks manual.
Anyone having write permission to the file I etc/ ini ttab can create new
run-levels or redefine existing run-levels. Make sure that the permissions on
/etc/ init and /etc/ inittab are:
-r-xr--r-- root other / etc/init
-rwxr-xr-x root root / etc/ inittab

Changing Run-Levels
When you start up your system, you will be in run-level 2. This means that at
startup your system executed the inittab commands associated with run-level
2. 6
You can move freely between run-levels as long as entering the new run-level
does not kill user or system processes that may have begun in the previous
run-level. (See the System Administration Tasks manual for details on
switching run-levels.) Two examples of possible problems are discussed below:
• If you change to a different run-level, the processes corresponding to entries
in inittab that do not explicitly include the new run-level will automatically
be terminated. For example, assume you are changing from run-level 2 to
run-level 3. If the getty entry for the console does not have the action
respawn for run-levels 3, entering run-level 3 from run-level 2 will cause
the console to die.
• A user logs off prior to a change in run-level. You then change run-levels
from run-level 2 to run-level 4. When the user attempts to log in, he cannot ,
unless his /etc/getty entry contains both run-level 2 and run-level 4.

Run-Levels 6-5
 7
Memory Management

Memory is high-speed data storage, implemented using various hardware

devices on the HP-UX system. Each device stores and retrieves data.
The data and instructions of any process ( a program in execution ) must be
available to the CPU by residing in physical memory at the time of execution.
RAM, the actual physical memory ( also called "main memory" ) , is shared by
all processes. To execute a process, the kernel executes through a per-process
virtual address space that has been mapped into physical memory.
The term "memory management" refers to the rules that govern physical and
virtual memory and allow for efficient sharing of the system's resources by user
and system processes.
Memory management allows the total size of user processes to exceed physical
memory by using an approach termed demand-paged virtual memory. Virtual
memory enables you to execute a process by bringing into main memory parts
of the process only as needed, that is, on demand, and pushing out parts of a
process that have not been recently used.
7
The system uses a combination of swapping and paging to manage virtual
memory. Swapping involves moving entire processes between main memory and
secondary memory ( usually a mass storage device, such as a disk ) , whereas
paging involves moving smaller units ( called pages ) between main memory and
secondary memory.
This chapter provides an overview of HP-UX memory management , including:
• Physical and virtual memory
• Mapping files into virtual memory
• Shared libraries
• Swapping

Memory Management 7-1

 Physical Memory
The physical memory of most interest to system administrators is the random
access memory (RAM) . RAM usually consists of memory cards that plug into
the computer's backplane. For the CPU to execute a process , the relevant
parts of a process must exist in RAM.
The more main memory in the system, the more data it can access and the
more ( or larger ) processes it can execute without having to page or swap.
This is because the system can retain more processes in main memory, thus
requiring the kernel to page less frequently. Memory-resident resources ( such as
page tables ) also take up space in main memory, reducing the space available
to applications.
During system startup , the system displays on the system console the amount
of physical memory installe d:
real mem no_of_ bytes
=

Refer to the HP Configuration Guide for your system's minimum, maximum,

and recommended RAM requirements.

Power and Data Permanence

A characteristic of memory is its volatility or nonvolatility-whether or not
a storage medium retains data when power is removed. Before the system
boots, only data or microcode stored in nonvolatile memory ( ROM, EPROM,
7
magnetic tape, disk ) is available.
Once the system is brought up , data is stored and used in RAM. Data held in
RAM is volatile; that is, it is not retained in RAM when the system is brought
down.

7 ·2 Memory Management
Transactions between RAM and Disk
At boot time, the system loads the operating system from disk. (This is
essentially what booting means.) The operating system then resides in RAM
until the system is shut down.
When the user runs programs and commands, they too are loaded from disk
into RAM. When a program terminates, it is usually flushed out of RAM (that
is, the operating system frees the memory used by the process) .
Complex swapping and paging algorithms determine when data and code for
currently running programs will get returned from RAM to disk.
User and system programs write data to disk (for example, to update the
password file or write a database record) . This data-to-be-written is either
written directly to RAM (if raw data) or buffered in cache and written to disk
in relatively big chunks. Programs also read files and database structures from
disk into RAM . Buffering algorithms try to minimize disk access by going to
disk as infrequently as possible; disk access is a bottleneck on all systems.
When you issue the sync command before shutting down a system, all modified
buffers of the buffer cache are flushed ( written) out to disk.
On the Series 800, if the system loses power for a short time and powerfa.il is
enabled, the powerfail routines will put the system in a consistent state and
bring it back up without the user having to reboot it .
There are two other characteristics of physical memory involved at system
start-up: available memory and lockable memory. 7

Available Memory
Not all physical memory is available to user processes. The kernel ( /hp-ux)
always resides in main memory ( that is, it is never swapped) , occupying
approximately 1 .6 MB on a Series 300/400 and 3 MB on a Series 700/800
system. (Note, however, these are static sizes only, and do not include kernel
tables, daemons, device drivers, processes, diagnostics , user interfaces, or other
executing code on a working system.)
The amount of main memory not reserved for the kernel is termed available
memory. Available memory is used by the system for executing user processes.

Memory Management 7-3

 Instead of allocating all its data structures at system initialization, the HP-UX
kernel dynamically allocates and releases some kernel structures as needed by
the system during normal operation. This allocation comes from the available
memory pool; thus, at any given time, part of the available memory is used by
the kernel and the remainder is available for user programs.
During system startup, the system displays on the system console the amount
of available memory:
avail mem = no_of_bytes

>- ,_ l Lockable Memory

,.... - Available Memory

>-- PHYSICAL
MEMORY

�---------------

{
7

HP-UX 1------1· �
Kernel at
Boot-up L.----- _

LG200172_01 1 a

Figure 7- 1 . Physical Memory Available t o Processes

7-4 Memory Management

Lockable Memory
As shown in Figure 7- 1, physical memory that can be "locked" ( that is, its
pages kept in memory for the lifetime of a process ) by the kernel or plock ( ) is
known as lockable memory .
User processes can lock memory using the plock or shmctl system call ( see
plock(2) or shmctl(2) in the HP- UX Reference). Locked memory cannot be
paged or swapped out . Typically, locked memory holds frequently accessed
programs or data structures , such as critical sections of application code.
Keeping them memory-resident improves system performance.
By default , lockable memory can be no more than three-fourths of available
memory during bootup . Available memory is a portion of physical memory,
minus the amount of space required for the kernel and its data structures. The
size of the kernel varies depending on the number of interface cards, users, and
values of the tunable parameters; thus, available memory varies from system to
system.
Lockable memory is extensively used in "real-time" environments, like
hospitals , where some processes require immediate response and must be
constantly available.
A Series 800 feature called pseudo-swap reservation ( discussed at the end of
this chapter ) also does not allow lockable memory to exceed three-fourths of
available memory.
If the default value of lockable memory is changed, care must be taken to allow 7
sufficient space for paging and swapping. As this space is decreased, the system
is more likely to deadlock. The system parameter to change the amount of
unlockable memory is unlockable_mem. ( See System Administration Tasks
manual for information on tunable operating-system parameters. )
During system startup, the system displays on the system console the amount
of its lockable memory ( along with available memory and physical memory ) .
You can display the values later by running /etc/dmesg.
lockable mem = no_of_ bytes
lockable_mem is the total amount of physical memory all users can lock
down at once. This value can be changed by modifying the operating-system
parameter, unlockable_mem.

Memory Management 7-5

 The system has boundary conditions requiring that unlockable_mem be
more than zero and less than the amount of memory available after boot . If
unlockable_mem is set to zero or a negative value, the kernel will compute an
appropriate default value at boot time.
A va.ilable memory minus the memory locked by processes is the memory
actually available for virtual memory demand paging.
You can determine the default value of unlockable_mem by subtracting the
amount of lockable memory from the available memory during boot-up .
available memory - lockable memory = unlockable_mem
The resulting value must be less than three quarters of free memory.
For example, the following boot information was displayed on a Model 835:
phys ical page s ize = 2 KB
avail mem = 9670656 bytes
real mem = 16772 16 bytes
lockable mem = 659456 bytes
The system's unlockable memory would be
9670656 - 659456 = 90 1 1200 bytes
For details on setting lockable memory, refer to the unlockable_mem tunable
parameter in the system parameters chapter of the System Administration
Tasks manual.
7
Secondary Storage
Ma.in memory ( RAM ) stores computer data required for program execution.
During process execution, data resides in two faster implementations of
memory found in the processor subsystem, registers and cache. Program files
are kept in secondary storage or secondary memory, typically disks accessible
either via system buses or network. Data is also stored when no longer needed
in ma.in memory, to make room for active processes.

7-6 Memory Management

 D
Processor subsystem
Registers

D
In CPU
Main
Cache Memory

Input/Output
(VO) Subsystem

1/0 Bus

�l
I
I
network
I
I I
I

I
, __ __ _ _ __ .,, , ..._ __ _ _ __ ..,,,.,
I Swap Sectlo
..._ - - - - -'!-
�wap_dlrecto.!¥
----- I

J
I

L--��ice��------file
d
���!_W� _ _ _ _ _ _ _ _

secon ary memory

LG20021 1 _001

Figure 7·2. Relationship between Main Memory and Secondary Storage

7
A transitory form of secondary data storage is termed swap, dating from
early UNIX implementations that managed physical memory resources by
moving entire processes between main memory and secondary storage. Besides
swapping, HP- UX uses paging, a more e:fliicent memory resource management
mechanism for virtual memory.
While executing a program, data and instructions can be swapped ( copied) to
and from secondary storage if the system load warrants such behavior.
Swap space is initially allo cated when the system is configured. Device swap
can take the following forms :
• an entire disk ( Series 300/400/700/800)
• a designated area on a disk ( Series 300/400/700)

Memory Management 7·7

 • a section or LVM logical volume of a disk ( Series 800 )
• a software disk-striped partition on a disk ( Series 700 ) ; see sdsadmin(lM)

You can also configure a file system so that remaining space not used for files is
used for swap; this is termed file-system swap. If more swap space is required,
it can be added dynamically to a running system, as either device swap or
file-system swap .
The swapon command is used to allocate disk space or a directory in a file
system for swap.

Note You cannot remove swap without rebooting the system.

The concepts of swap and swap-space management are further discussed in

"Swap Space Management ," later in this chapter. For procedures on allocating
swap space, see the HP- UX System Administration Tasks manual for your
system, Chapter 7, "Managing Swap Space" and Appendix B , "Swap Space
Computation" .

7·8 Memory Management

HP-UX Demand-Paged Virtual Memory
Early UNIX systems transferred only entire processes between the swap device
and main memory while executing a program. Complex algorithms governed
the priority by which systems moved processes to and from swap . To better
accommodate larger programs, HP-UX memory-handling algorithms can also
allow pages ( individual units of memory ) of a process to be read ( paged ) in and
out of memory.

Hardware Perspective on Memory Transactions

Most of the kernel's memory-management code operates independently of
hardware, but some code is specific to the Series 300/400 or 700/800 hardware.
When a program is compiled, the compiler generates addresses, called virtual
addresses, for the code. These virtual addresses must be mapped into memory
to a physical address for the compiled code to execute. User programs use
virtual addresses only. The kernel and the hardware coordinate a mapping of
these virtual and physical addresses for the CPU, called "address translation."
Figure 7-3 shows how the system uses the addresses to locate the process is in
memory:

Memory Management 7·9

 Virtual Address Space

Vi rtual Page Offset

N um ber in
Page

Address Translation in Physical Address

Translation Lookaslde Buffer In Cache

-
'--t Vi rtual Access P hysical P hysical Page Offset
i---
Pg. No. I nfo. Pg. No. __. N um ber
i---
in Page

Physical Address
in RAM

P hysical Page Offset

__. ......
N um ber in Page

LG20021 1 _002

Figure 7-3. Virtual Address Translation

When a process executes, it stores its code (text) and data in processor
7
registers for referencing. If the data or code is not present in the registers , the
processor goes out to the cache, a small high-speed memory between RAM
and the processor, to fill a cache line. If the data is not in cache, the processor
consults the translation tables in RAM that hold the mapping between virtual
and physical addresses of the data. If not in RAM, the data and code might
have to be paged into RAM from disk, in which case the disk-to-memory
transaction must be performed.
A memory management unit (MMU) manages the interface between blocks of
virtual address spaces and their physical memory locations. The MMU also
ensures that processes do not illegally access each others' address space. Access
to data in memory is thus facilitated and protected at a page-by-page level.
An important element of the memory management unit is the translation
lookaside buffer (TLB ) hardware in the processor. The TLB caches the most

7- 1 0 Memory Management
recently used virtual-to-physical address translations with their corresponding
access information. Once an address is translated, it can be used to reference
physical memory.
It is also interesting to compare the relative magnitude of these memory
components. The virtual address space can potentially be a thousand times
greater than the physical address space, while the physical address space might
be a thousand times greater than the TLB . These relationships enable the CPU
to execute programs much larger than the available physical memory. It also
lets you run many more programs at a time than you could without a virtual
memory system.

- - -
- - -
- - - -
- -
- - -
- - -

Virtual Physical
Address Address
Space Space
- - -
- - -
- - -
- - -
- - -
- - -

LG200 172_016a

Figure 7-4. Role and Magnitude of the Address-Translation Components

7

Memory Management 7-1 1

 Process-to-Page Mapping
When a process executes, various dynamic structures are used to locate the
memory segments, translate them from virtual to physical addresses, and
manipulate them to achieve the results desired. Figure 7-5 shows how these
structures point from a process to its segment page mappings .
When a process executes , the kernel maintains a proc structure, a data
structure for each active process. The proc structure points to a virtual
address space, which forms the header of a doubly linked list of several
per-process regions, each corresponding to some element of the process ( such as
code, data, bss , shared memory ) . In turn, each per-process region points to a
region, which points to page mappings of the code in physical memory.
Subsequent sections of this chapter describe the virtual address space and its
related structures .

Process-oriented

proc structure

per-process
virtual address space ---1 region
(preglon)
7 ------------------- ------------

System-oriented
region

page
mapping

LG2001 72_026

Figure 7 -5. Simplification of an Active Process Executing in Main Memory.

7-12 M emory Management

Pages
Pages are the smalle st contiguous block of physical memory that can be
allocated for storing data and processes . Pages are also the smallest unit of
memory protection. As of the current release, the page size of all HP-UX
systems is 4 KB.
A free page is a page of physical memory not mapped to a virtual address; that
is, not currently being used by any process. A used page is mapped to a virtual
address. The system keeps track of all free and used pages in physical memory.
Every physical page in the system is represented by an entry in the kernel's
pfdat array data structure. A pfdat entry keeps information for the
hardware-independent code on page conditions such as a page's validity, the
state of the page, and whether the page has been modified. Implementation of
the actual page mapping is hardware dependent .
You can lock critical pages of processes in memory to prevent their being paged
out by using the system call plock(2) as described in the HP- UX Reference ; see
"Lockable Memory" , earlier in this chapter.

Virtual Address Space

Virtual memory uses a structure for mapping processes termed
the virtual address space, described in the kernel header structure
/usr/ include/sys/vas . h. The virtual address space contains information and
pointers to the memory that the process can reference. 7
One virtual address space exists per process and serves several purposes:
• It provides the overall description of each process .
• It contains pointers to another element in the memory-management
subsystem-per-process regions, or pregions .
• It keeps track of pregions most recently involved in page faults ( explained in
"How the Kernel Executes Processes using Demand Paging" ) .
• It contains pointers to page tables that enable the CPU to access physical
addresses of data. ( Series 300/400 only ) .
On HP- UX systems , the entire virtual-memory system can handle addresses of
32 bits, allowing processes a maximum virtual address space of 4 Gigabytes .

Memory Management 7- 1 3
 The Series 300 / 400 has a linear address space without base and offset . All
memory is conceptually contiguous . Each process ( including the kernel ) shares
the entire 2 32 ( 4 GB) address range.
On the Series 700 / 800, space registers hold either 16 bits (for 48-bit
addressing ) or 32 bits ( for 64-bit addressing ) and are used to point to the
virtual space to be accessed. The specific location within that space is specified
by a 32-bit quantity called the byte offset .
The present Series 700 / 800 48-bit addressing implementation can be thought of
as having 2 1 6 32-bit spaces .

Configuring Virtual Address Space

Operating-system parameters limit the size of the code, data, stack, and
shared-memory segments for each individual process . These parameters have
predefined defaults, but you can reconfigure them in the kernel using the
procedures outlined in the System Administration Tasks manual. The following
list shows configurabl� system parameters for code, data, stack, and shared
memory:
maxts iz Limits the size of the code ( text ) segment.
maxds iz Limits the size of the data segment .
maxssiz Limits the size of the stack segment .
shmseg Limits the number of shared memory segments that can be
attached to a process.
shmmn i Limits the number of shared-memory identifiers.
7

Per-Process Regions (pregions)

The virtual address space structure points to per-process regions, or pregions.
Pregions are logical segments that point to specific segments of a process,
including code ( text , or process instructions ) , data, u_area and kernel stack,
user stack, shared-memory segments, and shared-library code and data
segments.
Pregions hold page protections and the number of pages mapped to each
segment.
Each segment also corresponds to the segments of virtual address space
illustrated in Figure 7-6, Figure 7-7, and Figure 7-8.

7-14 Memory Management

The text segment ( or code segment ) holds a process's executable object code
and may be shared by multiple processes. The maximum size of the text ( or
code ) segment can be changed by the configurable operating-system parameter
maxt s iz .
The data segment contains a process's initialized and uninitialized data
structures (bs s ) . It can grow as needed by a program's run-time logic ( using
sbrk (2), malloc(3C), or malloc(3X), as described in the HP- UX Reference).
The total allotment for initialized data, uninitialized data and dynamically
allocated memory can be changed by the configurable operating-system
parameter maxds iz .
The u_area contains information about process characteristics . The kernel
stack segment , which is in the u_ area segment , contains a process 's run-time
stack during kernel mode. Both u_ area and kernel stack segment are fixed in
size.
The user stack segment contains a process's run-time stack during user mode.
The user stack expands during process execution; its default maximum size can
be changed using the configurable operating-system parameter maxssiz.
Shared memory segments are typically used when multiple processes must
share data ( for example, in a windowing system, where all window processes
must be able to update common data structures ) . They can be created using
the shmop(2) system call ( see specifications in the HP- UX Reference).
Each shared library segment typically has three pregions-code, initialized
data, and bs s . 7

I/ 0 mappings are the ranges of addresses the operating system uses to deal
with 1/0 devices.
When a process executes, the memory management subsystem traverses the
pregions to find pointers to another data structure, called a region, which
points to the address of the desired page of code or data.

Per-Process Region Layouts

Per-process regions (pregions ) represent the virtual address space in memory.

The layouts of the Series 300/400 and 700/800 virtual address space differ in
specifics. This section shows the mappings of both Series 300/400 and Series
700/800.

Memory Management 7·1 5

 Placement of Segments in Address Space (Series 300/400)

The segments of virtual address space for a process reside in virtual memory as
shown in Figure 7-6. In the Series 300/400 memory design, the percentage and
boundaries of memory address are not critical.
As a process is executed, the stack segments grow in the direction shown by
the arrows in Figure 7-6 . Segments are arranged to allow for ample memory
allocation, to prevent overlap that would cause error (such as ENOMEM) and
kill the process.

7- 1 6 Memory Management
 ( high address) _________
oxFFFFFFFF kernel stack
-----------

�----------

-----------

1/0 mappings

shared memory

m emory-mapped files

shared libraries

-----------

=== m alloc (heap)] = l.

bss 7
maxdslz
initialized data
----------- .
-
( low address) - -- ��;{t;"xt) - - m axtsiz
oxOOOOOOOO

LG2002 1 1 _003

Figure 7-6. Serles 300/400 U ser Process Virtual Address Space

Memory Management 7- 1 7
 Placement of Segments in Address Space (Series 700/800)

In the Series 700 / 800, the virtual address space is divided into one-GB
quadrants and addressed by units of 32 bits. Each quadrant has several
segments associated with it , as shown in Figure 7-8 and explained on the next
page.
The basic segments ( quadrants ) of the Series 700 / 800 can be described as
follows:
• First one- GB quadrant always contains the process 's code and sometimes
some of the data.
Text is addressed from OxOOOOOOOO to Ox3FFFFFFF.
• Second quadrant contains data ( static data, stack, and heap ) .
Data is addressed from Ox40000000 to Ox7FFFFFFF.
• Third quadrant , which is addressed from Ox80000000 to OxBFFFFFFF, is
used somewhat differently on Series 700 and Series 800:
o On the Series 700, the third contains shared memory, shared mapped files ,
and shared library code.
o On the Series 800, the third quadrant contains memory-mapped segments
( such as shared-library code ) .
• Fourth quadrant , which is addressed from OxCOOOOOOO to OxFFFFFFFF,
contains shared-memory segments. On the Series 700, fourth quadrant also
contains shared memory-mapped files and shared library code. On PA-RIS C
architecture ( both Series 700 and 800 ) , addresses from OxFOOOOOOO t o
OxFFFFFFFF are used fo r I / O space.
7
On the Series 700, an EXEC _MAGIC user executable ( a . out ) format allows
data to start immediately after the code area in the first quadrant , instead of
at the beginning of the second quadrant , and grow to the beginning of the user
stack. Executables created with this format can handle more than 1 GB of
data; refer to ld( l ) in the HP- UX Reference Manual for information.

7-18 Memory Management

 (high address)
OxFFFFFFFF 1/0 mappings
oxFOOOOOO
�----------

OxEFF01 000 i-- - - - - - - - - - -

shared li brary code

shared memory 4th quadrant
shared memory-mapped
OxC0001 000 1-- - _ _ _ file§_ _ - - -
gateway page
OxBFFFFFFF

shared library code

shared memory 3rd quadrant
shared memory-mapped
files

Ox80000000
kernel stack

{ === J==
-----------
u area
m axssiz us; stack
private memory-mapped 2 nd quadrant
files
m alloc (heap)
bss
initialized data

7
-- - - - - - - - - - -
1 st quadrant

code (text)

OxOOOOOOOO
(low address)
LG20021 1 _004

Figure 7-7.
Series 700 User Process Virtual Address Space with EXEC_ MAGIC implemented.

Memory Management 7- 1 9
 (high address)
OxFFFFFFFF 1/0 mappings
i-- - - - - - - - - - -
OxEFFFFFFF

shared memory
4th quadrant

OxC0002000 -----------

gateway page
OxBFFFFFFF

shared li brary code

shared memory-mapped 3rd quadrant
files

{ === ]==
Ox80000000
m axssiz user stack
kernel stack
io-- - - - - - - - - - -

{
u area
- - - - ------ 1 -
2 nd quadrant
___ !!!_allo�_(he�)_ _

m a-. bss
�----------
i nitialized data
Ox40000000

1 st quadrant
io-- - - - - - - - - - -

7
code (text)

OxOOOOOOOO
(low address)
LG20021 1 _005

Figure 7 -8. Series 800 User Process Virtual Address Space

7-20 Memory M anagement

Regions
Regions are data structures unique to the file or chunk of memory being
accessed. Regions represent ranges of addresses and inform the process about
where the data exists in physical memory.
Each process segment 's per-process region maps directly to a segment of
memory in a region. Regions are associated with the system rather than a
process.
In turn, the region points to two kernel lists, which keep track of pages of data.
A region might be private or shared. A private region ( such as a process's data
segment , u_area, or stack ) has only one pregion pointing to it . A shared
region ( such as shared memory, code, graphics ) can have more than one
pregion point to it , because more than one process might be accessing the
same segment of memory.

Virtual Nodes (vnodes)

A region usually references a virtual node (vnode), an interface for reading

and writing pages of data between memory and disk. Vnodes are supersets of
inodes that the MMU stores to keep track of swapped or paged memory while
it is out on disk. Vnodes are categorized by file type-hfs , nf s , dux, cdf s , and
swap-and are defined in the header file, vnode . h. Kernel routines read the
stored file type when:
• Bringing requested and additional pages in from disk. 7
• Writing pages out to disk.
• Giving page information to the kernel for later computation.
• Duplicating page information for use by multiple processes.

Memory Management 7-2 1

 Memory-Mapped Files (Series 300/400/700)
Memory-mapped files allow applications to map file data into a process 's
virtual address space. Once mapped, the process can directly manipulate
the file data as a portion of memory. File data can be shared between
processes more efficiently, resulting in higher 1/0 throughput for processes .
Memory-mapped files are always page aligned. 1/0 to the files can be
performed through direct loads and stores of memory instead of using read ( 2 )
and write(2), avoiding copying of data between user and system buffers .
Memory-mapped files provide identical representation of in-memory and
on-disk data. The data can be selectively synchronized to flush out a portion of
a file range.
Files can be mapped into memory as either private or shared. Modifications
to a private mapping are not visible to other processes, regardless of whether
the other processes have created a private or shared mapping, or perform a
read or write of the same file. The modifications are never written back to the
file. Modifications to a shared mapping are visible to all other processes with a
shared mapping and the modifications are written back to the disk file.
Memory-mapped files are well suited to randomly dispersed 1/0 , such as in
database applications with data that is most frequently read, and which require
frequent , small, random updates. Likewise, multiple processes that read huge
data files can use memory-mapping for repeatedly searching and scanning the
same data.
7 Several system calls manipulate memory mappings to a file. Also, user-level
semaphores interface with the memory-mapped regions :
mmap Create a mapping from a range in the process 's address space
into a range in a file.
mprotect Modify the protections on memory pages mapped to a file.
munmap Unmap a mapping created by mmap .
msync Synchronize memory pages of a mapped file back to disk.
ms em_ init Initializes a semaphore in a memory-mapped region.
ms em_ lock Locks a semaphore in a memory-mapped region.
ms em_ unlock Unlocks a semaphore in a memory-mapped region.
ms em_remove Removes a semaphore in a memory-mapped region.
For detailed information on these system calls, refer to section two of the
HP- UX Reference Manual.
7-22 Memory Management
Limitations to Memory Mapping
Memory-mapped files cannot be locked into memory. Extensive use of
memory-mapped files (mmap) might result in some degradation of performance.
Access to the same file by read and write system calls, on the one hand, and
memory mapping, on the other hand, might produce inconsistent data, since
both approaches maintain separate in-memory caches of file data. Use one
access method exclusively for any given application.
The mmap interface does not map character device files and device-specific
addresses ( such as graphics framebuffers ) . Use graphics(7) for graphic devices
and iomap(7) for other devices on the Series 300/400; use framebu/(7) for
graphics devices on the Series 300/400/700/800.
The Series 700 memory-mapped file interface is defined to work in a 32-bit
process address and cannot share data with processes outside this address
space. The following limits apply to the Series 700 implementation:
• 0.9 GB is the maximum size of a single file that can be mapped entirely into
memory.
• 1.75 GB ( minus code size and data size ) is the maximum size of any
EXEC_MAGIC executable on a Series 700.
• 1 . 75 GB is the maximum combined size of all files mapped shared and shared
memory, by all processes on a system.
• Only one fixed, contiguous mapping can exist of a shared mapped file in the
shared virtual address space.
7

Memory Management 7-23

 How the Kernel Executes Processes using Demand
Paging
A compiled program has a header containing information on the size of the
data and code regions .
A s a process i s created from the compiled code ( by fork and exec) , the kernel
sets up its data structures and the process starts executing its instructions from
user mode.
A page fault occurs when the process tries to access an address that is not
currently in main memory.
The kernel switches execution from user mode to kernel mode and tries to
resolve the page fault by locating the pregion containing the sought-after
virtual address. The kernel then uses the pregion's offset and region to locate
information needed for reading in the page.
In main memory, the kernel also looks for a free page in which to load the
requested page. If no free page is available, the system swaps or pages out
selected used pages to make room for the requested page.
The kernel then retrieves ( pages in ) the required page from file space on disk.
It also often pages in additional ( adjacent ) pages that the process might need.
Then the kernel sets up the page's permissions and protections, and exits back
to user mode. The process executes the instruction again, this time finding the
7 page and continuing to execute.
Pages are not loaded in memory until they are "demanded" by a process­
hence the term, demand paging.

copy-on-write
An HP-UX enhancement to earlier UNIX implementations is the technique of
"copy-on-write." The system used to copy the entire data segment of a process
every time the process forked, increasing fork time as the size of the data and
code segments increased.
On the Series 300/400, HP-UX implements copy-on-write, which enables the
system to create processes more quickly. When a process forks, the parent
process makes a duplicate image of itself at child virtual addresses. But instead

7-24 Memory Management

of actually copying the pages, the parent marks them "copy-on-write." The
pages continue to reside in the parent address space, and are not copied until
the process actually writes on the page. Then, the old page is copied to a new
page, giving the process access to the new page. More specifically, data pages
being copied-on-write are shared until written, while for code pages, a private
copy is created on write.
On Series 700/800 systems, HP-UX implements a variation called
copy-on-access. Only one translation of a physical page is maintained; the
parent process can point to and read a physical page, but copies it only when
writing on the page. The child process does not have a page translation and
must copy the page for either read or write access .

Maintaining Page Availability-vhand a n d swapper Daemons

Two daemons (background processes )-vhand and swapper-are involved in
paging; the actual paging being performed by vhand, which is also known as
the pageout daemon.
vhand monitors free pages and tries to keep their number above a threshold
(shown in Figure 7- 10) . This ensures sufficient memory for the most efficient
operation of demand paging.

LG200211 _008

Figure 7 -9. The Pageout Daemon vhand

Memory Management 7-25

 As shown in Figure 7-9 the pageout daemon is like a "two-handed" clock, with
the distances between the hands representing the size of main memory available
to user processes. The first hand clears reference bits on a group of pages in an
active pregion. If the bits are still clear by the time the second hand reaches
them, the pages are paged out .
On the Series 800, the distance between the two hands is governed by a
factory-set parameter called vhandsk, or "vhand skew." The larger the
parameter, the longer pages can reside in_ main memory without being paged
out .
On the Series 300/400/700, the kernel automatically keeps an appropriate
distance between the hands .

1 00% free + ------.----­

m emory at
boot-up

Decreasing
Available
Memory

LG200211 _007

Figure 7 -1 o. Maintaining Memory Availability

vhand decides when to start paging by determining how much free memory is
available.

7-26 Memory Management

At the upper limit of the value gpgsl im, paging occurs . The swapper daemon's
upper limit is desfree . If free memory falls below desfree , the swapper
detects the condition and becomes active also. The swapper deactivates
processes and prevents them from running, thus reducing the rate at which
new pages are accessed. Once swapper detects that available memory has risen
above desfree, swapper reactivates the deactivated processes and continues
monitoring memory availability.

Thrashing
On systems with very demanding memory needs ( for example, systems that run
many large processes ) , the paging daemons can become so busy swapping pages
in and out that the system spends too much time swapping and not enough
time running processes.
When this happens , system performance degrades rapidly, sometimes to such a
degree that nothing seems to be happening. At this point , the system is said to
be thrashing, because it is doing more overhead than productive work.
Consider, for example, that your system might be thrashing because it is
handling a lot of disk activity. Perhaps a data base is in constant use, yet on
the same disk is swap space. ( On the Series 800, you can check system activity
using the sar(l) command; see the manual page in the HP- UX Reference.) If
so, you can minimize disk activity by moving a busy file system to a different
disk, so that you are distributing busy disk activity among different spindles.
In many cases, even this approach is insufficient . Thrashing is often eliminated 7
by adding more main memory to the system, thus alleviating the need to swap.
This reduces the amount of time the system spends paging and swapping.

Memory Management 7-27

 How the Memory-Management System Handles
Executable Code
The speed at which processes execute is related partly to how the operating
system accesses virtual address space segments of the compiled and linked
object code ( a . out or other executables) .
If you are doing applications o r system programming, you may b e aware that
a program's magic number and internal attributes determine which type of
executable code-standard, shared, and demand loaded-are possible. (A later
subsection, "Benefits and Shortcomings of Shared and Demand-Loaded Code,"
discusses this in more practical terms. Also see magic(4) .)
The following HP- UX Reference manual pages describe magic numbers and
internal attributes in detail:
• The chatr(l) command is used to change a program's internal attributes to
shared or demand-loaded.
• The link editor ld(l) produces executable files from one or more object files
or libraries.
• magic( 4) describes predefined file types and magic numbers for HP-UX
implementations .
• a. out(4) and its Series 300/400- and 700/800-specific manual pages describe
the output file format from the assembler ( as ( l ) ) , compilers, and link editor.
7 Series 300/400/700 and Series 800 computers have different capabilities . On
the Series 300/400/700 computers, executable code can be either standard,
shared or demand-loaded. Series 800 computers do not support standard
executable code, but code can be shared or demand-loaded.
Table 7- 1 and subsequent sections describe the types of executable code by
several criteria:
• Addressing in separate segments of code and data.
• Capability of virtual address code segments being shared among multiple
processes.
• Alignment of pages to corresponding address boundaries on both disk and in
main memory, for direct memory-to-memory copy. (Object code should align
on 4K page boundaries for all HP-UX systems .)

7 ·28 Memory Management

Although page size was 2 KB for some previous Series 800 releases, executables
generated during these releases will be handled properly by the current release,
but with some performance penalty. You might want to recompile for better
performance.

Table 7 - 1 . Characteristic Types of Executable Code

Type of Computer Architectures

Executable Code
Series 300/400 Series 700/800 Both

EXEC _MAGIC - single segment (code - EXEC_MAGIC is

and data) for read supported on Series
and write 700, unsupported on
- no restrictions on Series 800. 2 - 4
alignment 1 KB-page aligned on
- no sharing disk
- code and data faulted
in as needed

S HARE_MAGIC - not necessarily page- - 4KB-page aligned on - separate segments for

aligned on disk 1 disk3 code and data
- code is read-only
- code is shared
- page aligned in memory
- code loaded into memory
entirely at execution
- code and data faulted
in as needed

DEMAND_MAGIC - 4KB-page aligned on - separate segments for

disk code and data 7
- code is read-only
- code is shared
- page aligned in memory
- page aligned in disk2
- code and data faulted
in as needed

1 Executable code that is not page aligned on disk might suffer performance degradation, since it may
be faulted in by a less optimal path.
2 Series 700 EXEC_MAGIC executables will not run on Series 800.
3 Programs built on the Series 800 can be executed on the Series 700; however, footnote 1 applies on
a 4 KB-page size Series 700 for such executables. Note too, executables generated on previous
Series 800 operating systems featuring 2K page size will run, but with some performance penalty.
You might want to recompile for better performance.

Memory Management 7-29

 Standard Executable Code (EXEC_ MAGIC)

Note EXEC _MAGIC is not supported on the Series 800.

In earlier implementations of UNIX, compiled object files consisted of code

( machine instructions ) and data that occupied the same area of memory, with
read, write, and execute permissions.
Under those circumstances , each time a common program like vi, for example,
was run, a distinct copy of its code was read into main memory. The code was
not shared, even when several vi processes were run simultaneously.
Because the vi code occupied area segments with write permissions , it was
vulnerable to being overwritten. File header, code, data, and debugger code
were not aligned in main memory with their corresponding page boundaries on
disk, and so all code had to be read through a buffer cache before being copied
to the user's address space. This slowed the translation.
The EXEC_MAGIC user-executable ( a . out ) format on the Series 700 allows
creation of processes with more than 1 GB of data. This expanded data area
is implemented with a new option to the linker; refer to ld(l) in the HP- UX
Reference Manual for information.
Several factors contribute to the efficiency of EXEC_MAGIC on Series 700:
7 Pages are aligned at 4 KB, improving address translation. Text for processes
greater than 1 GB is relatively small, compared to data; thus , private code does
not pose a problem. Also, when processes are greater than 1 GB , normally only
one of them is run at a time.

Shared Code (SHARE_MAG IC)

Sharing of code segments reduces the amount of code kept in main memory.
With shared code ( also known as shared code), the code and data segments are
separated, so that code can be read-only and data can have write permission.
This protects code from being overwritten.
When several processes run the same program simultaneously, the processes
use the same copy of code in main memory via pointers to the code's virtual
address space. Only one copy of the code exists in memory, regardless of

7-30 Memory Management

how many processes run the program. The system keeps track of multiple
processes sharing code by maintaining a use count . These mechanisms decrease
dramatically the amount of memory required for each user's process space.
For example, when a shared program such as vi is first loaded into the user
code area, the use count for the program is set to one. While the first process
is executing vi, if another process invokes vi also, no additional memory
is allocated because the code already resides in main memory. The new
process merely executes the copy of vi's code residing in main memory; the
shared-code segment use count is incremented from one to two. When one of
these processes :finishes executing the code, the use count is decremented.
As long as the use count is greater than zero, the shared code remains in
memory. When the last process :finishes editing and terminates the vi program,
however, the system decrements the use count to zero and releases vi's shared
code data structure and its associated physical memory.
Shared code was also designed to facilitate page alignment between main
memory and disk, although pages are not guaranteed to be aligned.

Demand-Loaded Code (DEMAND_MAGIC)

Note For the Series 700/800, DEMAND _MAGIC code behaves

identically to SHARE_MAGIC , despite a difference in the
magic number.
7
Demand-loaded code encompasses some of the advantages of shared code
and provides additional optimizations. (See comparison of shared and
demand-loaded code in the next section. )
Like shared code, demand-loaded code addresses code and data separately.
Demand-loaded code is also shared; only one copy of code need be in main
memory for use by multiple processes .
When a user runs a demand-loadable program, pages of the program are read
into memory only as required. The system also anticipates (from prior page
usage) what subsequent pages might be required, and brings in additional
pages.

Memory Management 7-3 1

 Demand loading eliminates the need to allocate main memory to rarely
accessed routines and code, such as error handling routines, which in some
instances might account for a large percentage of a program's code.

D isk B uffe r Cache Memory � Disk Memory

Header
,--P .____.. I
I
Code

Data

I
I
I
Other

N on-Page-Aligned Code � Page-Al igned Code

LG200211 _008

Figure 7 - 1 1 .
Alignment of Page Boundaries Simplifies Transfer of Data and Code to Memory

Another demand-loaded code enhancement is guaranteed page alignment . This

is illustrated by Figure 7- 1 1 . Guaranteed page alignment is based on a loading
algorithm simpler for the system to implement . One-to-one mapping between
paging device and main-memory pages allows for direct disk-to-memory
7 transfer without an intermediary file-system buffer cache. Individu al pages can
also be copied faster.
When a executing process faults on a page, the process looks in the page cache
(a data structure that describes every page of physical memory) for the page.
If used in a recent process, the sought-after page is likely to be present . If the
page is present , the kernel maps in its address for use by the process. If the
page is not present , the kernel calculates the page's location from the inode
(which was read into memory when the program began to execute) and maps
in the page from disk.
Most executable code is now page aligned on 4KB page boundaries. Exceptions
are older a . out files of Series 300/400 systems and executables that were
compiled on Series 800 systems that were 2 KB-page aligned. When faulting in
pages for executables whose code is not page aligned, pages must be copied in

7-32 Memory Management

through the file-system buffer cache, a slower method than mapping through
the page cache.

Benefits and Shortcomings of Shared and Demand-Loaded Code

If you work in an environment where applications are written in-house, you
have some choice about how to link the object code for optimal performance.
You might consider the following questions:
• How is the program linked by default?
• How do you expect the process's pages to be used?
o At random?
o Serially?
• Are your programs running on a system with ample or limited memory?
• Are there many or few users?
If an application program is running more slowly than expected, using chatr(l)
or relinking to demand-loaded might improve its execution time.

Comparison of Shared and Demand-Loaded Code

The following points might help you determine which kind of executable to use:
• Most programs shipped as shared code by default .
You can tell how the code is shipped by running the f ile command on the
executable, for example:
Y. file /bin/cat 7
/bin/cat : s800 shared execut able
• Demand-loaded code gives flexibility; you can relink user code with ld(l) or
mark programs demand-loaded with chatr(l).
• Both shared code and demand-loaded code reduce the amount of memory
required for user code space when multiple process execute the same
program.
• For both shared code and demand-loaded code, less memory required to run
programs if only a subset of their pages are used, because pages are loaded
only as needed.

Memory Management 7-33

 Shared Code in an HP-UX Cluster (Series 300/400/700 only)
In HP-UX clusters, the benefits of shared code are reduced because each cluster
node uses its own RAM . In other words, even though a piece of code may
be marked as shared, if users on different cluster nodes invoke the program,
each cluster node will still have its own copy of the code in its own memory.
However, multiple users of a single cluster node would still share one copy of
the code.

Shared Libraries
A shared library is a collection of commonly used subroutines located in one
shared location in memory and which can be invoked dynamically at run-time
by programs that need it .
The libc, libm, l ibM ( the new PO SIX-conformant math library ) , as
well as some X and Starbase libraries are now provided in the /lib and
/usr/l ib directories as shared libraries. A set of shared library routines­
/usr/lib/libdld . sl - also provides a user interface to the dynamic loader.
Shared libraries are distinguished from archived libraries by suffix; for example,
the archive form of libc libraries are designated l ibc . a, whereas their
shared-library counterparts are designated libc . s l .

7
A shared library reduces the amount of memory occupied b y code during
execution because only one copy of its routines exists in memory.
When you include a call to an archived library, on the other hand, the library
code is copied to the executable file of the process at link time. Thus , each
executable using a routine has a copy of that routine both on disk and in
memory when running.

7-34 Memory Management

 user A running sh user A running sh

libc.a libc.sl

printf
-:::! printf stub to printf ,..... printf

user B running vi user B running vi

- printf printf
printf copied at stub to printf mapped at
link time run time
""'------

Archived Libraries Shared Libraries

LG20021 1 _01 0

Figure 7 · 1 2 . Shared Libraries Use Memory M ore Efficiently

Figure 7- 12 illustrates the use of printf in libc . A program using archived

libraries copies the printf library when linking printf . With shared libraries,
l ibc . sl is loaded once, as a single entity. Only stubs to the printf 's offset
within l ibc . sl are added to the address space of the user program. The
7
calling program maps to printf at run-time.
Processes using shared libraries no longer require their own copy of the library.
All processes calling the same library use the same image of that library, which
is dynamically loaded and linked into executing process at run time. Even if
several concurrent processes use the same shared library code, only one copy
of the code exists in core memory. The memory image of a shared library is
· shared among all programs using that library.

Memory Management 7·35

 How Shared Libraries are Dynamically Loaded
Shared libraries represent a change in the behavior of the compiled object-file
code that gets executed before main ( ) . That code ( crtO . o ) invokes the
dynamic loader (dld . sl), which consults a table stored at the beginning of
the process 's code segment to determine what libraries to load. The executing
program attaches all shared libraries specified on the linker command line. The
library source is compiled as position-independent code. The object modules of
this position-independent code are combined by the linker to form an object file
given the extension . sl, for shared library.
When loaded into a system, shared libraries reside in main memory. Like
demand-loaded code , only one physical copy of a shared library exists in
memory; code and data are faulted in as needed. Library code is shared among
all processes using that library. A private copy of library data is allocated for
each process.
On attaching a shared library, the start-up code maps the shared library code
and data into the process' address space, relocates any pointers in the shared
library data that depend on the actual virtual address, allo cates B S S , and
binds all references into and out of the shared library by filling in linkage table
entries in the program and the shared library.
To accelerate the program startup process, binding is normally deferred until
the program actually calls the shared library routine. Each linkage table entry
is initialized with a pointer to the dynamic loader. The first reference to each
7 routine causes the dynamic loader to intercept the call and bind the reference
to that routine. This deferred binding distributes the cost of symbol table
lookup across the execution time of the program, and is especially useful if the
program contains many references not likely to be executed.
The user can specify at link time that immediate binding be used. (On the
Series 700/800, the chatr command has also been enhanced to allow you
to specify binding mode without relinking. ) Immediate binding causes all
references to be bound at start-up time; thus, the cost of symbol table lookup
is taken at startup rather than spread across the execution of the program.
Immediate binding detects unresolved symbols at load time, rather than during
the execution of the program . The default is deferred binding.

7-36 Memory Management

 Before sh cal ls pri ntf
user running sh libc.sl

fopen
stub to dld.sl (fopen) � did.SI I
stub to dld.sl (print1) v printf

After sh calls pri ntf

user running sh libc.sl

fopen 7
stub to dld.sl (fopen) H did.SI I
stub to dld.sl (print1) printf

LG200211 _009

Figure 7 - 1 3. The Dynamic Loader at Work

Figure 7- 13 shows how the dynamic loader (dld . sl) handles processes when a
user invokes sh. A process is created with multiple stubs pointing to dld . sl.
In this example (more specifically for Series 700/800), two stubs are shown­
one stub for fopen(3S ) , other for printf (3S )-both of which are in the same
library. (All the platforms-Series 300/400/700/800-use a linkage table as
well; Series 700/800 use stubs also.)

Memory Management 7·37

 When sh requires printf , the process first uses its stub to dld . sl, which does
the work of establishing the pointer to printf and replacing its own stub with
a stub to printf . Thereafter, when the process needs printf , a stub is in
place to go right to printf , without having to point back to dld . sl .
Note that in the second half of the diagram, the stub intended for fopen i s in
place, although still pointing to dld . sl . However, it is ready to be accessed
and replaced by a stub pointing to fopen.
Pointers also exist from the libc routines to dld . sl , but are not illustrated.

Shared Libraries vs. Archived Libraries

Occasionally, you may prefer to use an archived library when you link your
application, in order to insulate your application from future modifications to
the library. You can prevent a library change from adversely affe cting your
application by linking in the archived version of the library.
Usually, however, an application program benefits from updates to a library.
Bug fixes and enhancements to a shared library result in automatic bug fixes
and enhancements to a program that uses the library, and new features in
a library should not affect an existing application. You can also use version
numbers described in Programming on HP- UX to ensure that you execute a
particular version when you do use shared libraries.
Another reason for using an archived library would be that the program will
execute on another system that does not have the shared library. A program
7
that uses a shared library cannot stand alone; the shared library must be
present at run time. ( An environment variable, SHLIB_PATH enables you to
specify search path for shared libraries; see Programming on HP- UX .) By
copying library code from an archive library into your program, you can
prepare a program that is independent of any other system file.
Another difference between shared and archive libraries is performance. For
object code to be shared, it must be position-independent ; that is, the library
cannot know where its code or data will reside while it is running, since
different programs may place the library code and data at different addresses .
( Text might vary the first time the library is loaded and stay fixed until the
last reference is gone; data will be at a different address for each program. )

7-38 Memory Management

In most cases, the performance improvement resulting from using shared code
and smaller executable files will outweigh the penalties of position-independent
code. However, if your library will be used by only one running application at
a time, it might perform better using an archive library.

Administering Shared Libraries

• Shared ( not archived ) libraries are the default implementation shipped. To
use archived libraries when a shared library exists requires a special linker
option.
• Consider carefully the path of shared libraries in the file-system tree. Unless
the environment variable, SHLIB_PATH is used, the shared library must be
found at run time at the same location ( that is, at the same pathname ) as it
is found at link time, or the program will not run.
o Be very careful where you put /lib , /usr/lib , /usr/local/lib and
/usr/contrib/lib. Mistakes can be harrowing!
For example, if /lib is not on the root volume, the system will try to
use it before it is mounted. As a result, no I etc/re commands ( or their
descendents ) that depend on any libraries in /lib will be executable.
Since /et c/re uses many commands that depend on /l ib/dld . sl and
/lib/libc . sl , /lib must be in the root volume. ( See fs(4) in the HP- UX
Reference for more information on volumes and file-system structure. )
• You can NFS mount /usr/local/lib , but consider that all programs that
use shared libraries from that directory will execute a little bit more slowly.
( However, once the shared libraries are used, they remain in memory for a 7
while; the NFS mount will then degrade performance only minimally. )
• For security, /lib and /usr/l ib should never to writable by all. If
/usr/local is writable by all, root should not be executing any programs
that use shared libraries from /usr/local/lib . This precaution will prevent
a potential security problem-actions of a malicious user, who could replace
a library in that directory, thereby compromising any program using it .
• Missing or incorrect critical shared libraries can be detected during the boot
procedure. If this occurs, the system administrator can recover the libraries
using recoversl( lM) or via a backup tape.
• For a more thorough discussion of shared libraries, refer to Programming on
HP- UX .

Memory Management 7-39

 HP-UX Memory-Management Features
Besides providing fundamental support for virtual memory, HP- UX provides
these important features:
• Shared memory for high-bandwidth interprocess communication ( refer to
shmget(2), shmat(2), and shmctl(2) in the HP- UX Reference).
• O n Series 300/400/700 systems , device mapping for mapping physical
addresses into virtual address space. This allows direct access to I/ 0 devices
( refer to iomap(7) and graphics(7) in the HP- UX Reference).
• Process locking for locking all or part of the user process space for real-time
application needs ( refer to plock(2) in the HP- UX Reference).
• O n Series 300/400/700 systems, memory-mapped files allow applications to
map file data into memory and perform 1/0 to files through direct loads
and stores of memory instead of read and writ e . ( Refer to mmap(2) in the
HP- UX Reference .)

7-40 Memory Management

Swap Space Management
Swap space is an area on a high-speed storage device ( almost always a disk
drive ) , reserved for use by the virtual memory system for swapping and paging
processes.
There are two types of swap space, device swap space and file-system swap
space, both of which can be configured using SAM or the swapon( lM)
command.
• Device swap space is located on a raw partition ( or raw logical volume on a
Series 800 system implementing LVM). Device swap is faster than file-system
swap . On a system with traditional disk partitions, device swap space is
fixed in size; on a system implementing LVM, the swap logical volume can be
increased as needed. Typically, at least one swap device ( primary swap ) must
be present on the system.
• File-system swap space is located on a mounted file system. File-system swap
is slower, but varies in size with the system's swapping activity.
HP-UX swap-space management allows you to allocate swap as needed ( that
is, dynamically ) while the system is running, without having to regenerate the
kernel.
Total available swap on a system consists of all swap space available on all
devices and file systems enabled as swap . The swapping subsystem reserves
swap space at process creation time, but does not allocate swap space from the
disk until swap time. Reserving swap at process creation protects the swapper
7
from running out of swap space.
In a cluster ( Series 300/400/700 only ) , multiple cluster nodes can share a single
swap area. ( Refer to Chapter 12 of Managing Clusters of HP 9000 Computers:
Sharing the HP- UX File System.)
For sample swap-space configurations and procedures on setting them up ,
consult the HP- UX System Administration Tasks manual fo r your system.
A variation on traditional swapping, called pseudo-swap reservation, allows
users to execute processes in memory without allocating physical swap .
Pseudo-swap reservation is described at the end of this chapter.

Memory Management 7 -4 1
 Swap Parameters
HP-UX deals with swapping in terms of several parameters :
swchunk The number of DEV _ BSIZE blocks in a unit of swap space, by
default , 2 MB on all systems. ( For information on DEV _ BSIZE ,
see Chapter 8 , "HFS File System" . )
maxswapchunks Maximum number of swap chunks allowed on a system.
min- Number of chunks the system reserves for swapping at boot-up .
swapchunks
swapmem_on Parameter allowing creation of more processes than you have
swap space for. ( See "Pseudo-Swap Reservation," at the end of
this chapter. )
During system startup , the size and location of each swap device is displayed in
512-KB blocks:
st art = xxxxxx indicates the swap space 's starting disk block number
s ize = xxxxxx indicates size of swap space

Device Swap Space

Device swap space resides in its own reserved area-an entire disk, or a section
or logical volume of a disk-and is not taken from file system space.
At least one swap device should be present on a system ( unless you use
pseudo-swap reservation ) . Device swap space is required during system startup .
7 You can also configure device swap space dynamically, without bringing the
system into single-user mode. Unless you are using LVM ( Series 800), device
swap is fixed in size and location.
Using LVM, you cannot reduce the size of active device swap , because virtual
memory cannot dynamically reconfigure the size of the swap device; thus,
reducing the size of active device swap might cause a system panic. You can,
however, increase the size of device swap by extending the swap logical volume
as needed. However, the system will not recognize the change until you reboot .
One or more swap device can be configured into the system, by specifying the
swap devices in the configuration file ( df ile for Series 300/400/ 700 or S800 for
Series 800). You can also configure one swap device in the configuration file
and dynamically configure more swap devices later, with swapon.

7-42 Memory M anagement

The first swap device listed in the configuration file is termed the primary swap
device; that is, the swap space used first by the operating system. Primary
swap defaults to the disk that contains the root file system.
By listing swap devices in I etc/ checkl ist , you ensure that the swap devices
are automatically enabled when the system is rebooted. The primary swap
device might be listed in / etc/checklist as , for example, the device file
/dev/dsk/c0d0 s 1 5 ( Series 800) where cO refers to the logical unit number
0, which is the first disk in the system, and s 1 5 is the section on that disk.
Or, on a Series 800 system using LVM, the device might be listed as a logical
volume, such as /dev/vg00/rlvol2 .
The Series 300/400/700 and 800 systems handle disks differently, and this
affects device swap space. On the Series 300/400/700, the entire disk is
handled as a single section, and the mkfs(lM) command is used to apportion
disk space for file system and swap . On the Series 800, disk space is divided
either by fixed sections or by logical volumes. swapon -a is run as part of
the /etc/re script , which executes when you enter multi-user state. The -a
instructs the system to read I etc/ checklist for swap information and mount
all swap sections.

Drive O . . ..
----
n
1
� Drive O
--6--- ----
Drive 1 Drive O
LIF
Drive 1

u
0 (swap) 7 (swap)
1
files root sr 7
Existing
13 13

usr
Device swap swap
(file system)

I
swap
- ­

Series Series BOO Series 800

300/400/700 with fixed sections with logical volumes
LG20021 1_0 1 1

Figure 7- 1 4. Series 300/40 0/700 and 800 Swap Space Compared

HP-UX allows you to configure swap on several disk drives , making it easy to
expand the swap space. You can have multiple swap volumes per disk.

Memory Management 7-43

 Having multiple swap devices on a system also increases throughput , by using
the principle of "interleaving swap ." Interleaving swap is achieved by assigning
identical priority to multiple swap devices to minimize disk-head movement
and enhance performance. ( See Chapter 7, "Managing Swap Space" in System
Administmtion Tasks manual. )
The required secondary storage for data, stack, and shared process segments is
reserved from swap space during bootup. Therefore, swap space must be large
enough to hold all segments of all existing processes . When it lacks sufficient
swap space, the system either returns an error ( such as ENDMEM) for some
system calls, or kills the user process.
If
you need more swap space, follow the procedures in the System
Administmtion Tasks manual to do one of the following:
• Enable an additional swap device or file-system swap to your running system.
• ( Series 300/400/700 only ) Rebuild the file system on the existing device
to reserve more swap space. To do this, you will also need to back up the
existing file system, remake the file system using mkfs(lM), then restore the
saved file system to the new configuration. For Series 700 only, run mkboot if
this is the root disk. This must be done from another system.
• ( Series 300/400/700 only ) Enable swap area to be shared within a cluster.
( See Chapter 12 of Managing Clusters of HP 9000 Computers: Sharing the
HP- UX File System for procedures. )
• ( Series 800 only ) Rebuild the kernel using SAM or uxgen to change primary
7 swap space. ( You would do this if your needs have changed, and rather than
continually allocating swap as needed, you decide to reorganize your disk
space and allo cate a different primary swap . )
• ( Series 800 only ) Create another logical volume or extend an existing swap
logical volume. Note: to make use of the extended swap volume space, you
must reboot the system.

7-44 Memory Management

File-System Swap
File-system swap, another form of secondary swap space, can also be configured
dynamically. File-system swap space allows a process to use an existing file
system if it needs more than the designated device swap space. The operating
system swaps to space in the file system, in addition to device swap space.
File-system swap is used only when device swap space is insufficient to meet
demand-paging needs.
Whereas a swap device is limited in size to a specific section of a disk ( or
logical volume), file-system swap consumes a variable amount of space. This is
an efficient use of disk space, because it enables the system to swap to unused
portions of a file system as needed.
A process might need 20 MB of swap space for a short time. If only device
swap is enabled, an entire 20 MB of swap space would have to be allocated
permanently just to handle that brief work. With file-system swap, the paging
system ( that is, swapper" daemon ) can fill these temporary needs from
file-system space. The paging system and the file system share file system
space. You can also limit file-system swap to a fixed size to prevent it from
consuming too much space.
To optimize system performance, file-system swap space is allocated and
de-allocated in swchunk-sized chunks . swchunk is a configurable operating
system parameter; its default is 2048 KB (2 MB). Once a chunk of file system
space is no longer being used by the paging system, it is released for file
system use, unless it has been preallocated with swapon. ( See swapon(lm) or 7
swapon(2) in the HP- UX Reference.)

Guidelines for Adding Swap Space

Swap configured into the system is always available to a booted and running
system. However, file-system and device swap can be allocated to a running
system for short-term use, and is used by the system like any other configured
swap device.
To make swap-space allocation automatic at boot time, you must include it in
the I etc/ checklist file. This ensures that the swap area is enabled when the
system is rebooted.

Memory Management 7-45

 If a swap area is not added to / etc/checklist , the system will no longer swap
to that area after a reboot . To disable swap , you must edit /etc/checklist
before rebooting. (These procedures can be performed using SAM . )
If you need more swap space but do not have any spare devices, we recommend
that you add file-system swap space. System performance is nearly as good
when using file-system swap space as device swap . If the system is using so
much file-system swap space that performance degrades badly, you might want
to increase device swap space.
If you have additional devices that can be used for swapping, enable these
before file-system swap , because the performance is better. If using both device
and file-system swap , give devices a higher priority (that is, a lower number in
the pri argument when executing the swapon command ) . File system swap ,
given equal priority, is always used after device swap .
For instructions on adding swap space, see Chapter 7, "Managing Swap Space"
in the System Administration Tasks manual and swapon( lM) in the HP- UX
Reference.
Sample /etc/checklist Entry for Device Swap

To enable your swap device each time the system is rebooted, be sure to
include an entry in /etc/checkl ist , as shown:
/dev/dsk/c1d0 s 1 /def ault swap pri= 1
or when using LVM (Series 800 only):
7
/dev/vg00/lvol3 /def ault swap pri= 1
Refer to the checklist(4) manual page of the HP- UX Reference for a thorough
discussion of the fields used when adding device swap.

Sample /etc/checklist Entry for File-system Swap

To enable file-system swap when the system is rebooted, include an entry in

/etc/checkl ist .
The following I etc/ checklist entry enables swap on the file system
containing the directory /swap . The size of a swap block is set by swchunk.
It is currently set to 2048 bytes , in units of dev _bs ize , which is 1024 bytes .
(2048/size of swap block x 1024 bytes = 2MB , which is the allocation size

7-46 Memory Management

of file-system swap-space chunks. ) A minimum of ten file-system blocks a.re
consumed immediately; a maximum of 4500 blocks can be used; 100 blocks a.re
reserved for file system use; and the priority is 2:
default / swap swapfs min= 10 , lim=4500 , res= 100 , pri=2
Refer to the checklist( 4) manual page of the HP- UX Reference for a thorough
discussion of the fields used when adding file-system swap.

Memory Management 7-47

 Comparing Device and File-System Swap
Device swap is faster than file-system swap. This is because the system can
write an entire request ( 64 KB on Series 300/400 or 256 KB on Series 700/800)
to a device at once. File-system swap must be done in file-system blocks, which
are typically 8 K B . As a result , the time it takes for one request in device
swap might take longer using file-system swap . For example, a 2 MB chunk of
file-system swap space is written to the disk in 256 requests in an 8 KB-block
file system, or 32 requests in a 64 KB-block file system.
File-system swap makes more efficient use of file-system space, but it might
degrade system performance somewhat , because its throughput is slower
than device swap . This is because free file-system blocks may not always be
contiguous; therefore, separate read / write requests must be made for each
file-system block.

Swap Space Priorities

Priorities, ranging from zero to ten, can be set for all devices or file systems .
The lower the number, the higher the priority. Thus , a device with a priority of
zero is used for swapping before a device of priority one or higher.
Swapping rotates among both devices and file systems of equal priority. Given
equal priority, however, devices are swapped to by the operating system before
file systems , because devices make more efficient use of CPU time.

7 We recommend that you assign the same swapping priority to most swap
devices , unless a device is significantly slower than the rest . Assigning equal
priorities limits disk head movement , which improves swapping performance.

The swapon Command

The swapon ( lM ) command allows you to enable additional device or file
system for paging and swapping. The swap space can be added while the
system is running.

7-48 Memory Management

 Sample Swap-Device Allocations

Example Description
/etc/swapon -e /dev/dsk/3s0 Enables swapping to a block device
on a Series 300/400/700 system using
the space after the end of the file
system ( -e option) for swap and
letting the priority default to l .
/etc/swapon -p 0 /dev/dsk/cOdOsO Enables swapping to a block device
on a Series 800 system with logical
number 0, CS/80 drive unit number
0, section 0, with the highest priority
(zero) .
/etc/swapon /dev/vg03/lvol4 Enables swapping on logical volume
l vol4 of volume group vg03 on a
Series 800 system using LVM . (See
the section on Device Swap for
limitations to changing the size of
device swap . Also, see chapter 7,
"Managing Logical Volumes" in the
System Administration Tasks manual
for use of LVM commands .)
/etc/swapon -p 2 -f /dev/dsk/c12d0s0 Forces swapping (-f option) to be
enabled on Series 800 block device
/dev/dsk/c 1 2d0s 0 , even if a file 7
system exists on the device. (Note
that the -f option makes this a
potentially destructive command, to
be used with caution.) The device is
assigned a swapping priority of 2.

Memory Management 7-49

 Sample File-System Swap Allocations

Example Description
/etc/swapon /swap -m 256 -1 1024 -r 0 Enables file-system swap to a
-p 3 directory named /swap . Initially, 256
file-system blocks are consumed; no
more than 1 024 file-system blocks
may be consumed. Also, 0 blocks are
reserved for file-system use and this
file system is assigned a swapping
priority of 3 .
/etc/swapon /disk2 -m 0 -1 2048 -r 0 Enables file-system swap t o a
-p 10 directory named /disk2 . Initially, no
( zero ) file-system blocks are
consumed; no more than 2048
file-system blocks maximum may be
consumed. No file-system blocks are
reserved for file-system use. The
swapping priority is 10, meaning that
/disk2 is the least likely file system
to be used for swap space .
/etc/swapon /bigdisk -m 1024 -1 4096 Enables file-system swap to a
-r O -p O directory named /bigdisk. Initially,
1024 file system blocks are consumed,
no more than a maximum of 4096
7 file-system blocks may be consumed.
No file-system blocks are reserved.
/bigdisk is assigned a swapping
priority of 0, meaning it will be used
most often for swap space.
/etc/swapon /dyndisk -m 0 -1 0 -r 2048 Enables file-system swap to a
-p 0 directory named /dyndisk. Although
no file-system space is allocated
initially, no limit is set on the amount
of space that can be used. 2048
blocks are reserved for file system
use . /dyndisk has a swapping
priority of 0-the highest .

7-50 Memory Management

Evaluating Swap-Space Needs
As a system administrator, you need to monitor your system's swap space
regularly, because swap space usage varies with system load. You want to
understand your system's swap requirements when demand is heaviest .
Swap space must be large enough to hold the sum of all shared memory, shared
libraries, stack, and data for the largest executable process.
The s ize command is a useful tool for acquiring information on process size:
Y. s ize /usr/b in/vi
243332 + 222992 + 147620 = 423944
s ize returns the code (text) , data, and bss (data uninitialized at the beginning
of process execution) for a program. Since no swap space is reserved for
code (the first figure) , you would combine only the second and third sizes to
estimate the size a program occupies in the swap area. Multiply the figure by
the number of users executing the program at the system's busiest time, since
each user is allocated data and bss. Repeat this calculation for each program
executed when the system is at its busiest.
Shared memory must also be included in your swap-space needs assessment .
To display the amount of shared memory your system is using at any given
moment , use the ipcs command:
Y. ipcs -b
!PC status from /dev/klnem as of Fri May 29 14 : 53 : 42 1992
7
T ID KEY MODE OWNER GROUP QBYTES
Message Queues :
q 0 Ox3c341834 -Rrw--w--w- root root 16384
q 1 Ox3e341834 --rw-r--r-- root root 264
T ID KEY MODE OWNER GROUP SEGSZ
Shared Memory :
m 0 Ox4134 1837 --rw-rw-rw- root root 512
m 1 Ox4134184f --rw-rw-rw- root root 7452
m 2 Ox4 1 1800ac --rw-rw-rw- root root 8 192
T ID KEY MODE OWNER GROUP NSEMS
Semaphores :
s 0 Ox4134184f - -ra-ra-ra- root root 2

Memory Management 7·5 1

 In this example, 512, 7452, and 8192 are shared memory segment sizes. You
would add them to the data and bss sizes. To account conservatively for page
boundaries, round up your estimate.
Because the swap allocation unit , swchunk , is by default 2MB , your figure
should be a multiple of 2MB . If you do not use a multiple of swchunk some
swap space is wasted.
The value of swchunk is derived as follows: A chunk is an integer with a
default value of 2048. Multiplying that integer by the default block size of 1024
bytes found in / etc/diskt ab , you derive the swap chunk size of 2 MB.
The figure of 2 MB is then multiplied by the operating-system parameter,
maxswapchunks , to derive the limit of swap space allowed on your system.
maxswapchunks is by default 256 (28), but can be set to any value up to 2 1 4 •
The value specified does not have t o be a power of 2.
One more parameter should be considered when determining swap-space
requirements. While overlaying the old process image with the new process
image, the exec system call uses an area of swap space to temporarily hold
arguments and environment variables. The size of this area is determined by
the configurable system parameter argdevnblks , whose default size is 256 KB .
If you need to change the amount of swap space to accomodate an application,
refer to the application's manual to see if it describes swap space requirements.
For details on setting and changing device and file-system swap space, see the
System Administration Tasks manual.
7

Swapping in an HP-UX Cluster (Series 300/400/700 only)

Swapping on a cluster server is identical to swapping on a standalone system:
the server must swap to a disk that is physically connected to it . A cluster
client can do any one of the following:
• Swap to the server's swap area.
• Swap to a swap area on the client's own local disk.
• Swap to a swap area on another client 's local disk.
To set up local swap in a clustered environment , see Managing Clusters of
HP9000 Computers: Sharing the HP- UX File System.

7-52 Memory Management

Pseudo-Swap Reservation (Series 800 only)
HP-UX ( Series 800 only ) has a new operating-system parameter to provide the
capability of using system memory for swap space. By default , swapmem_on is
set to 1, enabling pseudo-swap reservation.
Typically, some physical memory ( such as a disk ) must be configured for swap
space. When the system executes a process, swap space is usually reserved for
the entire process, in case the process must be swapped out . According to this
model, to run one gigabyte of processes, the system would have to have one
gigabyte of configured swap space. Although this protects the system from
running out of swap space, disk space reserved for swap is under-utilized if
minimal or no swapping occurs.
To avoid such waste of resources, HP- UX swap space is now configured to
access device swap space, file-system swap space, plus up to three-quarters
of system memory capacity. This means that system memory serves two
functions: as process-execution space and as swap space.
System memory used for swap space is called pseudo-swap space. By using
pseudo-swap space, a one-gigabyte memory system with one-gigabyte of swap
can run up to 1 .75 GB of processes. As before, if any process attempts to grow
or be created beyond this extended threshold, the process will fail.
Pseudo-swap space is set to a maximum of three-quarters of system memory
because the system starts paging once three-quarters of system available
memory has been used ( see "Maintaining Page A vailability-vhand and
7
swapper Daemons" ) .
For factory-floor systems ( such as controllers ) , which perform best when the
entire application is resident in memory, pseudo-swap space can be used to
enhance performance: you can either lock the application in memory or make
sure the total number of processes created does not exceed three-quarters of
system memory.
Note, however, that when the number of processes created approaches capacity,
the system might exhibit thrashing and a decrease in system response time. If
necessary, you can disable pseudo-swap space by setting the tunable parameter
swapmem_on in /etc/master to zero.

M emory Management 7-53

 8
H FS File System

HP-UX uses a file system called the High Performance File System ( HFS ),
which is also known as the McKusick ( or BSD) file system. A Beginner's Guide
to HP- UX discusses the file system at the user level; this chapter describes
the structure of the file system and its relationship to the disks on which file
systems reside.
The following additional resources are useful in developing an in-depth
understanding of the HFS file system and how to administer it:
• System Administration Tasks manual, for creating and managing file systems
and disk space.
• Solving HP- UX Problems for repairing file systems and dealing with problems
that arise in managing file systems and disk space.
• section ( 4) of the HP- UX Reference , for specifications of HFS file-system
formats.
To work effectively with file systems, you must understand their
interrelationship with physical disks. Every file of the HFS file system is
stored on a formatted mass storage medium, a disk. The disk is known to
HP-UX by specifying the path name to the disk's device file. Device drivers
in the operating system enable communication to the disk. Each architecture
supports a different set of disks, based on the device drivers written for that
architecture and disk. To access files in a file system, you mount the file system
on a disk, by associating the path name of its mount point to the disk's device 8
file. Once mounted, the file system is accessible to the operating system and
users.
This chapter discusses file-system creation, storage, modification, and
protection.

HFS File System 8-1

 Understanding File-System Creation
As a system administrator, much of what you do concerns file systems . System
files, application files, and user files are typically organized as file systems.
Also, although disks are the storage devices that hold data, the data must
reside in a file system to be available to the operating system. Thus, if you run
short of space, you can install a new disk and create a file system on it to hold
additional data.
Conceptually, the creating a file system involves:
• m aking the physical environment (the disk device) available to the file
system.
• creating the software entity (the file system) itself.
• establishing (by mounting) the "connective threads" between the physical
and software elements.
HP-UX uses the term "file system" to mean several things: A file system is
the HP-UX file-system (often several file systems mounted together) directory
tree, starting from root . File system is also a body of structures that exist
on each file-system device that enables you to keep data contiguous with the
existing data hierarchy. This second meaning of file system is the subject of
this chapter. This section summarizes the numerous aspects of file system
creation, to explain how a file system is connected to HP-UX as a whole.

Note All procedures for creating and maintaining file systems are
found in System Administration Tasks manual, Chapter 5,
"Managing the File System."

There are many reasons why you might add a new file system, including:
• You anticipate that your file system will soon exceed current maximum
8
capacity.
• Your current file system has already reached maximum capacity.
• You wish to separate portions of a file system physically, to restrict growth of
files on a portion of the file system or to increase concurrent access for better
performance.

8-2 HFS File System

To create a file system, you can use a sequence of HP-UX commands, or you
can invoke the SAM utility and perform the task interactively. In either case,
adding a file system involves:
• Installi n g the necessary device files for the new device ( done if disk is newly
connected )
• Preparing the storage medium ( the disk device ) for the file system ( if disk is
newly connected )
• Creating the file system itself.
• Mounting the file system to make it available for system use.
• Adding the file system to I etc/ checklist for automatic mounting.
If you are creating your new file system on a new disk drive, you first connect
the physical device to the system, referring to the device's installation manual.
Use a hard disk to hold an HP-UX file system. The capacity of flexible disks ,
cartridge and reel tape drives is too limited, slow, and subject to deterioration
from such constant use. Rewritable magneto-optical disks are slower than hard
disks, but substantially faster than flexible disks or tape, and are typically used
to back up a file system. If necessary, magneto-optical disks can be used to
hold an auxiliary file system.
Figure 8- 1 shows the physical and logical orientation of a file system to your
overall computing environment.

HFS File System 8-3

 .- - - - - - - - - ...,
- � �ev �--· ��
I Device Drivers I
I Device Special Files I
L - - - - - - - - -'

bin usr users

I I I
I I I I I I I
ada Is • • • yacc adm lib • • • tmp cindy joe • • •

LG20021 1 _0 1 9

Figure 8 - 1 . File System, Viewed Physically and Logically

Each type of disk is accessed physically via a compatible interface card that
connects the disk to the computer's bus architecture. Hard disk drives might
use any of the following interfaces-standard or high-speed HP-IB , fiber
8 link ( HP-FL ) , or small computer systems interface ( SCSI ) . The protocol for
each interface is encoded in a specific device driver. ( See Chapter 9 , "System
Configuration," for information on device drivers, device files, and configuration
concepts . )
The operating system accesses physical devices logically through both the
device driver and device special files.

8-4 HFS File System

• You can see the device drivers used by your system by reading the df ile (for
Series 300, 400, or 700 systems) or the S800 file (for Series 800 systems), or
by running the lsdev(lM) command.
• You can see device special files for disks by listing the /dev/dsk (for block
speci al files) and /dev/rdsk (for character special files) directories.
You create device files using the mknod(lM) command. For Series 800 systems ,
you can also create device files using the mksf( lM) or insf ( lM) commands.
(On the Series 800, device files for all devices the system finds are created by
insf , which is run as the -i option of ioinit by the /etc/ inittab file at
boot-up time. Thus, you should never create device files manually for a new
disk, unless it was not connected or powered on at boot time. )
Both character and block device special files are required for devices that hold
file systems. Thus, for Series 300, 400, and 700 systems, you need a character
and block device special file for the entire disk drive. For Series 800 systems,
you need a character and block device special file for each section of the disk
drive in use. Or, on Series 800 systems, if you are apportioning disk space
using the Logical Volume Manager (LYM), you need a character and block
device special file for each logic al volume. (LYM is the subject of chapters in
both this manual and the System Administration Ta�ks manual.)
The device special files are used when performing system administration tasks
involving the file system. For example,
• The mediainit( lM) command requires a transparent special file to reformat
a disk or tape for a file system. Use mediaini t if you suspect the media is
corrupted or worn. To use mediainit , you must create the device files using
·

the -t of mksf ( lM).

• The mount(lM) requires block device files to mount and unmount (umount )
the file system.
• The newfs(lM) command requires a character special file to create a file
8
system.
In Figure 8-2, you can see the role of newf s in preparing the disk to accept a
file system.

HFS File System 8-5

 /etc/disktab entry provides:
#sectors/track
#tracks/cylinder
#cylinders/disk
size of file system
block size
fragment size
#bytes/sector
rotational speed

LG20021 1 _021

Figure 8-2. newfs Builds the Disk Infrastructure for a File System

HP-UX cannot use media to store files until you place a file system on it. You
can create a new file system using SA M , mkfs(lM), or newfs(lM). Of the two
manual commands, newfs is easier to use. When you create a file system, you
create an environment to contain files, much like building a "file cabinet" for
paper files. When first built , the file cabinet is empty. Then you add files.
To create a file system, you specify disk type to the newfs command, which
then locates it in the I etc/ diskt ab file of disk characteristics and extracts
key values, including block and fragment size, number of bytes per inode,
8 percentage of reserved free space, and rotational delay. The newf s command
uses this information to create the infrastructure on the media necessary to
support the file system's data structures.
Procedures for building a file system are documented in System Administration
Tasks manual, Chapter 5, section entitled, "Creating File Systems Using the
newfs Command."

8-6 HFS File System

 I
I
Root File System
/dev/dsk/OsO
bin usr users

I
I
File System , - - - -, - - - - -1
/dev/dsk/1 so Cindy joe zack

LG200211 _022

Figure 8-3. File System /dev/dsk/1 s0 Mounted to Root File System at /users

After creating a file system, the file system has to be mounted ( attached ) to
the HP-UX file hierarchy, using the mount(lM) command. This incorporates
the file system into the existing file system's overall hierarchy. You do this by
logically associating the root directory of the new file system with a mount
point, a directory on the existing file system. Once a file system is mounted,
the mount points are seamless . You can access the new mounted disk space as
a contiguous part of the entire HP-UX file-system hierarchy. Figure 8-3 shows
a mounted file system.
To mount a file system:
• make a mount point directory ( using the mkdir command ) for the file
system.
• mount the newly created file system to the mount point ( using the mount
command ) .
If need be, an existing file system can be moved to a different location on the 8
HP- UX file hierarchy by unmounting ( detaching ) it from its current location
using the -u option ( or umount command ) of mount(lM) and remounting
the file system. A file system cannot be unmounted if any files are open or if
any user's current working directory is in that file system. You can use the
fuser(lM) command to identify which processes are using a file system or file
structure, and if necessary, terminate them. The shutdown command unmounts

HFS File System 8-7

 all mounted file systems before bringing a system down, so that the file systems
are not corrupted.
You cannot unmount the root file system or any file system that has dynamic
swap enabled. Likewise, be sure that the I etc , I dev, and /bin directories are
mounted directly on the root file system, so that they cannot be inadvertently
unmounted.
For mounting, you refer to the file system by its section's device file name ( or
with LVM, its logical volume ) and its mount point directory. For unmounting,
you refer to the file system by either the device file name or mount point ,
because unmounting breaks the link between the two.
As a system administrator, you maintain the I etc/ checklist file as a record
of mountable file systems and swap space. The I etc/ checklist file is read:
• by I etc/re , to mount all listed file systems when the system is booted up .
• by fsck(lM), to determine the order for conducting file-system checks.
• by shutdown( lM), to unmount all file systems before halting the system.
• by library calls such as getfsent(3X) and getmntent(3X), which enable
programs to make use of file system information.

Disk Layout
The disk layout is the geometry applied to a physical disk. For each platform,
the /etc/disktab file lists the disk geometry of each disk model available.
Typically, a disk is divided into areas that can accommodate file systems or
raw I/O , dump , and swap . A disk from which the system can be booted is
called a root disk, is organized somewhat differently from other disks, and is
8 discussed later in this chapter. Non-root disks typically contain a single swap
area, file systems , or a combination of both.
The following sections show the layouts of HP-UX disks for each architecture.
Note, only Series 800 disks support disk sectioning and logical volume
management .

8·8 HFS File System

Series 300/400 Disk Layout
On Series 300 / 400, a root disk consists of three areas, each discussed later in
this chapter:
• An 8-KB boot area, which contains a small file system in Logical Interchange
Format (LIF) and the secondary loader, which loads in the HP-UX kernel,
/hp-we.
• A file system, consisting of a primary superblock and one or more cylinder
groups. (Superblocks and cylinder groups are discussed in a later section
entitled "File system Layout." )
• An area reserved for swap .
Most Series 300/400 disks cannot be not partitioned; each disk has at most
one area for swap, file system, and boot . However, for Series 300/400 disks
that do support "hard" partitions, such as the HP 9133H, each partition can
be addressed separately; the volume field in the minor address indicates the
partition. Treat these partitions as though they are several separate disks, each
conforming to the layout shown in Table 8- 1 .

Table 8- 1 . Series 300/400 Root Disk Layout

Area Data Structure Size

Boot area LIF file system 8 KB
File system and Dynamic Swap Superblocks 1 8 KB
Cylinder group 1 varies
Cylinder group 2

Cylinder group n 8

Reserved swap area Swap tables2 0 or more

blocks

1 primary and redundant.

2 structures defined in /usr/ lll.clude/sys/svap . h

HFS File System 8-9

 Series 700 Disk Layout
Except for the ( optional ) boot area, the Series 700 disk is laid out much like
the Series 300/400, as shown in Table 8-2.
• The LIF directory, the first 8 KB of the Series 700, contains pointers to the
file system and to each boot program in the boot area.
• File system and swap areas resemble those of the Series 300/400.
• In the absence of a boot area, the swap area occupies the remaining space.
• The Series 700 boot program occupies the last 2 MB of the root disk layout .

Table 8-2. Series 700 Disk Layout

Area Data Structure Size

Boot pointers LIF directory 8 KB
File system and Dynamic Swap Superblocks 1 8 KB
Cylinder group 1 varies
Cylinder group 2

Cylinder group n

Swap Swap tables2 0 or more

blocks
Boot area3 LIF file system 2 MB

1 primary and redundant .

2 structures defined in /usr/mclude/sys/svap . h
8
3 optional.

A Series 700 feature, Software Disk Striping ( SDS ) , groups disks into arrays;
data on these disks can then be managed as if occupying a single disk. For
full information, see sdsadmin(lM) in the HP- UX Reference Manual and the
manual, Improving Performance with Software Disk Striping.

8- 1 0 HFS File System

Series 800 Disk Layout
Series 800 disks differ from those of Series 300/400/700 because they can be
apportioned into sections ( partitions ) to hold multiple file systems .
Using the logical volume manager ( LVM ) on the Series 800, you can also
configure your disks with logical volumes, which are similar to disk sections,
but are more flexible. ( You can span disks with logical volumes, and extend or
reduce their size as needs change. ) LVM is the subject of a separate chapter in
this manual, as well as in the System Administration Tasks manual.
Because Series 800 uses disk sections , both the boot area and primary swap
occupy their own disk sections.
Disk space can be partitioned on the Series 800 in a variety of ways, as
discussed in the next section.

Understanding Disk Partitioning (Series 800 only)

Disk partitioning is a Series 800 feature that allows you to divide a disk into
sections, which can be addressed like separate disk drives. Disk sections allow
the system administrator to manage disk space more precisely than on a
system without disk sections.
A section can used for:
• Boot area
• File system
• Swap area
• Raw 1/0
The layout of each section is nearly identical to the same areas on the Series 8
300/400/700. However, on the Series 800, the boot and swap areas reside in
their own sections instead of residing in the same section as the file system.
As of HP-UX 9.0, the Series 800 can be partitioned using two different
methods:
• HP-UX disk sections, the traditional disk partitioning method.
• Logical volumes, managed by the LVM.

HFS File System 8· 1 1

 HP-UX Disk Sections
Series 800 disks are partitioned into sixteen possible section choices. The size
and location of each section is dependent on the disk model. Section sizes and
locations are defined in the I etc/ disktab file ( described in further detail in a
coming section ) for most supported disk models. Figure 8-4 shows the choices
of hard-coded sections available for Series 800 disks. ( Note that disks smaller
than 350 MB have a slightly different layout . See /et c/diskt ab on your
system for layout . )

LG20021 1 _023

Figure 8·4. Disk Sections and R elative Sizes

When you create a new file system using the mkfs or newfs commands, or the
8
SAM program, you declare what section the file system is to be mounted on.
( For procedures on creating a file system, see the System Administration Tasks
manual. ) You must be careful not to use overlapping disk sections ( as shown in
Figure 8-4 and at the beginning of the / etc/diskt ab file ) .
For example, if you create a file system on section 1 1 on the 7935 disk, you
cannot allocate sections 2, 3, 4, 5, 8, 9, 10, 12, and 13 for other file systems,

8-12 HFS File System

because they would conflict with section 1 1 . You can put file systems in
sections 1, O, and 6, or 6 , 14, and 15.
Figure 8-5 shows two valid ways to set up disk sections . Compare this figure
with Figure 8-4 to understand how to make the most of disk sections.

6 1-------1
0 1-------t 7
1 1-------t
1 0 1-------t 1 0 ....,_____....

3 1-------t
8
4 1-------1
5 ._____....

LG2001 72_035

Figure 8-5. Examples of Yalid Sectioning Schemes

Logical Volumes
The Logical Volume Manager (LVM) , available on Series 800 systems, enables
you to partition disks in a more flexible manner than with traditional disk
sections. Using LVM, you combine one or more disks (physical volumes) into a
volume group , which can then be subdivided into logical volumes.
Logical volumes resemble disk sections, but with some important differences:
• You can define the size of a logic al volume according to need.
• You can extend or reduce the size of logical volume as needs change.
• Logical volumes can span disks . This enables you to create very large logical 8
volumes, or use small portions of disk space more efficiently.
• You can mirror logical volumes, using an optional product , LVM Mirroring.
Logical Volume Manager (LVM) is the subject of a separate chapter in this
manual and in System Administration Tasks manual.

HFS File System 8- 1 3

 File System Size
HP-UX supports file systems up to 4 GB. Note, however, that the size limit for
individual files is still 2 GB. Applications may also not use raw access to disk
sections larger than 2 GB .
For very large disks ( such as HP C2254B ) , the boot partition must lie within 2
GB of the beginning of the disk.
Protocols do not permit NFS-mounting file systems larger than 2 GB.

Disk and File System Tools

When working with file systems, you often have to understand how much
disk space you have and how large your file systems are. Two HP-UX tools­
/ etc/ disktab and I etc/ diskinfo-can help you determine available disk
space. To view how large a file system is that you want to mount , you can use
bdf , df or du. Each of these tools is discussed in the following sections.

Disk Geometry Database-/etc/disktab

The I etc/ diskt ab file exists on all systems as a database and informational
file about disks. It provides you with both reference information about the
many HP disks supported on a given computer system and tutorial information
about disk geometry.
Because it is a database ( and thus written for the system, rather than to be
read by people ) , the information in /etc/disktab appears in terse form, as
follows:
8 ty Type of disk.
ns Number of lK sectors per track.
nt Number of tracks per cylinder.
nc Total number of cylinders per disk.
so Size of file system ( Series 300/400/700) or section ( Series
800) in lK blocks. On Series 800, so designates section O ;
subsequent sections are listed s 1 through s 15.

8· 1 4 H F S File System
bO Series 300/400/700: Block size in bytes .
Series 800: Block size of section 0 in bytes. Subsequent
block sizes are listed bi through b15, corresponding to their
respective section. Default block size for all systems is 8K.
fO Series 300/400/700: Fragment sizes in bytes ( lK, 2K, or 4K
are supported ) .
Series 800: Fragment size of section 0 in bytes. Subsequent
fragment sizes are listed f 1 through f 15, corresponding to their
respective section. Default fragment size for all systems is l K .
se Number of bytes per physical sector.
rm Rotational speed of disk platters by revolutions per minute.
Not all abbreviations are used on all systems.
The contents of /etc/disktab are used by the newfs command. On the Series
300/400/700, / etc/diskt ab provides multiple entries for any given disk, to
enable you to specify to newf s whether you want portions of a disk used for
swap and boot .
If you are partitioning your disks using traditional disk sections, the Series 800
I etc/ diskt ab file is useful for determining valid disk sections because it shows
Figure 8-4, from which you can ensure that you neither overlap disk sections
or leave gaps. /etc/diskt ab 's listing of each disk with the size of each section
enables you to calculate available disk space.
If you are using the LVM , you have less cause to refer to / etc/disktab ,
although you might refer t o i t when you want t o use non-default settings for
file-system specification ( for example, to change the fragment size, customize
the various file-system sizes ) . Before adding a physical volume ( disk ) to a
volume group , you might consult / et c/diskt ab to get an idea of the disk size.
Be forewarned, however, that once you add the physical volume, the LVM uses
a portion of the disk ( as much as 1 MB ) for its data structures , so the size
8
provided by reading section 2 of I et c/ diskt ab for a given disk is larger than
the actual size you have available.
For full specifications, see disktab(4) in the HP- UX Reference.

HFS File System 8- 1 5

 Disk Characteristics Command-/etc/diskinfo
The diskinfo( lM) command displays characteristics of a disk device, when
given the device's character special file. I etc/ diskinfo is particularly useful
when setting up or managing logical volumes.
When used without options , / etc/diskinfo produces terse output :

Y. / etc/diskinfo /dev/rdsk/ c0d0s2

CS80 describe of /dev/rdsk/ c0d0s2 :

product id : 7937
type : f ixed disc
s ize : 57 1444224 bytes
bytes per sector : 256

With the -b option, /et c/diskinfo returns the size of the disk in 1024-byte
sectors.

Y. /etc/diskinfo -b /dev/rdsk/c0d0s2

55805 1

The verbose ( -v) option of I etc/ diskinfo displays different information,

depending on type of disk:
8 • vendor and product ID ( S CSI devices)
• device name (CS /80 and S CSI)
• number of bytes/sector (CS /80 and S C SI)
• geometry, interleave, and timing information (CS/80)
• size in bytes and logical blocks, revision level, SCSI conformance level (SCSI)

8- 1 6 HFS File System

% /etc/diskill£o -v /dev/rdsk/c0d0s2

CS80 describe 0£ /dev/rdsk/c0d0s2 :

controller : O:z:8001 installed unit vord
1 2SO max instant aneous :z:£r rate (Kbytes/sec)
0 CS/80 int egrated single unit controller
unit : 0 generic device type (£i:z:ed disc)
079370 device number (6 bed digits )
2 S 6 # 0 £ bytes per block
128 # 0£ blocks bu££ered
0 recommended burst s ize
179 blockt ime Cmicroseconds)
1000 cont inuous average trans£er rate (Kbytes/sec)
80 opt imal retry t ime (cent iseconds )
84 access t ime paramet er (cent iseconds)
1 maximum int erleave £actor
O:z:01 £i:z:ed volume byt e
O:z:OO removable volume byt e
volume : 139S maximum cylinder address
12 maximum head address
122 maximum sector address
2232203 maximum single-vector address
1 current int erleave £actor

HFS File System 8- 1 7

 Free Disk Blocks Command-bdf
The bdf command (Berkeley's variation of df) reports the number of free disk
blocks available on a file system. If no :file system is given as an argument , bdf
reports on all :file systems. Several options are available:
-b Displays information about file system swapping.
-i Displays used and free inodes.
-1 Local. Displays HFS and CDFS :file systems mounted on a
client . It does not display NFS-mounted file systems.
-L Displays :file systems that can be unmounted from a client ,
including NFS-mounted :file systems .
-t type Displays only information on mounted :file systems of a given
type.
Here is sample output of bdf :

% bdf

Filesystem kbyt es used avail capacity Kollllt ed on

/dev/dsk/c0d.Os13 484960 242946 193619 66% I
/dev/dsk/c6d0s8 237810 48481 166648 23% /graphics
/dev/dsk/c4d0s2 277964 130729 133327 60% /users
/dev/dsk/c3d0s10 121663 21878 87618 20% /tmp
/dev/dsk/c2d0s2 277964 196693 64466 78% /proj ects
/dev/dsk/c1d0s11 461664 320288 96209 77% /usr
/dev/dsk/c3d0s7 70979 2994 60887 6% /usr/spool
/dev/dsk/c1d0s7 71699 44662 23467 66% /usr/contrib

bdf reports its output in 1024-byte blocks.

df reports its output in 512-byte blocks .
8

8· 1 8 HFS File System

 Using LVM Device Files
All LVM components are represented by device files located in the I dev
directory. ( You might think of device files as agents for managing the
interactions with the disk space. ) The LVM device files are created by both
SAM and HP-UX commands and by default follow a standard naming
convention. However, you can choose more intuitive names for volume groups
and logical volumes.

Table 9-1 . LVM Device File Names (defaults}

LVM Component Block Special File Raw Special File

LVM disk 1 /dev/dsk/c nd0s2 /dev llJ;d sk/c ndOs22

fll ivol n 2
Volume Group /dev/vgn 3 N/A
Logical Volume /dev/vgn/lvol n 4 /dev/vgn

1 ( )
LVM disks physical volumes are named for section two, the entire
disk.
2 The r in the file name denotes "raw" (also called "character" ) and
refers to the way data is streamed, in characters rather than blocks,
or as raw data rather than blocks of data.
3 (This is not a block special file, but a directory. ) Directories for
volume groups are numbered vgOO , vg01 , vg02 , . . . vgnn , in order
created.
4 Device files for logical volumes are listed in the subdirectory for the
volume group to which they belong.

9·8 Logical Volume Manager

physical A configuration of physical volumes (LVM disks) within a
volume group volume group into a physical volume group , on the basis of the
1/0 channel or interface adapter to which they are connected
to achieve higher availability of mirrored data. (Also see "1/0
Channel Separation" , later in this chapter.)
quorum The requirement that a volume group have at least half the
configured LVM disks present to change or activate that
volume group. If there is no quorum, LVM prevents the
change.
Quorum is checked both during configuration changes (for
example, when creating a logical volume) and at state changes
(for example, if a disk fails) . If quorum is not maintained,
LVM will not acknowledge the change.
Quorum ensures the consistency and integrity of the volume
groups . The vgchange command with the -q n option can be
used to override quorum check.
root logical The logical volume containing the HP-UX kernel.
volume
synchroniza­ The process of updating "stale" (non-current) copies of
tion mirrored logical extents by copying data from a "fresh"
(current) copy of the logical volume. Synchronization keeps
mirrored logical volumes consistent by ensuring that all copies
contain the same data.
volume group A collection of one or more LVM disks from which disk space
may be allocated to logical volumes. A disk can belong to only
one volume group . A volume group is accessed through the
group file (for example, /dev/vg0 1 /group ) in that volume
group's directory.
You can use SAM or the vgcreate command to create a
volume group .
HP-IB devices may not be combined with devices of any other
interface in a volume group . 9

Logical Volume Manager 9-7

 logical volume A storage device, much like a disk section but of flexible
size, that can hold a file system (including root ) , raw data,
application program, or swap . Logical volumes can be mirrored
using an optional product , HP MirrorDisk/UX.
Because its data is distributed logically (rather than
physically) , a single logical volume can be mapped to one LVM
disk or span multiple disks, but the logical volume itself is used
as a single virtual disk.
The lvdisplay command can be used to verify distribution of
logical volumes on disks.
Logical An operating system software module that implements virtual
Volume (logical) disks to extend, mirror, and improve the performance
Manager of physical disk access.
( LVM )
LVM disk A disk that has been initialized for LVM. (Also see "physical
volume." )
mirroring Simultaneous replication of data (up to three copies) using an
optional product , HP MirrorDisk/UX (part number B249 1A).
(HP MirrorDisk/UX is not supported on HP-IB disks.)
Duplicated information storage ensures a greater degree of data
availability. Using HP MirrorDisk/UX, LVM maps logical
volumes to multiple LVM disks, thus providing the means to
recover easily from the loss of one copy (or two copies in the
case of three-way mirroring) of data.
Mirroring can provide faster access to data for database
applications using. more data reads than writes.
physical A disk that has been initialized by LVM for use in a volume
volume group; also calle d an LVM disk.
As with standard disks , an LVM disk (physical volume) is
accessed via a raw device file for section two (for example,
9
I dev I rdsk/ c3d0s 2).
You can use SAM or the pvcreate command to initialize a
disk as a physical volume (LVM disk).

9-6 Logical Volume Manager

LVM Terminology
The following LVM terminology is new to HP-UX:
allocation The LVM allocation policy governing how disk space is
policy distributed to logical volumes and how extents are laid out on
an LVM disk.
LVM allocates disk space in terms of strict vs. non-strict
and contiguous vs. non-contiguous. Strict allocation requires
that mirror copies reside on different LVM disks . Contiguous
allocation requires that no gaps exist between physical extents
on the disk.
bad block An LVM feature for recovering corrupted blocks of data from
relocation HP-FL and SCSI disks, by redirecting the data to another
block of media on the disk. Bad block relocation uses a
remainder of space in the disk layout . This feature can be
enabled or disabled with the lvchange ( lM) command.
This feature is not supported on HP-IB disks, nor is it
supported for the root logical volume.
block The smallest unit of space transferrable to and from a disk.
LVM transfers data in blocks of DEV_BSIZE ( 1024) bytes .
extent Fixed-size addressable areas of space on an LVM disk or in
memory. On disk, these contiguous areas are called physical
extents , and are by default 4 MB . Physical extents map to
areas in memory, called logical extents. (For more information
on extents, see "Allocating Disk Space for Logical Volumes"
later in this chapter.)
1/0 channel A configuration of disks useful for segregating highly
separation 1/0-intensive areas . For example, you might have a database
on one channel and file systems on another.
When mirroring logical volumes using HP MirrorDisk/UX, you
can spread the mirrored copies over different 1/0 channels to
increase system availability. 9

Logical Volume Manager 9-5

 All procedures for implementing LVM are detailed in Chapter 7 of the System
Administration Tasks manual. All LVM commands are documented in the
HP- UX Reference Manual.

Understanding Migration from . Sections to Logical Volumes

Before you can use LVM on a system whose file systems exist in fixed disk
sections, you must first migrate to LVM . You can migrate your entire system
( including the root disk ) or only the more dynamic file systems or sections of
raw data.
Migration is documented in Installing and Updating HP- UX , but briefly, the
procedure involves the following:
• Plan your new configuration to take advantage of LVM features. Use the
tool, lvmm igrate ( available on the system tape ) to plan your allocation of
space in logical volumes.
• Copy all customized files to a safe directory.
• Back up the contents of your disks . You must do this because LVM applies
new data structures to your disks and anything on your disks will be
overwritten. You can use fbackup or cpio to back up your file systems, and
dd to back up raw data.
• Install HP-UX on the root disk, which allows you to create new logic al
volumes for root and swap .
• Restore files and raw data, using frecover for file systems and dd for raw
data.
• Return customized files to the root directory.
• Use SAM to complete the migration of remaining data to logical volumes.
This involves creating and extending logical volumes to hold your data, then
recovering backed-up data onto the logical volumes.

Note LVM data structures consume a small, but significant amount

of overhead space. Therefore, be sure to plan thoroughly before
implementing LVM .

9-4 Logical Volume Manager

 Physical
Volumes
(disks)
/dev/rdsk/c0cl0s2 /dev/rdsk/c1 d0s2 /dev/rdsk/c2d0s2

/dev/vgOO
Volume
Group
(pool of disks)

-r-
_{> __ tdevtvsoo11vo11

Logical
Volumes
(apportioned
disk space)

LG200211 _027

Figure 9· 1 . Disk Space Apportioned into Logical Volumes

You can use SAM to create a file system in a logical volume of a specified size,
and then mount the file system.
Or using commands, you can create and then exten'd a logical volume to
allocate sufficient space for file system, user application, or raw data. You
would then create and mount new file systems or install your application in the
logical volume. You use the same approach when increasing the capacity of a
file system created on a logical volume. 9

Logical Volume M anager 9·3

 The LVM Paradigm
Implementing LVM requires a new way of thinking about disks and file system
configurations.
• In a traditional HP-UX configuration, you consider your disks individually
and in terms of their variously sized sections (or partitions) , each of which
holds a file system, swap space, boot area, or raw data.
• With LVM , you do not use disk sections. Instead, you consider the disks as a
pool (or volume) of data storage, consisting of equally sized ( 4 MB) extents,
multiples of which are allocated for file system, swap , and other purposes.
HP LVM lets you specify size in megabytes.
Setting up LVM requires you to follow a procedure to reorganize your
system. This is documented in Installing and Updating HP- UX and System
Administration Tasks manual.
An LVM system consists of groupings of disks initialized for LVM and
organized into volume groups. A volume group might consist of one or many
LVM disks; your entire system might consist of one or several volume groups.
Just as volume groups are groupings of one or more LVM disks, volume groups
are also subdivided into virtual disks, called logical volumes . Logical volumes
can
• encompass space on one or more LVM disk
• span LVM disks
• represent only a portion of an LVM disk.
You apportion disk space in a volume group by creating logical volumes. The
size of a logical volume is determined by its number of extents, each being
4 MB by default and is configurable. You then assign file systems, swap , and
dump to logical volumes.
In Figure 9- 1 , logical volume /dev/vg00/lvo11 might contain a file system,
/dev/vg00/lvol2 might contain swap space, and /dev/vg00/lvol3 might
contain raw data.

9-2 Logical Volume M anager

 9
Logical Volume Manager

The Logical Volume Manager (LVM) is a disk management subsystem ( Series

800 only) that lets you allocate disk space according to the specific or projected
sizes of your file systems or raw data. LVM logical volumes provide other
capabilities that are unavailable when using fixed partitions:
• LVM file systems can exceed the size of a physical disk. This feature is
known as disk spanning, because a single file system can span disks.
• Up to three copies of identical data can be stored and updated
simultaneously using LVM . This feature is known as mirroring, and requires
an optional product, HP MirrorDisk/UX (part number B2491A).

Note If you have a pre-installed 9.0 system, LYM is ready to use.

If you are upgrading to 9.0, you must "migrate" your disks

to LVM: You must backup all data from disks, set up logical
volumes on disks, and restore data to disks.

Migrating the root disk requires re-installation; see Installing

and Updating HP- UX.

LVM is most useful for:

• managing file systems whose size is likely to grow. With LVM, you can
specify the size of a file system based on need, expand a file system by
adding disk space from anywhere in your system.
• large-scale applications, such as databases and CAD/CAE systems , whose
data requirements often exceed the capacity of a single disk.
• disk mirroring for increased data integrity and high availability. LVM 's
mirroring capabilities enable you to perform some system administration 9
tasks , such as backups , without bringing the system down.

Logical Volume Manager 9-1

 Locking Activities
All file locking is controlled with the lockf or fcntl system calls. lockf
controls four file actions:
• Testing file accessibility by checking to see if another process is present on a
specific file record.
• Attempting to lock a file. If the record is already locked by another process,
lockf puts the requesting process to sleep until the record is free again.
• Testing file accessibility, locking the record if it is free, and returning
immediately if it is not.
• Unlocking a record previously locked by the requesting process.
When the locking process either closes the locked file or terminates, all locks
placed by that process are removed. For more details on how specific locking
activities work on HP-UX, refer to the lockf(2) and fcntl(2) sections of the
HP- UX Reference manual.

8-66 HFS File System

Enforcement Mode
Even if we use advisory locks in our example, Sarah would still be able to
overwrite the Marketing report if she uses commands or utilities that do not
check for advisory locks. She needs some way to ensure that no records are
written until George finishes accessing the report . HP-UX does this with
enforcement mode. When a process attempts to read or write to a locked
record in a file opened in enforcement mode, the process sleeps until the record
is unlocked. Enforcement mode can be used only on regular files .
Enforcement mode is enabled when the set-group-id bit ( sgid ) is set but the
group execute bit is not set. For example, if we opened a file which normally
has its file permission bits set to 644, a long listing of the file would look
something like:

-rw-r- -r-- 1 george fiscal 5 12 May 7 16 : 1 1 Market ing

Enforcement mode can be enabled by typing:
chmod g+s Market ing
This command turns on the sgid bit , resulting in file protection of 2644.
Enforcement mode can also be enabled by using the chmod system call. After
enforcement mode is enabled, a long listing shows:

-rw-r-Sr-- 1 george fiscal 512 May 7 16 : 1 1 Market ing

Using enforcement mode, George can prevent Sarah from overwriting his
changes, and S arah would have the data that George has added.
When attempting to access a file locked under enforcement mode, the process
sleeps until the file is released. This provides a means for one process to control
execution of another. Be careful when doing this , because a system deadlock is
possible. 8

HFS File System 8-65

 File Sharing and Locking
In a multi-user, multi-tasking environment such as HP-UX, it is often desirable
to control interaction with files . Many applications share disk files , and the
status of information contained in them could have serious implications to the
user (such as lost or inaccurate information) .
Imagine we are responsible for maintaining on-line technical reports for
a myriad of projects , and we have many different people who must have
simultaneous access to these reports. The content of a given report at a given
time could significantly affect a company decision, and so we want a way to
control how records are accessed.
One potential problem could arise if one person (let 's call him George) adds to
or modifies information in a report while someone else (Sarah) is working on
it . Sarah is unaware of changes that George has just made in the report . And
once she is done, Sarah overwrites the information George added. The result is
that we have lost all of George's information, and when Sarah added data she
was unaware of information that might have been pertinent .

Advisory Locks
A solution to this problem common to file sharing is called file locking . In
HP-UX, file locking is done with the lockf or fcntl system calls, which handle
two modes of functionality. Advisory locks are placed on disk resources to
inform (warn) other processes desiring access that a file is currently being
accessed or modified. Advisory locks are only valuable for cooperating
processes that are both aware of and use file locking.
In our example, the programs used to access the on-line reports can use
advisory locks. When George begins to work on the Marketing project his
8 program can call lockf and set an advisory lock. A few minutes later when
Sarah tries to access records in the Marketing report , she would get an error
message indicating that the report is busy. Her program could wait until
George is done and then access the report , by using the system call, lockf .

8-64 HFS File System

 Table 8-7. Explanation of File Permission Bits rwsr-sr-

Octal Binary Permission Meaning

Digit Form
1 set user ID Effective user ID of the process executing this file
is set equal to the real user ID of the file's owner.
6 1 set group ID Effective group ID of the process executing this
file is set equal to the group ID of the file's
group .
0 sticky bit The sticky bit is not set ; see "Protecting
Directories" .
1 read File owner may read the file .
7 1 write File owner may write to the file.
1 execute File owner may execute the file.
1 read Members of the file 's group (other than the file 's
owner) may read the file.
5 0 write Members of the file 's group (other than the file 's
owner) cannot write to the file .
1 execute Members of the file 's group (other than the file's
owner) may execute the file .
1 read Any other user may read the contents of the file .
4 0 write Other users cannot write to the file .
0 execute Other users cannot execute the program
contained in the file .

Access Control Lists

Access control list s ( A CLs ) offer a finer degree of file protection than
traditional file-mode protection bits. With ACLs, you can allow or restrict
file access to individual users, regardless of what group the users belong. For 8

additional information see acl(5) in the HP- UX Reference Manual manual

and/or HP- UX System Security, HP part number B 1862-90009 .

HFS File System 8-63

 File File's All
Owner Group Others

binary

' '
V ---....
v---1 '""'---..v---' '""'---..v---'
octal 6 7 5 4

where:
Q;::uo-ct-al di.git

Figure 8 - 1 2 . Permission Bits of an suid/sgid file set to rwsr-sr--

8-62 HFS File System

Setting Effective User and Group ID Bits (suid, sgid)
As discussed in Chapter 5, "Process Management ," a process has effective
user and group IDs that can be used to ensure file security. Using user and
group IDs, a file can be protected so that when executed, the process's effective
IDs are identical to the file owner's IDs. This capability is specified through
the most significant digit of the four octal file protection digits. The most
significant digit is represented by a three-bit binary value. When its most
significant bit is 1, the effective user ID of the process executing the file is set
equal to the user ID of the file's owner. This bit is called the set user ID bit
( suid or s etuid). Similarly, if the middle bit of the most significant octal digit
is 1 , then the effective group ID of the process executing the file is set equal to
the group ID of the file's group . This bit is called the set group ID bit ( sgid or
setgid) .
If the sgid bit is set for an ordinary file, and the file does not have group
execute permission, the file is in enforcement locking mode. Refer to the
section "File Sharing and Locking" later in this chapter, or to the lockf (2)
entry in the HP- UX Reference.
If the suid bit is set for a directory, the directory is hidden. A hidden
directory is part of the context-dependent file ( CDF ) structure used when
running an HP-UX cluster. For more information on CDFs, refer to Managing
Clusters of HP 9000 Computers: Sharing the HP- UX File System.
For example, suppose the file's permission bits are 6754. The binary equivalent
of octal 6754 is 1 10 1 1 1 101 100. These permissions are illustrated in
Figure 8- 12 and explained in Table 8- 7.

H F S File System 8-6 1

 File File's All
Owner Gro u p Others

inary 0 0

write write write

read read read

I I I
y y y
octal 7 5 4

where:
D =octa l di it

Figure 8- 1 1 . File Permission Bits of rwxr-xr--

Protecting Directories
Directories, like all files in the HP-UX file system, have permissions . The
format of a directory's permission bits is identical to that of an ordinary file;
however, the read, write, and execute permissions have a slightly different
meaning when applied to a directory.
• Read permission grants access to display the contents of a directory.
• Write permission grants access to add a file to the directory, rename a file
within the directory, and remove a file from the directory. Users (even
superusers) cannot write directly to the directory itself. Only the kernel can
write directly to directories.
• Execute permission grants access to search a directory for a file. If execute
permission is not set , the files below that directory in the file-system
hierarchy cannot be accessed, even when you supply the file's correct path
name.
8
Setting the sticky bit on a directory provides additional protection to files
within the directory: files cannot be removed from the directory except by
the owner of the file, the owner of the directory, or a user having appropriate
privileges. (See rm(l) in the HP- UX Reference Manual.)

8-60 H F S File System

significant digits. When the most significant digit is not specified, its value is
assumed to be zero (0).

Ignore File Fi le's All

for now Owner Group Others

inary

write write w rite

read read read

\ J J \ J I
y y y y
octa l D D D D

where :
D =octa l d i it

Figure 8 - 1 o. File Permission Bits

Each octal digit represents a three-bit binary value: one bit specifies read
permission, one bit specifies write permission, and one bit specifies execute
permission. If the bit value is one, then permission is granted for the associated
operation. Similarly, if the bit value is zero, permission is denied for the
associated operation.
For example, assume a file's permission bits are set to 754 ( octal ) . Octal
754 is equivalent to 1 1 1 101 100 binary. The 11 command represents this as
rwxr-xr-- . As shown in Figure 8- 1 1 , the file's owner may read, write, and
execute the file, while read and execute permission is granted to members of
the file-owner's group . This includes any user ( except the file's owner ) whose
effective group ID equals the ID of the file's group, or whose group access list
includes the file's group ID . All other system users may only read the file.
8
Note, if a file has associated Access Control List ( ACL ) entries , a + is displayed
following the permissions. By default , the chmod command deletes any ACL
entries , but you can use the -A option to preserve them. For more information
on ACLs, refer to A Beginner's Guide to HP- UX , HP- UX System Security, and
acl(5) in the HP- UX Reference .

HFS File System 8-59

 Table 8-6. Utilities and Services for File Transfer (continued)

Utility/ Service When to Use For More

Information . . .
t cio When transferring files between cartridge tcio ( l) in the
tape units (including autochanger) and a HP- UX Reference
controlling HP-UX computer. tcio is
typically used with cpio or tar.
fbackbup , When transferring (typically backing up fbackup( lM)
frecover and restoring) files to magnetic tape, andfrecover( lM) in
standard out , DAT tape, rewritable the HP- UX
magneto-optical disk, or to a file. Reference
Combines features of dump and ft io.

File Protection
When created, each file in the file system is assigned a set of file protections
stored in the file permissions bits ( often called the file's mode ) . The file
permission bits determine which classes of users may read from the file, write
to the file, or execute the program stored in the file. Read, write, and execute
permissions for a file can be set for the file's owner, all members of the file's
group ( other than the file's owner ) , and all other system users.
These three classes of users ( user, group , and other ) are mutually independent ;
that is, no member of one class of users is included in any other class of users.
When a file is created, it is associated with an owner and a group ID . For
example, a file created by pj w in group dbase is listed as being owned by user
8
pj w of group dbase . These values specify which user owns the file and which
group has special access capability.
The default permissions of a file are initially determined by umask ( set
systemwide, in the users' environment file, or on the command line ) , or by
parameters passed to creat , mknod, or mkdir system calls when the file is
created. The permissions can be changed with the chmod command.
File permissions are represented as the binary form of four octal digits as
shown in Figure 8- 10. The initial discussion deals with only the three least

8-58 HFS File System

 Table 8·6. Utilities and Services tor File Transfer (continued)

Utility /Service When to Use For More

Information . . .
NFS When users on different HP-UX ( and • Using NFS
other UNIX ) systems want to share files . Services
Explicit file transfers are unnecessary • Installing and
because file-system access is transparent . Administering
NFS Services
BIF Utilities When transferring files between Series bi/ ( 4) in the
300/400 systems and the HP Integral HP- UX Reference
Personal Computer or Series 200
computers.
SDF Utilities When transferring files between Series sdf( 4) in the
300/400/700/800 and Series 500 HP- UX Reference
computers.
cpio When transferring files by magnetic tape cpio ( l ) in the
( cartridge or reel-to-reel ) to another HP- UX Reference
UNIX system that supports the cpio
format . NO TE: cpio can be used with
the t cio command to ensure smoother
tape access.
ttio When copying files to magnetic tape ftio ( l ) in the
( cartridge, reel-to-reel, or DDS format ) . HP- UX Reference
Faster throughput than either tar or
cpio .
tar When transferring files by magnetic tape tar( l) in the
( cartridge, reel-to-reel, DDS format ) to HP- UX Reference
another UNIX system supporting the tar
8
format . NO TE: tar can be used with the
t cio command to ensure smoother tape
access. However, t cio only works on
cartridge tapes.

HFS File System 8-57

 Table 8-6. Utilities and Services for File Transfer

Utility/Service When to Use For More

Information . . .
kermit When both systems, connected by serial Kermit chapter of
lines, run kermit . kermit transfers data Remote Access:
between HP-UX and many incompatible User's Guide
operating systems.
LIF Utilities When transferring files between HP-UX lif (4) in the HP- UX
and systems that support the LIF file Reference
format , including HP-UX Basic , Pascal,
and other HP-UX systems.
uucp When the other system is a UNIX system U UCP chapter of
(including HP-UX) , connected by modem Remote Access:
lines, direct connection, or X.25 network, User's Guide
and with UUCP utilities installed . uucp
automatically reconciles differences in file
format between systems.
ARPA/Berkeley When the both systems are connected via • Using ARPA
Networking Services LAN to the same services. The other Services
system can be an HP-UX or UNIX • Installing and
system, or an MS-DOS personal A dministering
computer. ARPA/Berkeley Networking A RPA Services
Services reconcile automatically any • .ftp ( lM) in the
differences in file format between systems . HP- UX Referen ce
HP FTAM/9000 When both systems are networking via • HP FTA M/9000
OSI and using FTAM (OSI File Transfer , User's Guide
Access , and Management) . OSI is a • Installing and
multi-vendor standard compatible with A dministering HP
8
UNIX and non-UNIX operating systems . FTA M/9000
FTAM handles binary and ASCII file
transfers, but does no data conversion.
Network When transferring files over LAN to any Network
Services/9000 HP-UX platform. Services/9000
User's Guide

8-56 HFS File System

File-System Connectivity
f s ck checks the general connectivity of the file system. If it finds directories
not linked into the file system, fsck links the directory back into the file
system by placing them in the /lost+found directory.

Uncorrectable File System Corruption

In certain instances , fsck may be unable to check and repair the file system
( for example, if all copies of the superblock are lost ) . The fsdb ( file system
debugger ) command is provided for such situations.

Caution fsdb should be used only by an HP-UX file system expert ,

since it can easily destroy the entire file system. Refer to the
fsdb( lM) entry in the HP- UX Reference for details.

Transferring Files between HP-UX and Other Systems

Not all computers use the HFS File System. Among HP products, for example,
the Series 200 HP-UX 2.x used the Bell System III file system ( BFS ) , and
the Series 500 HP-UX system used Structured Directory Format ( SDF ) . To
accommodate variances, HP-UX supports several utilities and services for
transferring files , including to other vendors' operating systems .
Table 8-6 shows what to use when transferring information between HP-UX
and various systems. In some cases ( such as with networking products ) ,
optional products must be present .

HFS File System 8-55

 Block Count

fsck checks the block count of two types of data blocks:

• Ordinary data blocks containing information stored in a file. f s ck does not
attempt to check the validity of the contents of an ordinary data block.
• Directory data blocks containing directory entries.
Indirect blocks are owned by an inode; thus, inconsistencies in indirect blocks
affect the inode that points to the block. fsck checks indirect blocks for the
following block-count inconsistencies:
• Blocks already claimed by another inode.
• Block numbers outside the range of the file system.
fsck detects and corrects the indirect-block inconsistencies iteratively, by the
same scheme used for direct blocks.
fsck checks each directory data block for inconsistencies involving:
• Directory inode numbers pointing to unallocated inodes. If a directory entry
inode number points to an unallocated inode, fsck removes the directory
entry.
• Directory inode numbers larger than the number of inodes in the file system.
If a directory entry inode number points beyond the end of the inode list ,
fsck removes the directory entry. This occurs if bad data is written into a
directory data block.
• Incorrect directory inode numbers for "." and " . . " (current and parent
directories, respectively) .
The directory inode number entry for "." should be the first entry in the
directory data block. Its value should be equal to the inode number for the
directory data block.

8 The directory inode number entry for " . . " should be the second entry in the
directory data block. Its value should be equal to the inode number for the
parent of the directory entry (or the inode number of the directory data
block if the directory is the root directory) .
If the directory inode numbers for "." and " . . " are incorrect , fsck replaces
them with correct values .

8-54 HFS File System

To resolve duplicate blocks , fsck makes a partial second pass of the inode list
to find the duplicated blocks ' inodes . fsck prompts the operator to clear both
inodes . Often clearing only one inode solves the problem, but the data in the
other inode is suspect .

Bad Blocks

Contained in each inode is a list or pointer to lists of all the blocks claimed
by the inode. fsck checks each block number claimed by an inode for a value
outside the range of the file system (lower than that of the first data block or
greater than the last block in the file system ) . If the block number is outside
this range, the block number is a bad block number.
fsck prompts the operator to clear the inode.
Series 800 systems using LVM have another mechanism for relocating bad
blocks . LYM bad block relocation is discussed in Chapter 9, "Logical Volume
Manager" .

lnode Size

Each inode contains a 64-bit ( eight-byte ) size field indicating the number of
characters in the file associated with the inode. f s ck uses the inode size field
to check for size inconsistencies.
fsck calculates the number of blocks that should be claimed by an inode by
dividing the number of characters in the file by the number of characters per
block and rounding up to get the number of direct blocks. fsck then counts
actual direct and indirect blocks associated with the inode. If the actual
number of blocks does not match the computed number of blocks, f s ck warns
of a possible file-size error. This is only a warning because HP-UX does not fill
in blocks in sparse data files.
A directory inode within the HP-UX file system has the mode word set to 8

directory . The directory size of a file system using 14-character filename

limits must be a multiple of 32 characters, because a directory entry contains
32 bytes of information. The number of blocks actually used for the directory
should match that indicated by the inode size. f s ck reports any directory
misalignment , but cannot correct it.

H F S File System 8·53

 Format and Type

fsck verifies inodes classifications-regular file, directory, block special file,

character special file, network device, FIFO , symbolic link, or continuation
inode. It also examines the inode state, as:
• Unallocated
• Allocated
• Neither unallocated nor allocated
This last state indicates an incorrectly formatted inode. An inode can get into
this state, for example, if bad data is written into the inode list through a
hardware failure. To correct such an ambiguous state, fsck clears the defective
inode.

Link Count

Contained in each inode is a count of the total number of directory entries

linked to the inode. fsck verifies the link count stored in each inode by
traversing the total directory structure ( starting from the root directory ) and
calculating an actual link count for each inode.
If the stored link count is non-zero and the actual link count is zero, no
directory entry appears for the inode; f s ck links the disconnected file to the
/lost+found directory. If the stored and actual link counts are non-zero and
unequal, a directory entry may have been added or removed without the inode
being updated. fsck replaces the stored link count in the inode by the actual
link count .

Duplicate Blocks

Duplicate blocks can occur from using a file system with blocks claimed by
both the free-block list and other parts of the file system.
8
Each inode contains a list ( or for large files, pointers to lists in indirect blocks )
of all blocks containing its file's data. fsck compares each block number
claimed by an inode to a list of allocated blocks. f s ck updates the list of
allocated blocks to include the block number. If a block number is already
claimed by another inode, fsck adds the block number to a list of duplicate
blocks.

8-52 HFS File System

system. If the figures do not agree, fsck replaces the count in the summary
information by the actual free-block count .
If anything is wrong with the free-block maps, fsck rebuilds them, excluding
all blocks in the list of allocated blocks.

lnode Checking

The superblock summary information contains a count of the total number of

free inodes within the file system. fsck compares this count to the number
of free inodes it finds within the file system. If the figures do not agree, fsck
replaces the count in the summary information by the actual free inode count .

lnode Consistency
Individual inodes are less likely than superblock summary information to be
corrupted. However, because of the great number of active inodes, it is possible
that a few inodes might become corrupted.
The inodes list is checked sequentially, from inode 2 (inode 0 marks unused
directory slots and inode 1 is reserved for future use) to the last inode in the
file system.
The inode structure is defined in /usr/ include/sys/ inode . h. There are two
major types of inodes: primary and continuation. Continuation inodes contain
only a mode (which is of type continuation) , a link count , and access control
list (ACL) entries. Continuation inodes exist only if a file has optional ACL
entries associated with it . fsck checks the continuation inode's mode, link
count , and the reference from the primary inode. It does not examine the ACL
information itself. fsck checks each primary inode for inconsistencies in the
following areas:
• Format and type 8
• Link count
• Duplicate blocks
• Bad blocks
• Inode size
• Block count

HFS File System 8-5 1

 Superblock Consistency
The superblock's summary information can become inconsistent because
every change to the file system's blocks or inodes modifies it. Most often, the
superblock and its associated parts become corrupted when the computer
is halted and the last command involving output to the file system is not a
reboot , shutdown, sync , or umount command. fsck checks the superblock for
inconsistencies involving:
• Free block count-this is fairly common
• Free inode count-this is fairly common
• File system size-this rarely happens.
If fsck detects corruption in the static parameters of the primary ( default )
superblock, fsck requests the system administrator to specify the location
of an alternate superblock. The alternate superblock addresses are listed
during file-system creation. An alternate superblock is always found at block
number 16. If this superblock is also corrupted, you must supply the address of
another superblock. If the last time you created a file system was during the
installation, a list of superblock addresses can be found in the / etc/sbtab file.

File System Size

f s ck examines the superblock for inconsistencies involving file system size,

number of inodes, free block count , and the free inode count . The file system
size must be larger than the number of blocks used by the superblock and the
number of blocks used by the list of inodes . The file system size and layout
information are critical pieces of information to the f s ck program. While
there is no way to actually check these sizes, f sck can verify that they are
within reasonable bounds. All other checks of the file system depend on the
correctness of these sizes.
8
Free-Block Checking

f s ck checks that all data associated with files and directories can be found.
The superblock summary information contains a count of the total number of
free blocks within the file system. fsck checks that all the blocks marked as
free are not claimed by any files. When all the blocks have been accounted for,
fsck compares this count to the number of free blocks it finds within the file

8·50 HFS File System

In performing its checks, fsck examines the file system several times, each time
examining different characteristics, including:
• Block and file size
• Path names
• Connectivity ( parent-child relationships )
• Reference count links
• Cylinder groups
fsck checks intrinsically redundant file-system data. The redundant data is
either read from the file system or computed from known values.
The file system should be in a unmounted state when you check it . The root
file system should only be run from init run-level s, the system administrator
run-level. ( Thus, you can check the root file system after performing a system
shutdown. ) Do not run fsck for the root file system when the system is busy.
You can check non-root file systems any time, but be sure they are unmounted.
You can run fsck interactively or non-interactively. When invoked
without options, fsck runs interactively on file systems marked "hfs" in
I etc/ checklist and queries you for a response when it finds an inconsistency.
In non-interactive mode ( typically, in the p or preen mode ) , fsck reports
-

inconsistencies, corrects many problems, but does not remove data. If it

cannot solve a problem, fsck terminates. If this happens, you should run fsck
interactively to fix the problems.

Note When running fsck -p before a backup, if the command

completes successfully, perform your backup . If it aborts with
errors, back up the bad file system, repair it , then back up the
file system again.

Do not issue the reboot command in its default form after fsck has repaired 8
a mounted file system. By default , reboot executes sync on the disks, thus
writing out inconsistent data. If you must reboot , use reboot -n, which does
not issue a sync .
Refer to Solving HP- UX Problems ( Chapters 6 and Appendix A ) for full
discussion of f s ck options and actions. The following subsections describe the
interaction of f s ck on various elements of the file-system.

HFS File System 8-49

 • Never take a mounted file system off-line (for example, by shutting its power
off or by disconnecting it) while it is in use.
Follow proper startup procedures (described in the System Administration
Tasks manual) :
• Always check the file system for inconsistencies . (The fsck command runs
automatically when the system reboots.)
• Always repair inconsistencies, using f s ck. Allowing a corrupted file system
to be further modified in such circumstances can be disastrous.

The /lost+found Directory

Every file system should have a lost+found directory at its root . fsck, the file
system check command (discussed in the next section) , places any problem files
or directories in lost+found. After fsck completes , you should examine each
file in lost+found, to determine its name and location, and attempt to return
it to its rightful place.
lost+found is created by both mkfs and newf s when they create file
systems. However, if your system lacks lost+found, you can rebuild it using
mklost+found(lM). mklost+found creates several empty file slots for f s ck.

Understanding Use of fsck to Detect and Correct

File-System Corruption
The f s ck command is the principal file-system maintenance tool for checking
system consistency and making repairs.
Never run fsck on a mounted fi.le system. However, f s ck should be run
8
regularly to ensure the file system's structural integrity:
• fsck is invoked during system boot-up by the /etc/bcheckrc script run by
init .
• For preventative maintenance, f s ck should be run weekly (before each full
backup) on all file systems , but particularly on file systems that have been
unmounted.
• You should run f s ck any time you suspect problems with the HP-UX file
system. Be sure to unmount the file system first !

8-48 HFS File System

Although asynchronous disk writes increases system performance for most
applications, if a system crashes , file-system data structures are likely to be left
in an inconsistent state. For this reason, we do not recommend that you turn
on f s _async on a production system.
Normally, file-system recovery is performed automatically by fsck in
the reboot process and does not require any intervention by the system
administrator. However, using asynchronous disk writes might require system
administrator intervention in the event of a crash. For further information,
refer to fsck(lM) in the HP- UX Reference.

Minimizing File-System Corruption

Although the HFS file system is very reliable, hardware failures, accidental
power loss, or improper shutdown procedures can cause its corruption.
Problems, such as a bad block on a disk, power loss, or a non-functional
disk controller, can occur and cause the hardware to fail. By following
recommended hardware preventive maintenance procedures and by keeping
regular backups ( as defined in the System Administration Tasks manual ) , you
can avoid most serious problems and be prepared for any that might occur.
As a system administrator, you are responsible for preserving users' data. Since
the file system is the HP-UX data structure that stores the data, it is essential
that you safeguard the file system by performing maintenance tasks ( such as
regular backups ) , following proper startup and shutdown procedures, and by
checking the file system when necessary using the fsck command. (fsck( lM) is
discussed shortly, in this chapter, in Solving HP- UX Problems , and in HP- UX
Reference Manual.)
8

System Shutdown and Startup Guidelines

To ensure file system integrity, always follow proper shutdown and startup
procedures ( described in the System Administration Tasks manual ) :
• Always shut down the system using reboot or shutdown.
• Never physically write-protect a mounted file system, unless it is mounted
read-only.

HFS File System 8·4 7

 Synchronous vs. Asynchronous Disk Writes
When HP-UX writes a file-system data structure to disk synchronously, any
file-system activity must complete to the disk before the program is allowed to
continue; the process does not regain control until completion of the physical
I/O ( regardless of whether the I/O is user data or operating-system data) .
Synchronous writes include some file-system structures and whatever an
application writes with O_SYNC set .
When HP-UX writes to disk asychronously, I/O is scheduled at some later time
and the process regains control immediately, without waiting.
By default, some critical changes to the structure of the file system are posted
to disk synchronously. Synchronous writes ensure file system integrity in case
of system crash, but this kind of disk writing also impedes system performance.
Run-time performance increases significantly ( up to roughly ten percent )
on I/0-intensive applications when all disk writes occur asynchronously;
little effect is seen for compute-bound processes. However, if a system using
asynchronous disk writes crashes , recovery might require system-administrator
intervention using fsck and might also cause user data or directories to
disappear.
As a system administrator, you can specify whether some disk writes are
performed synchronously or asynchronously. The fs_async parameter in the
S800 ( Series 800) or df ile ( Series 300/400/ 700) enables and disables the
feature regarding inodes. ( You cannot modify whether or not other types
of disk writes occur synchronously. They are asynchronous by default and
synchronous if O_SYNC flag is set by the application. )
• On the Series 300/400/800, the fs_ async default value of 0 specifies that the
writes should be performed synchronously. Setting fs_ async to 1 causes
fewer writes to be performed synchronously. Typically, this causes file-system
8 performance to improve.
• On Series 700 systems only, the default value of 0 specifies that writes be
performed asynchronously.
Note too, fs_async , deals with inodes and directories, while O_SYNC deals with
files and data. If a file is opened via O_SYNC , the file continues to be written
synchronously, regardless of what method is specified. O_SYNC also causes
inodes to be updated synchronously.

8-46 HFS File System

I mmediate Reporting
Numerous S C SI disk devices are shipped with a feature calle d immediate
reporting ( disabled by default ) . Immediate reporting speeds status notification;
its implementation is handled by the disk controller and disk device. However,
immediate reporting also has some associated risks.
With immediate reporting, when a device driver sends a write request to a
device, the device accepts the data, places it in its buffer or its cache, and
reports to the SPU that the write completed successfully. Without immediate
reporting, status is not returned until the data goes to the media itself.
In a power ( or other ) failure, data might not have been written successfully
to disk, but in fact , still reside in a buffer. An application, writing to the raw
device or to the files system using O_SYNC , continues processing as though the
data has been written. If data remains in the buffer at the time of a system
failure, the database is left in an inconsistent state.
Under rare circumstances, immediate reporting might also cause delayed errors
or system panics. This can occur in the following scenario: A user has a write
request and the system returns good status immediately. If the next request is
a kernel request and an error occurs ( such as a write failure ) caused by the
user's write request, the error might get associated with the kernel request. If
the kernel request cannot tolerate the error, the kernel might panic.
Immediate reporting can be set or disabled using the scsictl(lM) command.
If it is critical that your system never go down, you might want to disable
immediate reporting. Although S CSI disks available for Series 800 systems can
be set for immediate reporting, the feature poses greater risk of inconsistent
data; the disks are shipped with the feature disabled.

HFS File System 8-45

 Primary Superblock The superblock of a mounted file system is written to
the disk whenever a umount command is issued, or
when a sync command is issued and the file system
has been modified.
!nodes An inode contains information describing the file.
The inode is written to disk after every modification,
unless the fs_async parameter is set in the
configuration ( SBOO or df ile ) file. ( See "Synchronous
vs. Asynchronous Disk Writes" , later in this chapter. )
Data blocks In-core blocks (including directories, indirect blocks,
files, pipes, symbolic links, and FIFOs ) are written
to the file system after being modified and released
by the operating system. Upon release, data blocks
are buffered or queued for eventual writing. Physical
1/0 takes place when the buffer is needed by HP-UX,
when a sync or fsync command is issued, or when
O_SYNC is set for the file. If a file is opened with the
O_SYNC or O_SYNCIO flag set , the writ e system call
does not return until completed.
Cylinder group The cylinder group information is updated whenever a
information sync is executed, or when the system needs a buffer
and the cylinder group is written.

Caution • Always unmount a file system before executing fsck.

• Always reboot the system without syncing ( that is, use
reboot -n) after altering the root device with fsck.
A file system can become inconsistent if you execute fsck on a
mounted file system other than the root file system; you risk
8 missing buffered information not yet written to the file system.
If this information is then flushed from the buffer cache, it
might overwrite corrections that f s ck had made.

8-44 HFS File System

The dynamic buffer cache shrinks by use of vhand, the virtual-memory
subsystem's pageout daemon. ( See "Maintaining Page Availability-vhand and
swapper Daemons" in Chapter 7, "Memory Management ." )
vhand reclaims pages of memory from the buffer cache as well as virtual
memory, by using reference bits, much as it does through the virtual memory
subsystem's regions. Its first ( age ) hand clears the status bits of any buffer
pages not recently accessed. If the status bit remains clear by the time the
second ( steal ) hand traverses it, vhand reclaims ( pages out ) the associated
page.
The dynamic buffer cache provides the operating system flexibility to
accommodate both small application programs that do a lot of 1/0 and large
programs that do little 1/0 but require many pages of memory for data.

How the HFS File System Modifies Files

Every time a file is modified, the HP-UX operating system updates the file
system to ensure its consistency.
When a process updates ( writes to ) the file system, the data being written
is copied into an in-memory buffer cache. The physical disk is updated
asynchronously from the buffer write. The data, along with the inode
information reflecting the change, is written to the disk sometime later, unless
the file was opened in the synchronous mode ( see the description of O_SYNC
and O_SYNCIO in open(2) and fcntl(2) in the HP- UX Reference). The process
continues, even though the data has not yet been written to the disk. If the
system is halted without writing the buffer to disk, the file system on the disk
is left in an inconsistent state. Such inconsistencies are flagged and corrected, if
possible, by the f s ck command at system startup . ( Discussions of f s ck, the 8
file-system check command, appear later in this chapter, in Solving HP- UX
Problems , and in fsck(lM) of the HP- UX Reference Manual.)
The sync command can b e used to force synchronization. However, the syncer
command routinely updates the file system's superblock, inodes , data blocks,
and cylinder group information, as described below. ( For further information,
see sync(lM) and syncer( lM) in HP- UX Reference Manual.)

H F S File System 8-43

 Dynamic Buffer Cache (Series 300, 400, and 700 only)
The buffer cache is implemented on Series 300/400/700 systems to grow and
shrink in size, depending on operating-system and virtual-memory need. Note
that demand for memory is generated not only by the file system, but also by
other objects, including processes, data regions , memory-mapped files.
From a system-administration perspective, using the dynamic buffer cache
is simple: the operating system is shipped with it set up . The nbuf and
bufpages operating-system parameters are not specified in the df ile ; in their
absence, operating system determines how many buffer-cache pages are needed
for optimal system performance. You can choose to configure these parameters,
but if you do, the buffer cache will no longer function dynamically.
Both buffer cache and virtual-memory subsystem access the same body
of RAM in main memory. The dynamic buffer cache is allowed to grow
considerably larger than a statically configured buffer cache, permitting more
data to be held in memory. When the virtual-memory system requires more
memory, the dynamic buffer cache is reduced to yield memory for processes.
The dynamic buffer cache functions like a large memory-mapped file that is
shared among all the processes running on the system. ( Note that there are a
number of subtle interactions between the buffer cache and memory-mapped
files, which can streamline bringing data into the virtual-memory subsystem.
For information on memory-mapped files, see Chapter 7, "Memory
Management" . )
The dynamic buffer cache uses a different algorithm for reusing existing buffer
pages and allo cating more pages from memory. It still reuses pages according
to "least-recently used" (LRU) order. However, instead of maintaining
three free lists (LRU, AGE, and EMPTY ) as in the standard buffer-cache
implementation, the dynamic buffer cache uses two free lists (LRU and
8 EMPTY ) .
The LRU lists buffers in most-recently to least-recently used order. This list
may grow as long as the buffer cache is growing. When a buffer is read for
the first time, its buffer is inserted mid-list , in fairly high priority. If accessed
again, its priority is increased. Other buffers might decrease in priority ( such
as file-system writes to an entire block, which typically do not get referenced
again ) .

8-42 HFS File System

Requests for buffers come from many sources, including file-system reads and
writes, and device driver allocations. If a buffer is requested and not already in
the cache, the operating system obtains the buffer header, allocates memory for
the pages of the buffer, and then gives it to the part of the operating system
making the request .
In the standard (non-dynamic) implementation, the buffer headers are
allocated in a single contiguous block and treated as an array. Inactive buffer
headers are placed on one of three doubly-linked lists-LRU (least recently
used), AGE, and EMPTY:
• Although its name suggests otherwise, the LRU list actually points to
blocks of most frequently accessed data, representing no more than 403 of
total buffers. If data in a buffer is dirty (that is, its contents changed since
accessed from the file system) , its pages must be written to disk before the
pages can be reallocated.
• The AGE list contains buffers accessed less frequently and the overflow of the
LRU list.
• The EMPTY list contains unallocated buffer headers.
If a buffer requires more than one page, HP-UX ensures that the pages are
assigned in consecutive addresses.
As code and data move into the buffer cache, the system copies the information
from the buffer cache into user's main memory. If a user requests information
already in the buffer cache, the information is copied from the cache to user's
main memory, eliminating the I/O operation to bring it in from the file system
disk.
When data is written through the buffer cache, any data in the virtual-memory
system's page cache (in main memory) with the same vnode and block address
is purged. Virtual addresses used by the buffer cache are in kernel space.
When a pagein occurs , both the buffer header (on one of the buffer lists) and 8
associated data in the buffer cache are flushed.

HFS File System 8-4 1

 If you have just created and compiled a program, all transactions occur from
the buffer cache. For an existing program, however, data might exist on both
disk and buffer cache. When a page is faulted in from disk to memory, HP-UX
also ensures that the process executes using the most current copy of the data.
During a file-system write, HP-UX ensures that only the most current copy of
the data, whether in the virtual-memory system's page cache or in the buffer
cache, is written to disk.
The HP-UX file-system buffer cache is currently implemented in two ways:
• On Series 300/400/700 systems, the buffer cache is dynamic; that is, it can
change in size depending on system demand for virtual memory vs. buffer
cache. (This dynamic capability is described in the next section. )
On HP-UX clusters, the dskless_fsbuf s parameter governs the maximum
number of buffers that you can create for incoming cluster traffi c .
dskless_fsbufs is a separate pool of buffers, which are allocated as needed.
(Clusters are not supported for Series 800 systems in the HP-UX 9.0 release. )
• On the Series 800, the number of buffers in the cache is set by two
operating-system parameters in the S800 file-nbuf and bufpages . When
you power up your system, these parameters reserve memory for buffer
headers (nbuf ) and for pages of memory for buffer-cache use (bufpages)
based upon the amount of available RAM. Of the two parameters, bufpages
is more critical, defining the amount of memory in buffer cache, which can
vary depending on block size.
You can use SAM to change the default buffer-cache values and then reboot
to implement the changes. For information on nbuf and bufpages, refer to
the System Administration Tasks manual.
The buffer cache consists of two parts:
• buffer headers, which have pointers to the buffer and describe its contents.
8 • buffer data area, which reside in data blocks ranging in size from DEV _BSIZE
to MAXBSIZE.
Just as a file system must be created in a multiple of DEV_BSIZE, so must a
buffer always be some multiple of DEV _BSIZE. MAX_BSIZE is the largest buffer
size, in bytes. The smallest unit of memory assigned to a buffer is one page.
The data structures used for allocating and managing buffers are defined in the
/usr/ include/sys/buf . h header file.

8-40 HFS File System

The File-System Buffer Cache
The file-system buffer cache manages data fl.ow between main memory and
secondary memory ( principally disks ) , by temporarily holding ( buffering )
information about data being transferred to and from disk. The buffer cache
speeds data transfer from the file system to main memory; once buffered, data
is accessed by a process's executing space in main memory much faster than
from the file system on disk. The buffer cache is used for all file system I/O
operations, plus all other block I/O operations in the system ( for example,
mount , inode reading, LVM management , and some device drivers ) .

The role of the buffer cache is illustrated in Figure 8-9. When you execute a
program, the shell passes the file path name to exec, finds the file on disk,
and reads the a . out header into the buffer cache. The a . out header contains
preliminary information about the executable, including the size of the text
and the uninitialized data ( bss areas ) .

Secondary Storage M ain Memory

B uffe r Cache
.- - - - - - - ,

1.;-1
I * a. out header I
P rogr�m L - - - - - - -'

File ,
*
_ _
Program
Executable

LG20021 1 _020

Figure 8-9. Buffer Cache Holds the a.out Header of Executing Programs

8
As the code executes , the virtual-memory system reads the pages of data
directly from the disk into memory. ( Some additional pages might also be read
in, based on the probability that they will be needed ) . The file's a . out , which
is only needed to begin the "demand-paging" , might ( or might not ) remain in
the buffer cache throughout the process execution , depending on whether its
buffer is needed.

HFS File System 8-39

 the new block. If the size of the old and new data is less than a full block,
a block with enough contiguous fragments (or a full block ) is located and
allocated.
When a block or fragment has been located, the address is recorded in the
inode table and the free block bit map is updated.

Allocation Policies
Allocation is performed globally to place new directories and files and locally to
place data in blocks.
A global decision determines which cylinder group will contain a given file or
directory. HP-UX attempts to put all files from a single directory in the same
cylinder group . Newly created directories are put in the cylinder group with
the greatest number of free inodes and the smallest number of directories .
Global policy specifies that once the file size reaches maxbpg (maxbpg is defined
via the tunef s command ) , HP-UX allocates blocks from another cylinder
group . This helps to enforce grouping of all files within one directory into a
single cylinder group by spreading the less common larger files over several
cylinder groups .
Global allo cation routines call local allocation routines with requests for
specific data blocks. Blocks are allocated by the following priorities:
• Allocate block requested.
• Allocate a block on the same cylinder that is rotationally closest to the
requested block.
• Allocate any block within the same cylinder group .
• Use a quadratic hash to find a new cylinder group; allocate a block
somewhere in the new cylinder group .
• Use sequential search to find an available block.
8
Speed in allocating blocks is the most important characteristic of this strategy.
For this reason, the percentage of free space must be maintained.

8-38 H F S File System

How Disk Space is Al located
Free space availability is determined from a bit map associated with each
cylinder group . The bit map contains one bit for each fragment . To determine
if a block is available, the system examines consecutive fragments. A piece of
the bit map from a file system using 1024-byte fragments and 8192-byte blocks
is shown in Table 8-5.

Table 8-5. Sample Free Block Bitmap i n a n 8KB/ 1 KB File System

bit map 00000000 000000 1 1 1 1 1 1 1 100 11111111

Fragment numbers 0-7 8- 1 5 16-23 24-3 1
Block numbers 0 1 2 3

Fragment numbers 14-21 and 24-31 in this example are free, indicated by 1 s in
the bit map . Fragment numbers 0- 13 and 22-23 are allocated, as indicated by
Os in the bit map. Fragments in adjacent blocks cannot be used to create a
full block; only eight contiguous fragments starting on a block boundary can
be used to allocate a full block. Fragments 24-31 can be coalesced to form a
full block, but not fragments 14-21. Also, if a partial block is allocated, the
fragments must be consecutive and not cross a block boundary. For example,
if three fragments are needed, fragments 16- 18 can be allocated , but not
fragments 14- 16.
Every time data is written to an existing file, the system checks to see if file
size must increase. If so, one of three conditions exists:
• Sufficient space exists in the existing block or fragment ; the new data is
written into the already allocated space.
• The file contains only whole blocks; the last block contains insufficient space 8
to hold additional data. If more than a full block of data must be written, a
new block is allocated and written. This process is repeated until less than a
full block of new data is needed. At that point , a block containing enough
contiguous fragments is located and the new data is written there.
• The file contains fragments, but not enough to hold the new data. If the size
of the existing data in fragments plus the new data exceeds the size of a full
block, a new block is allocated. Both the old and new data are written to

HFS File System 8-37

 Minimum Free Space
To ensure the availability of free blocks near one another, a certain percentage
of free space must always be available in the :file system. This minimum free
space percentage is specified at :file system creation using the -m option of the
newfs command or the minfree argument of the mkfs command. The default
is 10 percent . Values lower than 10 percent may severely degrade system
performance, by causing the :file system to search harder for free space.
The percent of free space can be changed at any time using the -m option of
the tunefs command. The reserved free space is inaccessible to the normal
user; once this threshold is met, only the superuser can continue to allocate
blocks. When the percentage of free space drops below the threshold, system
throughput ( to and from newly created :files ) drops because the :file system can
no longer localize the blocks for a :file. Accessing a :file is quicker if the whole
:file is grouped together ( localized ) .

8-36 HFS File System

 I node
•
•
•

File 20 K
Size
•
•
•

1. I I. I I I

I
I I I I I I I I I I I 1 I I I I I I

8 15 24 31 40 43 46 48
8 ,......I

2 24

3 43
Direct
Blocks
4 0
•
•
•

12 0
•
•
•

Figure 8-8. lnode Addressing Example

All indirect blocks are referenced only as full blocks ; no pieces of the file are
addressed at the fragment level beyond the 12 direct pointers.
When a block or fragment is needed, the disk is searched for free blocks .
Ideally, free blocks should be found throughout the disk, for searches to locate
a free block close to related blocks. When the file system is full, there are long
linear searches to find the block, and when a block is allocated, it is likely to be
placed far from the previous block of the file, resulting in long seeks and slow 8
performance.

HFS File System 8-35

 Inode pointers hold the address of a fragment . The address can be interpreted
as referencing an entire block or one or more fragments , depending on the
number of bytes stored at the address.
All blocks but the last have a full block of data allocated to them. If the
amount of data in the last block is less than the :file system block size, only
the number of consecutive fragments needed to actually store the actual data
are allocated. For example, in an 8-KB/ 1-KB file system, a 15-KB file is
stored as 2 8-KB blocks and 3 consecutive 1-KB fragments. ( The latter might
also be referred to as a 3-KB fragment. ) This allocation scheme provides
the performance advantage of large blocks with the space savings of small
fragments.
Figure 8-8 shows an example of a 20-KB :file stored in 8-KB blocks with 1-KB
fragments. The number of blocks needed is 20 + 8 (file size + block size ) : 2
full blocks with a remainder of 4 fragments. Therefore, the :first and second
pointers point to full blocks, but the third pointer points to the remaining 4
fragments.

8-34 HFS File System

 1 st level 2nd level File's Contents Here
mode & file type
indirection indirection

# links to file

owner/ g roup info

file size in bytes

time sta m p s

1 e-11----'
direct 2 -ti------_.
blocks

single indirect

double ind irect

tri ple indirect •

• 1K p o in te rs if F.S. Block Size 4 Kbytes ;

2K poin ters if F.S. Block Size 8 Kbytes 8

Figure 8-7. Mapping from lnode to File Data Blocks

HFS File System 8 -3 3

 How a File is Accessed from l node to Data Blocks
As described earlier, the inode in the cylinder group contains pointers to the
locations of a file's actual data. Depending on the size of a file, its data might
be reached through pointers to direct blocks or indirect blocks , which are
pointers to a block containing more pointers to the data. HP-UX allows for up
to triple indirect pointing for enormous files.
Figure 8- 7 shows the mapping from an inode to a file's data blocks .
The first 12 pointers in an inode point directly to the first 12 blocks or
fragments containing the file's data. If the file is larger than 12 blocks ( greater
than 12 x fs_ bsize ) , indirect reference is made to the file's data. A group of
indirect pointers is contained in one data block. Each pointer is 4 bytes long,
so there can be either 1024 pointers (4096/4) or 2048 pointers (8192/4) in each
block of indirect pointers.
The thirteenth block address in the inode points to a block containing 1024
or 2048 additional pointers to data blocks ( from now on, the number of
indirect pointers in a block will be called num_ip). Thus , the thirteenth
( single indirect ) block address handles files up to 4,243,456 bytes in a 4-KB
block file system or 16,875,520 bytes in an 8-KB block file system (fs_bsize
X (12 + num_ip)) . If the file is larger than this, the fourteenth inode block
address points to num_ip indirect blocks, each of which contains pointers to an
additional num_ip actual data blocks.
If the file cannot be contained in this space, the :fifteenth inode block address
points to num_ip double-indirect blocks. With the :fifteenth ( triple-indirect )
block address, the size of a file is limited to fs_ bsize x ( 12 + num_ip +
num_ip2 + num_ ip3).

8-32 HFS File System

Block size is set using the mkfs or newf s command, when you construct a file
system. See mkfs(lM) and newfs(lM) in the HP- UX Reference for details.
Larger block sizes are faster for sequential access to the file system, while
smaller block sizes use space more efficiently and are better for random 1/0 .
Having a large block size has both benefits and costs . For big files, a large
block size significantly reduces the number of disk accesses, thereby increasing
file system throughput . The problem is that most HP-UX files are small; thus ,
using a large block size for small files might waste space.
In the fs . h header file, the size of blocks is referred to as fs_bsize , depending
upon what block size your file system uses.
To minimize wasted space, a block can be divided into either 1 KB, 2 KB , or 4
KB fragments.
Fragment size is bounded on the lower end by 1024 and on the upper end by
the block size. Fragment size is specified at file system creation, can never be
less than one-eighth of a block, and must be one, two, four, or eight times 1024.

HFS File System 8·3 1

 If the file described by the inode is not a regular file, some of the inode fields
differ as follows:
• FIFO and pipes
The space reserved for indirect block pointers contains information about the
current state of a FIFO or pipe.
• Character or block device files
The first direct block address is actually the major and minor number of the
device. The rest of the direct block addresses are 0.
• Directories
The pointers point to regular file system data blocks, but the blocks
contain specifically formatted data, described in the header file
/usr/ include/sys/dir . h.
When you create a file system (using newf s or mkf s ) , the system creates
inodes. The number of created inodes limits the number of files that you can
have in a file system. Each time you create a file, an inode is allocated for that
file. Both commands default to 2048 bytes per inode, meaning the system
assumes that the average size of your files will be 2048 bytes.
Although uncommon, an inode error message, inode : t able is full, might
require changing the size of the inode table. This message refers to the kernel's
in-core inode parameter. A configurable parameter, ninode , defines the
maximum number of open, in-core inodes.
You can use SAM to change these configurable parameters , which are
documented in System Administration Tasks manual for your system.

Data Blocks

Disk space before or after the superblock, cylinder group information, and
inode table is filled with data blocks. (The specific locations of data on each
8 platter is different , due to the cylinder-block offset .) The blocks are used to
store data for regular files, directories, and symbolic links.
HP-UX provides support for file systems in several block sizes:
• On the Series 300/400/800, the file system can use blocks of either 4 KB or 8
KB .
• On the Series 700, the file system can use block of 4 KB , 8 KB , 16 KB , 32
KB , or 64 KB .

8-30 HFS File System

lnodes

Besides maintaining information about the state of the file system, the cylinder
group also maintains key information about the inodes of the file system-the
system's index to the actual files of data. !nodes contain the locations of the
actual file data.
The cylinder group maintains an inode table, which provides summary
information about each file in the cylinder group ( see Figure 8-7). In addition,
the "disk inodes" appear in an expanded version ( "in-core inodes" ) in memory
for inodes currently ( or recently) used.
A disk inode includes the following information:
• mode and file type
• number of links to the file
• owner and group information
• file size in bytes
• time stamps
• pointers to the file's actual blocks of data on disk
When a file is read into memory, its "in-core inode" contains the following
additional fields:
• status of the in-core inode, including if the inode is locked, if a process is
waiting for the inode, if the disk inode now differs from the in-core copy due
to file modification, if the file is a mount point
• numeric address of the file system containing the file
• inode array number by which the kernel identifies the disk inode
• pointers to other in-core inodes linked on buffer hash list and free list
The header file, /usr/ include/sys/ inode . h, defines the in-core inode;
the header file, /usr/ include/sys/ ino . h, defines the disk inode.
When the operating system accesses a file, it finds the file by means of the 8
inode's pointers to the file's blocks of data. The scheme for finding the file data
is discussed later in this chapter.
A static number of inodes is allocated for each cylinder group when the file
system is created. HFS uses a default that provides sufficient inodes per
cylinder group for average usage.

HFS File System 8-29

 Table 8-4. Cylinder Group Layout

Data Structure Size

Boot block 1 8 KB
Primary superblock 1 8 KB
Redundant superblock 8 KB
Cylinder group information 1 block (4 or 8 KB)
lnode table varies2
Data blocks 0 or more blocks3

1 First cylinder group only. The beginning of all

subsequent cylinder groups might be filled by data
blocks, depending on offset .
2 See "!nodes" section.
3 Due to offset; see "Data Blocks" section.

A redundant copy of the superblock is located in each cylinder group. This

ensures that if any single track, cylinder, or platter is damaged, the file system
itself can be repaired by executing fsck and specifying an alternate superblock.
Further, each successive cylinder group is laid out offset by one track in
relation to the previous cylinder group , so that the redundant copies of the
superblock spiral down the platters.
The cylinder group information contains the dynamic parameters of the
cylinder group:
• Number of inodes and data blocks in the cylinder group
• Pointers to the last used block, fragment , and inode
• Number of available fragments
8
• Used inode map
• Free block map.
The cylinder group information data structure's size is one block ( a block can
be defined when running newfs as either 4 KB or 8 KB ) . The layout of the
cylinder group information is defined in /usr/include/sys/fs . h.

8-28 HFS File System

 LG20021 1 _024

Figure 8-6. Track, Cylinder, and Cylinder Group on a Disk

The cylinder group controls all access to a file and its associated data. Each
cylinder group contains a copy of the superblock, a cylinder group information
structure, an inode table, and data blocks (see Table 8-4) .

HFS File System 8-27

 The Cylinder Group
The cylinder group is the term used to describe a further internal organization
of disk sections.
Here is the principle, illustrated in Figure 8-6-A cylinder is a collection
of tracks located the same distance from the edge of a disk platter. Since
all the tracks in a cylinder are accessed by the read/write heads of the disk
drive simultaneously, the blocks of space on each track can be accessible with
minimum rotational latency; that is, requiring no seek time.
For performance reasons, small groups of adjacent cylinders ( sixteen by default ,
see newfs(lM)) are grouped together as cylinder groups. Each cylinder group
has its own set of inodes and local mappings of free space in the group . This
internal organization results in both bringing to closer proximity file-system
inodes and their associated data without long seeks and dispersing data and
inodes across cylinders. Minimum time is lost seeking file data within a
cylinder group .

8·26 HFS File System

The Superblock
The superblock is a contiguous, 8-KB block of disk space near the beginning
of the file system's disk section. The superblock contains a record of the static
information about the state of the file system at the time of its creation ( or
extension, if using LVM):
• file system size
• number of inodes it can store
• locations of free space on the file system
• number of cylinder groups
• location of superblocks, cylinder groups, inode blocks, and data blocks
• size and number of blocks and fragments.
The primary superblock also keeps track of file system update information in
its summary information area.
HP-UX uses information in the superblock for various file system maintenance
procedures-for example, when you mount a file system or perform a file
system check by executing f s ck.
Because the superblock is so important , HP-UX always keeps redundant copies
on disk in each cylinder group. One copy is brought into main memory when
you boot up. A primary superblock is at the beginning of the file system, and
each cylinder group has a copy of the superblock. This redundancy further
ensures the integrity of file system data.
The non-redundant superblocks on the disk are updated whenever the sync
command is executed and when a file system is unmounted ( see sync(lM) in
the HP- UX Reference).
Record of all superblock locations can be found i n two files-/etc/sbtab for
Series 300/400/700 and / etc/super_blocks for the Series 800.
8

H F S File System 8-25

 File System Layout
With the exception of disk drives used for raw data, every disk drive contains
some file systems. All HFS file systems are laid out in a common format , with
the following structures:
• Primary superblock
• Multiple cylinder groups
The many data structures governing the superblock and cylinder group
are defined in several header files, particularly /usr/ include/sys/fs h
. .

Superblock headers, defined in fs . h , also include absolute disk addresses for

the first boot block and definitions of numerous file system attributes, including
cylinder-group characteristics ( such as rotational positions, number of inodes
per group , number fragments per block ) , file length, and mirror states of root
and primary swap .
A description of the file-system format is found on fs( 4) of the HP- UX
Reference Manual.

8-24 HFS File System

Although a single disk can have multiple swap sections , it is not recommended
because performance will degrade as the system attempts to do interleaved
writes to swap areas on separate areas of the disk. Instead, configure multiple
swap areas on separate disks . ( For discussion of interleaving, see Chapter 7,
"Memory Management" . )
On Series 800 systems using LYM , the primary swap area resides in the
root volume group in a designated logical volume. As with a disk-section
configuration, you can set up multiple swap areas in logical volumes that are on
separate disks ( physical volumes ) .
You can list all swap areas on your system using the swapinfo( lM) command;
see the HP- UX Reference Manual for complete specifications.
Procedures for managing primary swap space ( as well as secondary swap ) are
found in two chapters of the System Administration Tasks manual: Chapter 6,
"Managing Swap Space" and "Managing Logical Volumes."

H F S File System 8-23

 Series 700 Boot Area Implementation

Whereas the Series 300/400 has three locations hard-coded into the system,
the Series 700 has an additional loader, which understands the layout of the
file system. The Series 700 boot area has pointers to the actual bootstrap
programs.
The lifls command on a Series 700 reports presence of FS, SWAP , ISL , AUTO ,
HPUX, IOMAP , EST, and PAD files. Its reportage of FS and SWAP indicates that
Series 700 LIF has knowledge of the entire disk, including the file system and
swap .
When the system is booted, the loader goes off and can find /hpux at a default
or designated location, using the boot console user interface.
For more information, see the owner's guide for the Series 700 systems or
hpux_ 700 ( 1M ) in the HP- UX Reference.
Series 800 Boot Area Implementation

On Series 800, the LIF header contains ISL, HPUX, AUTO, RDB , and IOMAP files.
ISL uses the AUTO file to locate the HP-UX kernel.

Primary Swap Area

The primary swap area is the first swap device specified in the S800 or df ile ;
i t i s also specified i n /etc/checklist . The primary swap area i s a contiguous
area of the root disk used by the virtual memory system ( see Chapter 7,
"Memory Management" ) to temporarily store a process image. Until /etc/re
executes swapon, primary swap is your only swap device.
As device swap space ( in comparison to file-system swap ) , primary swap is fast ;
the system accesses it directly, without going through a file system.
8
On Series 300/400/700 systems, the primary swap area occupies blocks after
a file system area or an entire disk dedicated as a swap disk. If you have
multiple disks, each one can contain its own swap area, but there is still only
one primary swap area on the entire system.
On Series 800 systems using traditional disk sections , primary swap space
occupies its own section, separate from the file system and boot area sections.
Like Series 300/400, there can be a boot area section on each physical disk.

8-22 HFS File System

 Table 8-3. Series 3 00/400 Boot Area Layout

Subarea Size
LIF volume header 256 bytes
( unused ) 256 bytes
Volume directory information 256 bytes
Bootstrap program 7.25 KB

Unless you create your file system using the -n option to newf s , the boot area
will be present on your disk. If you create your file system using mkf s the boot
area will not be present by default . In this case, if you will use the file system
on this disk to boot your system, you must explicitly put the boot area on your
disk using the following command (replace OsO for your disk's actual device file
name) :
d d if=/etc/boot of=/dev/dsk/OsO count= 1 bs=Sk
Each boot disk must have a volume header in the boot area to identify the
volume format (in this case, LIF). The volume header is checked by the boot
ROM in its examination of bootable mass storage media when the computer is
powered up.
The volume directory information contains SYSHPUX, SYSB CKUP, and
SYSDEBUG. All three are object files .
SYSHPUX corresponds to the file /hp-ux, which is your kernel. SYSB CKUP
corresponds to /SYSBCKUP , which is used as a backup kernel. SYSDEBUG
corresponds to the file /SYSDEBUG. This file is used only when writing device
drivers ; its use is described in Series 300 HP- UX Driver Development Guide .
8
On Series 300/400, the last 7.25 KB on the boot area contain the secondary
loader. The boot ROM loads and passes control to the secondary loader which
in turn loads and passes control to the file /hp-ux (or the backup kernel if you
are using SYSB CKUP ) . /hp-ux (or the backup kernel) then completes the task
of bringing up HP-UX.

HFS File System 8-2 1

 Boot Area
The boot area is the portion of the disk that holds the code used to bring
the system into an operational state. The boot code initializes and tests the
hardware, then loads into memory a secondary loader. The secondary loader
is the program that loads /hp-ux ( the operating system ) into memory to
enable you to use your system. ( For detailed information about boot code, see
Chapter 2 , "System Startup ." )
The boot area is reserved on the mass storage medium ( usually a disk ) during
the installation process . Information in the boot area is used only if the disk is
used for booting ( boot disk ) , but the space can be reserved on all disks.
On Series 300/400/700, the boot area precedes the file system on the disk.
On Series 800 systems using traditional disk sections, the boot area must reside
in its own disk section ( section 6) distinct from the file system and swap area
sections .
Typically, section 6 is used since it is one of the smallest sections, yet large
enough to contain the boot area. On Series 800 systems using the Logi cal
Volume Manager ( LVM ) , section 6 might not be large enough, so section 2 is
always used.
Using LVM , the boot data is contained in a Boot Data Reserved Area, which is
created using the pvcreate -B command. ( For detailed information on LVM
and the Boot Data Reserved Area, see Chapter 9 , "Logical Volume Manager
( LVM ) ," of this manual. For procedures on using LVM , see Chapter 7 of
System Administration Tasks manual. )
Although the disk layouts for the various HP-UX platforms differ, all systems
( Series 300/400/700/800) use a small file system for the initial system booting,
written in Logical Interchange Format ( LIF ) . ( LIF is described in li/(4) in
8 the HP- UX Reference. The manual page also contains pointers to the LIF
utilities. )

Series 3 00/400 Boot Area Implementation

On the Series 300 / 400 only, the boot area contains a volume header, volume
directory information, and a small secondary loader used when the system is
loaded. This area is reserved exclusively for use by the boot ROM. Table 8-3
shows the layout of this area, and the size of its components.

8-20 HFS File System

Disk Usage Command-du
The du command reports disk usage in 512-byte blocks for all files or
directories specified; if none is specified, du reports on the current directory. Its
report traverses the file tree recursively.
Here is sample output using du on one of the file systems listed in the previous
example:

Y. du /usr/contrib
16 /usr/contrib/lost+found
448 /usr/contrib/bin+/HP-PA/respond+
2832 /usr/contrib/bin+/HP-PA
3384 /usr/contrib/bin+/HP-MC68020
6218 /usr/contrib/bin+
2 /usr/contrib/etc
2 /usr/contrib/games
2 /usr/contrib/ include

486 /usr/contrib/system
8 /usr/contrib/xseethru
89 124 /usr/contrib

The final number reported is the total of 512-byte blocks for the /usr/contrib
file system, and therefore the number is twice as large as that reported by bdf
in 1024 -byte blocks. 8

Note If it encounters a protected directory (that is, one whose

file permissions are set to prevent access) , du cannot report
the number of blocks contained in that directory or its
subdirectories.

HFS File System 8- 1 9

Table 9-2 shows which type of device file to use for common LVM tasks.

Table 9-2. When to Use Block vs. Raw Special Files for LVM

Special File Type

LVM Task Block Raw
Creating a physical volume x
( SAM or pvcreat e )
Adding a physical volume to a volume group x

Creating a file system in a logical volume x

( SAM or newfs )
Adding or extending disk space to a logical volume x
( SAM or ext endfs )
Mounting or unmounting a file system x

Accessing raw data in a logical volume x

Logical Volume Manager 9-9

 Major and Minor Numbers of LVM Data Structures
The LVM device drivers find data via the major and minor numbers of the file
system data structures . The device file for an LVM disk is shown in the long
listing in Figure 9-2:

Y. 11 /dev/dsk/c3d0s2
brw-r- - - - - 1 root sys 7 Ox000302 Dec 20 15 : 38 /dev/dsk/c3d0s2

Figure 9-2. Device File for an LYM Disk (physical volume)

The 7 represents the major number (index to the device driver) , 3 is the LU
number (physical device) , and 2 is analogous to the section number of a Series
800 disk.
Now compare the minor numbers (shaded) for a volume group in Figure 9-3

Y. 11 /dev/vg3
crw-rw-rw- 1 root other 64 Jan 23 14 : 35 group
brw-r- ---- 1 root other 64 Jan 24 17 : 02 lvo1 1
crw-r----- 1 root other 64 Jan 24 17 : 02 rlvo11
brw-r- ---- 1 root other 64 3 1 1 : 53 lvo12
crw-r----- 1 root other 64 Feb 3 1 1 : 53 rlvo12
brw-r----- 1 root other 64 Mar 6 12 : 0 1 sales
crw-r----- 1 root other 64 Mar 6 12 : 0 1 rs ales

Figure 9 - 3 . Volume Group, Showing Minor Numbers (shaded)

with the overall hexadecimal format by which they were created, in Figure 9-4:

9- 1 0 Logical Volume Manager

 31 23 15 7 0
+ - - - - - - - - - - - - +- - - - - - - - - + - - - - - - - - + - - - - - - - - - +

l maj or number l VG number l reserved l LV number !

+------------+- - - - - - - --+--------+---------+

l < - - - - - - -minor number-- - - - - - > I

Figure 9-4. Hexadecimal format of LYM Major and M inor Numbers

The logical volume (LV) number is encoded into bits 0- 7; the volume group
number (03 ) , comparable to the LU number of a non-LVM device file but in a
different position, is encoded into bits 16-23. The major number is encoded
into bits 24-31. Note too, the block and raw special files created for the logical
volume /dev/vgOO/sales .
Within each volume group directory is a special file, group , with the volume
group major number 64, logical volume number O , and volume group number.
Volume group numbering begins with zero (vgOO), while logical volumes begin
with one (lvo1 1 ) . This is because the logical volume number corresponds to
the minor number and the volume group 's group file is assigned minor number
0.
HP-UX accommodates u p t o 256 volume groups and 255 logical volumes per
volume group .

Logical Volume Manager 9-1 1

 Understanding Volume Groups
If you install LVM on your root file system, you have /dev/vgOO , your first
LVM volume group . ( If you have not installed LVM on your root file system,
you must use SAM or manually create /dev/vgOO .) You can then apportion
disk space from the volume group into logical volumes.
As we have seen , volume groups are visible in HP-UX as device files in the
I dev directory. The directory includes device files for the logical volumes
belonging to the volume group . The volume group directory also contains a
group file, which provides LVM data structures with information about the
entire volume group . ( These data structures are described in "Characteristics
and Layout of LVM Disks" , later in this chapter. )
To see the contents of a volume group , run the vgdisplay command. See
vgdisplay( lM) in the HP- UX Reference Manual for full explanation of the
command and its options. ( For a very graphic representation of a sample LVM
configuration, showing volume groups organized functionally , see the LVM
chapter of Solving HP- UX Problems .)
Figure 9-5 shows verbose output of vgdisplay on a system comprised of one
volume group, one LVM disk ( physical volume ) , and two logical volumes. You
can also see that volume group /dev/vgOO consists of one physical volume
(PV) out of a maximum 32, two logical volumes (LV) out of a maximum 255,
and that each physical extent (PE) is four megabytes . Of a total 632 physical
extents available, 170 have been allocated, leaving 462 free. Status information
reveals that the logical volume contents are synchronized, or up-to-date.

HP-18 Limitations
HP-IB disks must be used in their own volume group and cannot be combined
in volume groups with HP-FL or SCSI disks . This is because HP-IB disks
can handle only limited LVM capabilities: HP-IB does not support bad block
relocation or disk mirroring.

9- 1 2 Logical Volume Manager

Y. vgdisplay -v
--- Volume groups
VG Name /dev/vgOO
VG Status available
Max LV 255
Cur LV 2
Open LV 2
Max PV 32
Cur PV 1
Act PV 1
PE Size (MB) 4
Max PE per PV 1016
VGDA 2
Total PE 632
Alloc PE 170
Free PE 462
Total PVG 0
- - - Logical volumes
LV Name /dev/vg00/lvo l 1
LV St atus available/syncd
Current LE 10
Allocat ed PE 10
Used PV 1

LV Name /dev/vg00/lvol2
LV St atus available/ syncd
Current LE 20
Allocated PE 20
Used PV 1
--- Physical volumes
PV Name /dev/dsk/c2d0s2
PV Status available
Total PE 632
Free PE 462
9
Figure 9-5. Sample Verbose Output of vgdisplay

Logical Volume Manager 9- 1 3

 Consider newfs Limitations when Organizing Volume Groups
Best system performance is achieved if you organize your volume groups by
identical disk type. This means that if you create a file system that spans LYM
disks, be sure that the logical volume in which the file system resides spans
identical disk types .
This guideline derives from the behavior of newfs(lM): when you run the
newfs command to construct a file system, you can specify only one disk type,
which newfs then uses to lay out the disk. If more than one disk type is used
for the file system, the logical volumes on the specified disk type will perform
well, but the unspecified disk type will undoubtedly be less efficient . Thus,
try to maintain consistency of disk type among physical volumes used for any
logical volume that holds a file system.
You might already know the disk type to which your file system is being
mounted. However, on a system set up with LVM, the lvdisplay -v command
is the most accurate starting point for finding the correct disk type.
Another consideration when grouping disk drives is their relative size: The size
of disk drives within a volume group should be similar (within a factor of two
between largest and smallest) to avoid wastiiig space on LVM data structures.

9-14 Logical Volume Manager

Consult /etc/diskinfo for Disk Attributes
When setting up logical volumes, you might find the information presented by
I etc/ diskinfo helpful in determining how to allocate your file systems.
The I etc/ diskinfo command describes disks attributes such as number of
bytes per sector, device type, and timing information. You can also use it for
more information about a device file listed in /etc/checklist . The following
example shows relevant output from /etc/diskinfo for a SCSI disk:
# / etc/diskinfo -v /dev/rdsk/c3d0s2
SCSI des cribe of /dev/rdsk/ c3d0s2 :
vendor : HP
product id : 22 13A
type : direct access
s ize : 648 192 Kbytes
bytes per sector : 512
blocks per disk : 12965 1 1

Logical Volume Manager 9-15

 The Root Volume Group
• In a traditional configuration, the root disk contains all the attributes
required for initial booting, plus system files, dump , and primary swap-all
on one disk.
• In LVM , the concept of a single root disk is replaced by a root volume group .
The root volume group contains all the same elements as a root disk, and I ,
swap , and /usr fi t on a single 300-MB disk.
The root logical volumes can be dispersed over more than one LVM disk, as
shown in the lower example in Figure 9-6. In practice, however, your root
volume group might be nearly identical to a traditional root disk. By default ,
the HP-UX installations programs automatically fill one disk at a time.
• You can mirror root logi cal volumes, using an optional product ,
MirrorDisk/UX.
Figure 9-6 shows a root volume group whose logical extents are mapped to
either one or two LVM disks.

9- 1 6 Logical Volume Manager

 logical extents LVM disks

root LE

swap LE

root volume group

usr LE

/dev/dsk/c0d0S2

root LE

swap LE

usr LE

/dev/dsk/c1 d0S2

LG20021 1 _028

Figure 9-6. R oot Volume Group

When you install the system, having decided to implement LVM, the migration
and installation processes set various options ensuring that your root volume
group is properly configured. Later, if you choose to establish another root
volume group, you might want to refer to the guidelines found in "Guidelines
for Configuring the Root Volume Group" , later in this text .
9
Underlying the root volume group is an LVM disk data structure different from
that of non-bootable LVM disks. Bootable LVM disks have sectors reserved
for a boot data reserved area ( BDRA ) and a LIF volume containing boot

Logical Volume Manager 9- 1 7

 programs. For an LVM disk to be bootable, it must be initialized using the
pvcreate(lM ) command with the -B option (see "Managing Logical Volumes"
in System Administration Tasks manual for the procedure) . The boot data
reserved area and characteristics common to both bootable and standard
physical volumes are discussed later in this chapter, in "Characteristics and
Layout of LVM Disks" .

Contiguous vs. Non-Contiguous Allocation of LYM Disk Space

Just as traditional disk space is allocated contiguously if possible, and then by
fragments, for efficient use of disk space, so LVM LVM allocates disk space
with regard to contiguity.
As illustrated in Figure 9-7,
• Contiguous disk space allocation means that the logical extents (LE) of a
logical volume must be mapped to adjacent physical extents (PE) on an LVM
disk (example a) ) .
• Non-contiguous disk space allocation means that the logical extents of a
logical volume need not be mapped to adjacent physical extents (example
b)).

9- 1 8 Logical Volume Manager

 LE

a)

LE

b)

LG20021 1 _029

Logical Volume Manager 9- 1 9

 Disk space used by the root volume group must be contiguous. That is,
physical extents mapping to the logical volumes of the root file system, primary
swap , and dump must be contiguous . The root volume group 's contiguous
extents also adhere to the following requirements:
• Physical extents must be allocated in ascending order.
• No gap may exist between physical extents.
• When mirrored, all physical extents of a mirrored copy must reside on the
same LVM disk.
Contiguous allocation is less flexible than non-contiguous allocation, and
therefore uses disk space less economically. Non-contiguous mapping can result
in a logical volume's physical extents being dispersed onto more than one LVM
disk, since logical volumes can span multiple disks.

9-20 Logical Volume Manager

Understanding Logical Vol umes
Logical volumes are allotments of disk space within a volume group . Like disk
sections or partitions , logical volumes hold file systems, swap , or raw data;
unlike disk sections, you can set the capacity of logical volumes according to
your need.
Logical volumes consist of logical extents, which map to the physical extents of
LVM disks. This mapping gives logical volumes great flexibility:
• to span more than one LVM disk.
• to be larger than a single LVM disk
• to be reduced or expanded to meet changing file system needs. ( The file
system must be unmounted to reduce or expand a logical volume. )
• to be mirrored, using an optional product , HP MirrorDisk/UX.
Each logical volume has a block special device file and raw special device file.
These device files are listed in the subdirectory of the volume group to which
the logical volumes belong.
To see a logical volume, type the 1 vdisplay command, giving as the argument
the pathname of the logical volume:

Y. lvdisplay /dev/vg00/lvol1
- - - Logical volumes
LV Name /dev/vg00/lvo1 1
VG Name /dev/vgOO
LV Permis s ion read/write
LV Status available/syncd
Mirror copies 0
Cons istency Recovery MWC
Schedule parallel
Current LE 10
Allocated PE 10
Bad block on
Allocation strict
9
Figure 9-8. Sample Output of 1 vdisplay

Logical Volume Manager 9-2 1

 Much of the information given-LV status, Mirror copies , Consistency
Recovery, Schedule, bad block, and Allocation-pertain to mirroring
capabilities provided by the optional product, HP MirrorDisk/UX. Mirroring is
described later in this chapter.
Verbose output of the lvdisplay command shows the mapping of physical
extents (PE) and logical extents (LE) of the logical volume on the LVM disk.
This information can be useful when creating a file system.

--- Distribut ion of logical volume - - -

P V Name LE on PV PE on PV
/dev/dsk/c2d0s2 10 10

- - - Logical extents ---

LE PV1 PE1 Status 1
0000 /dev/dsk/c2d0s2 0000 current
000 1 /dev/dsk/c2d0s2 000 1 current
0002 /dev/dsk/c2d0s2 0002 current
0003 /dev/dsk/c2d0s2 0003 current
0004 /dev/dsk/c2d0s2 0004 current
0005 /dev/dsk/c2d0s2 0005 current
0006 /dev/dsk/c2d0s2 0006 current
0007 /dev/dsk/c2d0s2 0007 current
0008 /dev/dsk/c2d0s2 0008 current
0009 /dev/dsk/ c2d0s2 0009 current

Figure 9-9. verbose Output of lvdisplay, Showing Logical Extents

9-22 Logical Volume Manager

Allocating Disk Space for Logical Vol umes
• Using traditional HP-UX, once you partition the disk into fixed disk sections
of various sizes, you make your file systems in the context of these sections.
Supported section sizes are shown in / etc/diskt ab.
• Using LVM, you create a logical volume and allocate disk space for file
systems, swap, or raw data in either megabytes or an LVM unit of measure
called an extent .
By default , LVM extents are 4 MB . Extent size is configurable, provided the
size chosen is a power of two ( for example, 1 ,2 ,4,8) ; valid extents can range in
size from 1 MB to 256 MB .

Note When calculating the size of a logical volume to contain a file

system, base your calculations on how many megabytes a file
system needs, but choose a quantity divisible by the extent size.
Any quantity not divisible by extent size is rounded up .

Under most circumstances, you need never change the 4-MB default extent
size. Exceptions might arise for extremely large disks. Large extents are more
wasteful of disk space and smaller extent sizes allow a finer granularity of
allocation. ( For more information on these considerations, see "Volume Group
Descriptor Area ( VGDA ) " , later in this chapter. )
As you can see in Figure 9-9, LVM uses two kinds of extents-physical extents
and logical extents. LVM stores data on disks as sets of addressable, disk
blocks called physical extents. Logical volumes consist of logical extents, which
map to the disks ' physical extents.
An unmirrored logical volume has an identical number of physical and logical
extents; a doubly mirrored logical volume has three physical extents to each
logical extent .
Logical extents can be mapped non-contiguously to physical extents on LVM
disks. This means that LVM can disperse data in non-consecutive physical
extents on disk. An exception to this policy is the root volume group. ( See
"Contiguous vs. Non-Contiguous Allocation of LVM Disk Space" , earlier in 9
this chapter. )

Logical Volume M anager 9-23

 How LVM Maps Extents to Logical Volumes
The LVM subsystem maps logical extents to physic al extents via a translation
table that resides on the LVM disk. When the volume group is activated,
the table resides in real memory. LVM translates incoming read and write
requests to the correct address of the physical extent , then sends the request
to the corresponding physic al block . Thus , the extent serves as a translation
mechanism between the incoming request and underlying device drivers.
By default , LVM selects available physical extents from LVM disks in the order
the disks were added to the volume group .
The physical-to-logic al extent mapping of a logical volume can be
discontiguous; a single logical volume (and therefore file system) can exist on
several different LVM disks. File system performance is best , however, when its
physical extents are contiguous , whether on one or multiple LVM disks .
As a system administrator, you can control to which LVM disks a logical
volume is mapped (bypassing the default LVM's distribution) , by using a
two-step process of lvcreate and lvext end. This is documented in System
Administration Tasks manual, Chapter 7, "Managing Logical Volumes."

LVM Disk Space Allocation Commands

The following commands let you create and view disk-space allocation:
• You can use SAM or the vgcreate( lM) command to set the size and number
of physical extents in a volume group .
• You can use SAM or the lvcreate(lM) command to assign the number of
physical extents in a logical volume.
• You can increase the number of physical and logical extents in a logical
volume by using the lvextend( lM) command.
• You can display the size (in bytes) of physical extents (PE) in a volume group
by using the vgdisplay(lM) command.
• You can display the number of logical extents (LE) in a logical volume by
using the lvdisplay( lM) command.

9-24 Logical Volume Manager

Understanding Physical Vol umes (LVM Disks)
Physical volumes (also called LVM disks) are disks that have been initialized
(using SAM or pvcreate) for LVM. Whereas traditional disks are partitioned
into sections , space on LVM disks is allocated for logical volumes.
Physical volumes use the same device speci al files as traditional HP-UX disk
devices. Once a disk has been initialized, you can view its characteristics as
an LVM disk by running the pvdisplay command, giving as an argument the
disk's block special file.

Y. pvdisplay /dev/dsk/c2d0s2
--- Phys ical volumes
PV Name /dev/dsk/c2d0s2
VG Name /dev/vgOO
PV St atus available
Allocat able yes
VGDA 2
Cur LV 4
PE Size (MB ) 4
Total PE 632
Free PE 462
Allocat ed PE 170
Stale PE 0

Figure 9· 1 0. Sample Output of pvdisplay

A verbose listing (pvdisplay -v) of the same LVM disk shows the mapping
status of all available physical extents. (In the following example, physical
extents 0 170 to the end are free, meaning unallocated.)

Logical Volume Manager 9-25

 --- Distribut ion of physical volume ---
LV Name LE of LV PE for LV
/dev/vg00/lvol 1 10 10
/dev/vg00/lvol2 20 20
/dev/vg00/lvol3 100 100
/dev/vg00/lvol4 40 40

--- Physical extent s ---

PE St atus LV LE
0000 current /dev/vg00/lvo1 1 0000
000 1 current /dev/vg00/lvol 1 000 1
0002 current /dev/vg00/lvo1 1 0002
0003 current /dev/vg00/lvo1 1 0003

0 170 free 0000

0 1 7 1 free 0000
0 172 free 0000

Figure 9 - 1 1 .
Verbose Output of pvdisplay, showing Mapping of Physical and Logical Extents

Sector Size of Physical Volumes (LVM Disks)

In HP-UX, different disk types have different sector sizes ( shown in
Figure 9- 12). HP-IB and HP-FL disks have 256 bytes per sector; SCSI disks
have 512 bytes per sector; multi-spindle disk arrays have 2048 bytes per sector.
LVM deals with this diversity without user intervention.
The disk device drivers ( such as disc3 for SCSI) understand the language
of the physical disk as well as a requirement of HP-UX, which specifies that
device drivers be able to receive data in units of DEV _BSIZE ( 1024) bytes and
pass the data to the disk in its own terms.

9-26 Logical Volume Manager

 DEV BSIZE
(1 024 bytes)

H P-1 8 H P-FL Multispindle

SCSI disks disk arrays
disks disks
256 bytes 256 bytes
51 2 bytes up to
per sector 2048 bytes
per sector per sector
per sector

LG20021 1 _030

Figure 9- 1 2. Relationship Between Disk Sector Size and DEV_BSIZE.

When using the raw interface to a device ( as with a database application ) , data
is read and written in DEV BSIZE units. This can pose a performance loss for
_

multi-spindle disk arrays, whose writes are required in units of 2 KB, the disk
array's underlying sector size. For example, to transfer 1 KB ( that is, one
DEV BSIZE unit ) of data, the controlling device driver reads in 2 KB of data
_

( even though only 1 KB is changing ) in order to write in the 2-KB sector size
required by the disk array. ( This is called "read modify write." )
For block interface ( as with a file system ) , there is no performance loss as long 9
as the fragment size is 2 KB or greater.

Logical Volume M anager 9-27

 Note Any file system that resides , even partially, on a multi-spindle
disk array should have its fragment size be a multiple of 2 KB .

LVM handles all data transfers in terms of DEV BSIZE units . In most cases ,
_

LVM communicates with the disk driver directly, to ensure that 1/0 block
size is a multiple of DEV BSIZE and the beginning 1/0 address begin on the
_

DEV BSIZE boundary. LVM allocates disk space in units of extents. Extent
_

sizes range from 1 MB to 256 MB and must be a power of 2.

9-28 Logical Volume Manager

Characteristics and Layout of LVM Disks
There are two kinds of LVM disk layouts-for boot disks and other LVM
disks-and they differ in their data structures . Non-bootable disks have two
reserved areas-the physical volume reserved area (PVRA) and the volume
group reserved area (VGRA) , while bootable disks have additional sectors
reserved for the boot data reserved area (BDRA) and LIF.
Figure 9- 13 shows the layouts of bootable and non-bootable LVM disks . The
text that follows describes each of the component parts and then discusses
LVM overhead.

LI F Header (8 blocks) Physical Volume Area

(PVRA)
Physical Volume Area (PVRA)
Volume Group Reserved Area
Boot Data Reserved Area (BORA)
(VGRA)
LI F Volume
(Boot Programs)
Volume Group Reserved Area
(VGRA) User Data Area

User Data Area

.• � . .�
, ,

v
,
• •

Bad Block Relocation Pool Bad Block Relocation Pool

LG20021 1 _031

Figure 9· 1 3.
Bootable and Non-Bootable Physical Volume Layouts, Showing Organization of 9
Data Structures

Logical Volume Manager 9·29

 Boot Data Reserved Area (BDRA)
To boot the system, the kernel activates the volume group to which the
system's root logical volumes belong. The location of the root logical volumes
is stored in the boot data reserved area ( BDRA ) . The boot data reserved area
contains the locations and sizes of LVM disks in the root volume group and
other vital information to configure the root, primary swap , and dump logical
volumes, and mount the root file system.
The BDRA contains the following records about the system's root logical
volumes:
• Timestamp , indicating when the BDRA was last written
• checksum. for validating data.
• Root volume group ID .
• Number of LVM disks in the root volume group.
• A list of hardware addresses of the LVM disks in the root volume group , and
indices into that list for finding root , swap , and dump .
• Information needed to select the correct logical volumes for root , primary
swap , and dumps.
Information about the LVM disk data structures in the BDRA is maintained
by using the lvlnboot(lM) and lvrmboot( lM) commands. Here is sample
output, followed by explanation:
# . /lvlnboot -v -d /dev/vg00/lvol2
Boot Def init ions for Volume Group /dev/vgOO :
Phys ical Volumes belonging in Root Volume Group :
/dev/dsk/ c3d0s2 Boot Disk
/dev/dsk/c4d0s2 -- Boot Disk
/dev/dsk/c5d0s2
/dev/dsk/c 12d0s2 -- Boot D isk
Root : lvol 1 on : /dev/dsk/c3d0s2
/dev/dsk/c4d0s2
Swap : lvol2 on : /dev/dsk/ c3d0s2
/dev/dsk/ c4d0s2
Dump : lvol2 on : /dev/dsk/c3d0s2 , 0
9
• The physical volumes (LVM disks ) designated "Boot Disk" are bootable,
having been initialized with mkboot and pvcreat e -B.

9-30 Logical Volume Manager

• Multiple lines for lvol 1 and lvol2 reveal that the root and swap logical
volumes are being mirrored.
• Notice that lvol2 is being used for both swap and dump , but that mirroring
applies to only swap .

LIF Information

Much like non-LVM boot disks, the LVM boot disk contains a LIF volume, in
which are stored ISL, HPUX, AUTO , IOMAP , RDB , and LABEL .
The LIF LABEL file is created and maintained by lvlnboot and lvrmboot . The
LABEL file has pointers to the starting point and size of boot-relevant logical
volumes, including the file system containing the kernel. With information in
LABEL , HPUXboot can locate and gain access to the logical volume for root .
The support and install tapes also use the information to access the root,
primary swap, and dump logical volumes without actually using LVM.

Caution If the LIF or any reserved area becomes corrupted, the support
tape can be used to recover the operating system. If that fails,
or is impractical, the system must be re-installed.
The LIF volume cannot be copied directly onto LVM disk
using dd( IM) . It must be set up by the install process or by
mkboot(IM), on LVM disks for which the pvcreate -B option
was used, due to a gap between the LIF header and LIF
directory. When using mkboot(IM), you should specify section
2, not section 6.

Logical Volume Manager 9-3 1

 Physical Volume Reserved Area (PVRA)
Everything LVM needs to know about an LVM disk is contained in the
physical volume reserved area ( PVRA ) . The physical volume reserved area
stores information about the LVM disk configuration and operation, starting
addresses, and sizes. It also contains a bad block directory for the LVM disk.
The PVRA contains the following record of the LVM disk in relation to its
volume group :
• LVM identification field, signifying that the LVM record is present and valid.
• Unique physical volume ID , checked by the LVM when it generates new
physical volumes.
• Physical volume number within the volume group.
• La.st physical sector number, used in determining the amount of space
available on the disk.
• Size of each physical extent on the LVM disk, expressed exponentially in
DEV BSIZE blocks. All physical extents are identical in size for all disks in a
_

volume group.
• Amount of space ( expressed in number of DEV BSIZE blocks ) allocated for
_

each physical extent on the LVM disk.

The PVRA also contains the starting addresses ( physical sector number and
length ) of the following areas:
• Boot Data Reserved Area ( for boot volume groups )
• Volume Group Reserved Area
• Volume Group Descriptor Area
• Volume Group Status Area
• Mirror Consistency Records
• User Data Area
• Bad Sector Relocation Pool
The bad block directory of the PVRA contains record of all software sectors
that are faulty, have been relocated, or are awaiting relocation. Hardware
sectors are not included.

9-32 Logical Volume Manager

Volume Group Reserved Area (VGRA)
The volume group reserved area ( VGRA ) describes the volume group to which
the disk belongs. This information is organized in three subareas:
• Volume Group Descriptor Area ( VGDA )
• Volume Group Status Area ( VGSA )
• Mirror Consistency Record

Volume Group Descriptor Area (VGDA)

The volume group descriptor area ( VGDA ) contains the information the LVM
device driver needs to configure the volume group for LVM, including:
• Volume group header with timestamp to indicate when the VGDA was
last updated, volume group ID number, and three configurable operating
system parameters ( see System Administration Tasks manual for further
information ) :
o maxl vs-maximum number of logical volumes allowed per volume group.
o maxpxs-maximum number of physical extents allowed per LVM disk.
o maxpvs-number of LVM disks that can be installed in the volume group .
• List of logical volume entries, one for each logical volume within the volume
group , recorded as follows:
o Maximum number of logical extents permissible per logical volume.
o Current status and capabilities of the logical volume.
o Mirroring schedule policy, if set.
o Maximum number of mirror copies allocated, if set .
• List of LVM disks, including:
o Header with global information ( ID , number of physical extents, status )
about the disk.
o Map of physical extents to logical volumes.
• Volume group trailer, timestamped when the VGDA was last updated. ( Both
header and trailer timestamps are compared to verify the consistency of the
VGDA. )
The VGDA is the area checked to ensure a quorum of total LVM disks in the
volume group . A volume group cannot be activated unless half the disks in
a volume group are present and available, that is, containing the latest state 9
information. The vgchange -q n option can be used to override the system's
quorum check.

Logical Volume Manager 9-33

 Caution Overriding quorum can result in a volume group whose
configuration is inaccurate (for example, missing recently
creating logical volumes) . This configuration change might not
be reversable.

Since each physic al extent is recorded in a YGDA entry, the extent size has a
direct bearing on the YGDA. In most cases, the default extent size will work
perfectly well. However, if you run into problems , you might consider the
following circumstances:
• Since the YGDA is a fixed size, a high-capacity LYM disk might exceed the
total number of physical extents allowed. As a result , you might need to use
a larger-than-default extent size on high-capacity LYM disks.
• Conversely, if all LYM disks in a volume group are small, the default number
of extents might make the YGDA too large, wasting disk and memory space.
A smaller-than-default extent size or number of physical extents might be
preferable.
• A high-capacity LYM disk might be unusable in a volume group whose
extent size is small or set with a small number of physical extents per disk.

Volume Group Status Area (VGSA)

The volume group status area (YGSA) tracks the validity and availability
of LYM disks and the current state of physical extents (stale or fresh) . The
LYM reads this information when it initializes the volume group , updates it as
a result of configuration changes to the volume group, or as a result of I/O
errors.
The YGSA can be considered an extension of the YGDA, because it duplicates
some of YGDA information, such as quorum, maxpvs , and maxpxs .

Mirror Consistency Record (MCR)

When the OSF Mirror Write Cache is used, the mirror consistency record
(MCR) minimizes the amount of 1/0 required to bring all mirrors into a
consistent state following a system crash or power failure.
9 The MCR contains records of:
• Timestamp when the mirror consistency record was last written.
• Minor numbers of the logical volumes involved.

9-34 Logical Volume Manager

• Number and size of the logical track group to which the logical volume is
associated.

User Data Area

The user data area is the region of the LVM disk used to store all user data,
including file systems, virtual memory system ( swap ) , or user applications.
When you create the volume group , you apportion the user data area into
fixed-size physical extents. Physical extents map to logical extents that hold
user data-file systems, virtual memory subsystem, or a user application.
By default , physical extent size is 4 MB . On high-capacity LVM disks, the
default limit on the total number of physical extents per LVM disk might
necessitate a larger extent size, or a larger number of physical extents per LVM
disk.

Bad Block Relocation Policy

Bad blocks can cause serious data transfer problems. The LVM device driver
checks each physical request for bad blocks in the address range of the request .
If it encounters a bad block on a mirrored LVM disk, LVM can recover from
the error by reassigning the data to other blocks on the disk. This feature,
called bad block relocation, is enabled by default .
With this feature, LVM supports two types of bad block relocation:
• Soft relocation, which assigns a new block to replace the defective one.
• Hard relocation, in which LVM instructs the disk device driver to spare the
bad sector ( s ) .
The following commands deal with LVM bad block relocation policy:
lvcreate(lM) Sets bad block relocation policy for logical volumes being
created in a given volume group .
lvchange(lM) Allows you to change the policy to prevent or allow bad block
relocation for logical volumes in a given volume group.
pvcreate(lM) Lets you specify the minimum number of bad blocks that LVM
should reserve to perform software bad block relocation. ( By 9
default , one block for every 8K of data blocks is reserved. )

Logical Volume Manager 9-35

 It is good practice to run diagnostics testing for bad blocks on any device
before initializing an LVM disk.

LVM Overhead
The data structures that enable you to use LVM consume a modest amount of
overhead from your disk space. This overhead is set at a fixed boundary (2912
KB) for bootable LVM disks and may vary in size in non-bootable LVM disks
( typically 400 KB).
Overhead required by non-bootable LVM disks depends on the parameters
listed below, which can be modified using SAM or the vgcreate command. If
you set a small extent size or create many physical volumes, your LVM data
structures ( overhead ) will be larger.
-e Sets the maximum number of physical extents allocatable for
LVM disks in a volume group ( by default , 1016).
- 1 Sets the maximum number of logic al volumes allowed in a
volume group ( by default , 255).
-p Sets the maximum number of LVM disks ( physical volumes )
allowed in a volume group ( by default , 32) .
-s Sets the size, in megabytes , for each physical extent in a
volume group ( by default , 4) .
An operating-system parameter, maxvgs , can be configured to set the
maximum number of volume groups LVM will recognize. Its default ( 10) is in
the I etc/master file and can be overridden ( permissible range is 0 to 256) by
adding a statement to the S800 file.

9-36 Logical Volume Manager

Understanding LVM Configuration Maintenance
While LVM provides desirable flexibility on your HP-UX system, it also adds a
level of complexity for system administration. Be sure that you safeguard your
LVM configuration.

Guidelines for Configuring the Root Volume Group

The following options are recommended when setting up the root volume group
using lvcreate(lM) command. The same conditions can be applied when using
SAM :
-
c y Makes extents contiguous.
-r n Turns off bad block relocation.
For any swap logical volume, the following options should be used:
- c y Makes extents contiguous.
-M n Turns off the Mirror Write Cache ( used for mirroring ) .
-c n Turns off all synchronizaton.
A dump logical volume requires only the following speci al options, as long as it
is not also a swap logical volume, in which case it requires the same options as
swap .
-c y Makes extents contiguous.

The /etc/lvmtab File

At the heart of the LVM configuration is the / etc/lvmtab file, which is read
by all LVM commands. / etc/lvmtab is not readable or editable on-screen.
The /etc/lvmtab file is run-time generated; that is, it is generated the
first time you create an LVM entity using SAM or commands such as
vgcreat e , and updated every time you change the LVM configuration. Every
configuration update or query reads the I etc/l vmtab file.
LVM disk file names are recorded in /etc/lvmtab . If /etc/lvmtab file is
destroyed, all configuration operations involving LVM data structures become
9
impossible.

Logical Volume Manager 9-37

 You can recover the / etc/lvmtab file using vgscan. To safeguard against LVM
configuration problems , use the vgcfgbackup(lM) and vgcfgrestore(lM) utilities
to back up and restore the LVM configuration.

Note vgcfgbackup and vgcfgrestore back up and restore only the

LVM configuration, not the user data the LYM data structures
contain. You still have to back up and restore user data using
HP-UX utilities, such as fbackup and frestore.

The vgcfgbackup command backs up volume-group configuration information

into binary files , one file per volume group .
The vgcfgrestore command restores LYM disk configuration in case of disk
loss or data corruption.
vgcfgbackup backs up the following:
• LVM record of each LVM disk.
• BDRA record for each bootable LVM disk.
• One copy each of the current VGDA and YGSA data structures.
• LIF LABEL files.
vgcfgbackup does not back up the following:
• LIF header and files
• bad block directory.
Another command useful in maintaining the /etc/lvmt ab file is vgscan(lM).
If I etc/l vmtab is deleted or corrupted, running vgscan recreates the
I etc/l vmtab file. vgscan searches every LYM disk on the system for logical
volumes, then groups them into volume groups by searching the /dev directory
and matching major and minor numbers.

9-38 Logical Volume Manager

Dealing with Removable Media and Volume Groups
Before a volume group created on another computer system can be accessed,
the volume group must be associated with its new system by being written
into / etc/lvmtab . Two commands, vgexport ( lM ) and vgimport ( lM ) deal with
these circumstances.
The vgexport command removes the definition of a volume group from the
system (/etc/lvmtab and device files ) without removing it from the disks.
This allows the disks to be brought to another system and used there ( after
you execute vgimport ) .
The vgimport command scans a specified list of LVM disks, rebuilds volume
group information, updates I etc/l vmtab, and sets up LVM special files for the
new volume group.

LYM Maintenance Mode

Maintenance mode boot provides a means , during system installation, or when
critical LVM data structures have been lost , to boot from an LVM disk without
the LVM data structures.
After the support tape has made the LVM disk minimally bootable, the
system can be booted in maintenance mode using the lm option to the hpux
-

command at the ISL> prompt . This causes the system to boot to single-user
mode without primary swap , dump, or LVM to access the root file system.
For further information, refer to the manual, Installing and Updating HP- UX
or hpuxboot ( lM ) in the HP- UX Reference Manual.

Caution The system must not be brought to multi-user mode ( that is,
init 2) when in LVM maintenance mode. Corruption of the
root file system might result .

Logical Volume Manager 9-39

 Mirroring

Note Mirroring requires the optional product , HP MirrorDisk/UX,

part number B 2491A.
Mirroring is not available on HP-IB disks.

Mirroring is the capability of storing identical copies of data on separate LVM

disks or on separate locations of the same disk. This redundancy has several
advantages :
• The system can survive LVM disk crashes if you mirror the root file system
and swap .
• Valuable data is available on more than one LVM disk. This benefit is known
as high availability. If an I/O channel fails, LVM can recover the data from
the duplicate source.
• Mirror-write recovery mechanisms enable the system to synchronize data.
• Mirroring speeds read-intensive applications by enabling the hardware to
read data from the most convenient LVM disk, thus optimizing I/O .
• Backups can b e done on one copy of the data while another copy continues
to run.
• Mirroring of data to areas of the same LVM disk allows you to overcome
local media errors.

9-40 Logical Volume Manager

 ~
/dev/dsk/c0d0s2

~
/dev/dsk/c1 d0s2
/dev/dsk/c1 d0s2

LG20021 1 _032

Figure 9- 1 4. Strict Mirroring of Logical Volume lvol2

Figure 9- 14 shows logical volume /dev/vg00/lvol2 mapped to physical

extents on LVM disks /dev/dsk/c0d0s2 and /dev/dsk/c 1d0s2 (left side ) or
to different areas of physical extents on the same LVM disk, /dev/dsk/c 1d0s2
( right side ) .
In Figure 9- 15, you can see how mirrored logical volumes (LV) might distribute
over three LVM disks . ( In this example, all logical volumes but LV5 are
mirrored. )

Logical Volume Manager 9-4 1

 LV3

LV3

LV4 LV4

LG20021 1 _xx

Figure 9- 1 5. Three LYM Disks with Mirrored Logical Volumes

Both examples in Figure 9- 14 demonstrate single mirroring. That is, in both

cases , 1 vol2 maps from one logic al volume to two sets of physical extents on
LVM disks . Another kind of mirroring is possible. Double mirroring maps one
logical volume to three sets of physical extents on LVM disks. (Sets of physical
extents can map strictly to separate LVM disks or non-strictly to different
areas of the same disk or disks .) Double mirroring is shown in Figure 9- 16.

9-42 Logical Volume Manager

 LG20021 1 _033

Figure 9- 1 6. Three LVM Disks Mirror a Single Logical Volume

Each copy of mirrored data maps to the same logical volume; the number of
logical extents remains constant , but the number of used physical extents ( and
therefore, occupied disk space ) changes, depending on the number of mirrored
copies.
Mirrored logical volumes must belong to the same volume group ; you cannot
mirror across volume groups.

1/0 Channel Separation

1/0 channel separation is an approach to LVM configuration requiring that
mirrored copies of data reside on LVM disks accessed via separate device
adapters ( interface cards ) and cables. 1/0 channel separation achieves higher
availability and better performance by reducing the number of single points
of possible hardware failure. If you mirror data on two separate disks, but 9
through one card, you are vulnerable to failure if the card fails.

Logical Volume Manager 9-43

 You can separate 1/0 channels on a system with multiple disk adapters but a
single bus ( such as HP Model 8578 ) by mirroring disks across different interface
cards. This is shown in Figure 9- 17, in which the LVM disks at addresses
44 . 0 . 0 and 48 . 0 . 0 are accessed through different I/ 0 channels ( slots 1 1 and
12 ) on the HP-PB bus .

SCSI
host adapter
@ slot 1 1

H P-PB
bus
SCSI
host adapter
@ slot 1 2

/dev/dsk/c3d0s2
@ 48.0.0

LG20021 1 _034

Figure 9- 1 7. 1/0 Channel Separation via Separate Interface Cards

9 On systems such as the Model 8708 , you can mirror disks across separate
buses, as shown in Figure 9- 18. The logical extents ( LE1 , LE2 , and LE3 ) in
memory map to the physical extents ( PE 1 , PE2, and PE3 ) in LVM disks labeled
PV1 and PV5 .

9-44 Logical Volume Manager

 LE1
Processor(s) LE2 Memory
LE3

System Memory Bus

Central Bus O Central Bus 1

CIO Bus O CIO Bus 1

slot 2 with slot 2 with

H P-FL interface card H P-Fl interface card

- --
C\/? C\/1 PVr::
- - -l!:i._7
PE1 PE1
-
- PE3
PE2 PE2
PE3
-
- -
- - - �

PV� PV.4 PVR - - PVR :::::

- � �

Physical Volume Group Physical Volume Group

Volume Group

LG20021 1 _035

Figure 9- 1 8.
1/0 Channel Separation via Separate Buses, and using Physical Volume Groups

You can further ensure channel separation by establishing a policy caled

PVG-strict allocation, which requires logical extents to be mirrored in separate 9
physical volume groups. Physical volume groups are subgroups of LVM disks
( physical volumes ) within a volume group . An AS CII file, /etc/lvmpvg
contains all the mapping information for the physical volume group , but the

Logical Volume Manager 9-45

 mapping is not recorded on disk. Physical volume groups have no fixed naming
convention; you might name them PVGO, PVG 1 , and so on, or something more
intuitive. The /etc/lvmpvg file is created and updated using the vgcreat e ,
vgextend, and vgreduce commands.
I/O channel separation is particularly useful for databases, because it heightens
availability (LVM has more flexibility in reading data on the most accessible
logical extent) , resulting in better performance. If you define your physical
volume groups to span I/O devices , you ensure against data loss, even if one
card fails.
For maximum performance, group the LVM disks by type (for example, all
HP-FL or all SCSI) . When using physical volume groups, you might also want
to use a PVG-strict allocation policy for logical volumes.
As with other mirroring features, physical volume groups are not supported on
HP-IB .

9-46 Logical Volume Manager

How Mirrored Logical Volumes are Written
Three sets of policies govern how mirrored logical extents are written to
physical extents:
• Allocation policy.
• S cheduling of disk writes.
• Synchronization policy for crash recovery.
Allocation, scheduling, and synchronization can be set using SAM,
lvcreate(lM), or lvchange(lM).
Allocation Policies for Mirroring

Mirrored data can be distributed on LVM disks by strict or non-strict ,

contiguous or non-contiguous allocation. By default , allocation of mirrored
logical volumes is set to strict , non-contiguous.
• Strict allocation requires logical extents to be mirrored to physical extents on
different LVM disks.
• Non-strict mirroring allows logical extents to be mirrored to sets of physical
extents, not necessarily on different LVM disks .
• Contiguous allocation policy has three characteristics:
o Physical extents are allocated in ascending order.
o No gap exists between physical extents within a mirror copy.
o All physical extents of a mirror copy reside in a single LVM disk.
• Non-contiguous allocation allows logical extents to be mapped to
non-consecutive physical extents.
When the root volume group is mirrored, it must be set up with contiguous
extents. That is, the physical extents used by logical volumes for root , primary
swap, or dump devices must consecutive for all mirrored copies.

Scheduling Policies for Disk Writes

The LVM scheduler converts logical requests into one or more physical
requests, then schedules them through the physical layer. S cheduling occurs for
both mirrored and non-mirrored data.
9
To ensure maximum control, two 1/0 schedule policies are available-parallel
and sequential.

Logical Volume Manager 9.47

 • Parallel. The parallel scheduling policy is used by default with mirroring for
maximum 1/0 performance. Parallel scheduling causes mirror operations
to write simultaneously to all copies . LVM performs reads in an optimized
fashion, by reading from the LVM disk with the fewest outstanding 1/0
operations .
• Sequential. The sequential policy causes mirror write operations to proceed
sequentially; that is, one write completes before the next write begins.
Likewise, LVM mirrors are read in a predefined order. On a practical level,
sequential policy would be used only for extreme caution in maintaining
consistency of mirrors.

Synchronization Policies for Recovering Mirrored Data

Maintaining consistency of mirrored data can be accomplished by configuring

your system using any of three different approaches:
• Mirroring with the OSF Mirror Write Cache enabled.
• Mirroring with the OSF Mirror Write Cache disabled and LVM Mirror
Consistency Recovery enabled.
• Mirroring without using an LVM mirror consistency mechanism.
The OSF Mirror Write Cache provides a fast resynchronization of data
following a system crash or failure, but at a potential performance cost for
routine system use.
The OSF Mirror Write Cache keeps track of all available space on the LVM
disk to which you are writing, periodically recording the activity in an on-disk
data structure called the Mirror Write Consistency Record. An extra disk
write is required for every mirrored write not already marked on the LVM
disk. This slows down on-line processing and degrades performance when you
access the disk at random; when writing to an area of the disk that is already
marked, performance is not impaired. Upon system reboot after crash, the
operating system uses the relatively small record of the Mirror Write Cache to
resynchronize inconsistent data blocks quickly.
The frequency of extra disk writes is small for sequentially accessed logical
volumes ( such as database logs) , but increases when access is more random.
g Therefore, logical volumes containing database data or file systems with few
or infrequently written large files ( greater than 256 KB) should not use the
OSF Mirror Write Cache when run-time performance is more important than
crash-recovery time.

9-48 Logical Volume Manager

When mirroring with the OSF Mirror Write Cache disabled and LVM
Mirror Consistency Recovery enabled, LVM does not impact run-time 1/0
performance. However, for any logical volumes using Mirror Consistency
Recovery, the entire data space is resynchronized when the volume group is
activated. Synchronization is performed in the background without interfering
with reboot or access; however, while during this time, I/O performance is
degraded. This approach is useful if optimal run-time I/O performance is
required.
When mirroring without using any LVM mirror consistency mechanism, the
operating system's run-time behavior is identical to that of the previous
approach. This approach is useful when running an application program ( such
as a database ) that has its own means of maintaining or recovering consistent
data, such as transaction logfi.les. Note, however, that the database logfi.les
themselves can be configured as a mirrored logical volume to use effectively the
OSF Mirror Write Cache.
Manual Synchronization.
An HP-UX command, lvsync(lM), can be used to manually refresh extents
that have been marked "stale." This approach recovers "stale" extents by
reading from a "current" mirrored physical extent and writing to the "stale"
physical extent .

Backing Up Mirrored Data

B ackups can be performed on one copy of an off-line logical volume, while
another copy is operational. This is done using the 1vspli t command to take
a mirrored logical volume off-line, backing it up, then using lvmerge to return
the backup copy of the logical volume on-line. Double mirroring ( mirroring
onto three LVM disks ) enables you to continue mirroring while performing the
backup . ( This is especially useful for high availability environments. )

Useful Mirroring Commands

Although you can perform all mirroring tasks using SAM, the commands listed
below deal with synchronization and scheduling in a mirrored environment.
Their manual pages, in HP- UX Reference , provide further information about 9
LVM capabilties. Procedures for the tasks to which they are related are

Logical Volume Manager 9-49

 found in Chapter 8, "Managing Logical Volumes" of the Series 800 System
Administmtion Tasks manual.
lvcreate(lM) Creates a logical volume in a volume group , at which time you
can set the scheduling policy, set the number of mirror copies ,
enable or disable the O S F Mirror Write Cache, and enable or
disable the LVM Mirror Consistency Recovery mechanism.
lvchange( lM) Changes the characteristics of a logical volume. This includes
scheduling policy, enabling or disabling the Mirror Write
Cache, enabling or disabling the Mirror Consistency Recovery
mechanism.
lvextend( lM) Increases the number of physical extents allocated to a logical
volume.
lvdisplay( lM) Shows information about logical volumes, including the mode
set for mirror consistency recovery and scheduling policy.
lvsync(lM) Synchronizes physical extents that are stale in one or more
mirrored logical volumes.
vgchange(lM) Sets volume group availability to the logical volume manager.
lvsplit( lM) Divide a mirrored logical volume into two logical volumes.
lvmerge( lM) Combine two logical volumes into one logical volume.

9-50 Logical Volume Manager

I nternal Representation of LVM
Figure 9- 19 illustrates the relationship between the LVM subsystem and the
HP-UX system architecture in general:

system mgt application

com m ands programs

• •
File
System
User Space
System Call Boundary
Kernel

�
,, ' �.

logical P rocess context

volume data (top halt) "Iv" pseudo-device
structures · - - - - - - - -
driver 1/0 and
1/0 context configuration routines
(bottom halt)

• T
physical device
}
drivers ("discn'1

LG20021 1 _0 1 2

Figure 9· 1 9. Architecture of the LVM Subsystem

The LVM pseudo-device driver (lv) handles all 1/0 operations for the LVM
components on behalf of applications, file systems, and other subsystems.
System management commands perform LVM configuration operations by
opening the volume group 's control device and issuing LVM ioctl c alls. When
applications issue open, read, write, ioctl, and close calls, the top h alf of
1 v does everything required to send a request to the underlying disk drivers . 9
This includes ensuring that requests do not overlap , choosing how to access
the mirrors, translating logical addresses to physical addresses, and calling the
driver to start the 1/0 .

Logical Volume Manager 9-5 1

 The bottom h alf of the LVM handles everything associated with request
completion from the underlying disk driver. This includes finding locating a
mirror if one is not already available, restarting requests previously blocked
due to overlap, restarting requests waiting for the Mirror Write Cache to be
written, and other functions.
The bottom half executes in interrupt context ; thus, it is always interrupting
user processes . However, since it is not running in the context of a process , the
LVM cannot access user process information. It also should complete as quickly
as possible and cannot go to sleep (for example, to wait to allocate memory) .
The 1 v pseudo-device driver maintains the on-disk data structures that
describe the volume groups. Each volume group has a separate entry in the
kernel device switch table, with entry points for the logical volume device
driver and a pointer to the volume group data structure. Each volume group's
control device file, called group (logical volume number 0), provides the means
for manipulating the volume group 's data structures. Each logical volume
in the volume group is accessible through a device node with its own unique
(non-zero) logical volume number.
To translate from a logical to physical operation, an 1 v _xlat e function
converts the logical address consisting of the volume group , logical volume,
logical block, and mirror number, into a physical address consisting of physical
volume and physical block. A mirror number and logical block index entry
maps into the logical volume's extent map , which can then be used to find the
LVM disk.

9-52 Logical Volume Manager

 10

10
System Architectures

This chapter surveys the various HP-UX system platforms , with figures of card
cages and bus structures. Visualizing the layouts of the specific architectures
will help you understand HP-UX addressing and configuration.

System Architectures 1 0- 1
10

Architecture of the Series 300 Bus

All HP 9000 Series 300 computers use either the DIO (Direct 1/0) or DI0-11
internal bus.
Buses are circuitry-the pattern of conductors (traces or wires) on a circuit
board-through which data travels between hardware modules (CPU, memory,
channel adapters, device adapters) within the system. Cards are plugged into
these buses.
The number of data conductors equals the number of data bits of the bus. A
16-bit DIO bus has sixteen conductors; a 32-bit DI0-11 bus has 32 conductors.
The 16-bit DIO bus transfers data at 1/0 bandwidths of 3 MB/second; the
32-bit DI0-11 bus transfers data at 6 MB/second.
DIO and DI0-11 cards come in various sizes:
• DIO system cards are designed in pairs. The upper card is designated as the
accessory card and the lower is the 1/0 card. (One exception: the Model 340
has a single DIO 1/0 card, and no accessory slot .)
• DIO system cards for processors or video interfaces are twice as wide as DIO
1/0 and accessory cards.
• DI0-11 cards are similar in size to the DIO system card, but include a 32-bit
backplane connection. Only DI0-11 cards are supported in DI0-11 slots,
unless a card is identified as supported in both types of slots. DIO adapters
are available for backward compatibility.
• Multiplexer (MUX) cards can be plugged into an internal DIO or DI0-11 slot
to connect terminals , modems , printers, plotters , and other RS-232C devices.
• Daughter boards are, strictly speaking, not interfaces; instead, they screw
onto cards to add RAM on a processor or RAM board, additional interface
on the system interface board, or an accelerator on a video board.
Both Model 350 and 370 workstations have an additional internal RAM bus,
fitted as a top plane opposite the DI0-11 connector. The RAM bus provides
higher speed, and therefore better performance, by bypassing the DI0-11 bus,
and thus avoiding bus collisions with anything running on the backplane.
External buses (high speed and standard-speed HP-IB , VME, S C SI, and
HP-HIL) connect peripheral devices to HP 9000 Series 300 computers.

1 0 ·2 System Architectures
 10

Except for Models 318 and 319 workstations, and other lower-priced
non-expandable units, you can enlarge your Series 300 memory or I/O
capability by adding expander boards, either piggy-backed to the System
Processor Unit ( SPU ) or HP-HIL card, or caged in a VME Expander bus
connection. The VME Expander is supported on all DIO-II hosts.

Model 375-a DI0-11 I mplementation

The elements depicted in Figure 10- 1 of a Model 375 workstation show the
basic organization of its boards and is typical of Series 300 models .

Processor-1/0 Board
(occupies 2 adjacent DI0-11 slots
SCSI Disk Interface
8MB ECC RAM (expandable to 32MB)
Power
Supply
Video Board

Available DI0-11 Slots

Figure 1 0- 1 . Backplane of a Model 375 Workstation.

Figure 10-2 depicts the functional relationship between the DIO-II bus on the
Model 375 and its attached boards.

System Architectures 1 0-3

10

Processor 1/0 Board

MC68882 Floating-Point Coprocessor

MC68030 C P U/Memory Management Unit -

32 KB C8che
S MB ECC RAM
4, 8, 1 2, 1 6, 20, or 24 MB add-on ECC RAM

Built-in HP-HIL, LAN, RS-232, HP-IB, SCSI Disk,

Centronics (Parallel) Audio Interfaces, and OMA

32-bit 010-11 Bus

Video Board �
r - - - - - - - - - - - - - - - - - - - ,

I Series 300 010-11 Interfaces 1--

� - - - - - - - - - - - - - - - - - - - �
r - - - - - - - - - - - - - - - - - - - ,

I Series 300 010-11 Accessories 1--

� - - - - - - - - - - - - - - - - - - - �

LG200 1 72_030

Figure 1 0-2. Model 375 Workstation Functions

To learn more about architectural differences and configuration possibilities ,

consult the HP 9000 Workstations Configuration Guide for each specific model.

1 0-4 System Architectures

 10

Processor-1/0 Board

The Processor-I/O board occupies one double-width slot on the DI0-11 bus of
the Model 375, and contains the following essential elements:
• Motorola MC68030 CPU /Memory Management Unit .
• MC68882 floating-point coprocessor.
• 32 KB cache (high-speed memory) .
• Error checking and correcting RAM (minimum 8 MB ) . Small SIMM RAM
cards mounted perpendicularly onto the CPU can increase the Model 375 to
a maximum of 128 MB of RAM.
The following elements of the processor-I/O board are specifically relevant to
configuration:
• SCSI (Small Computer System Interface) interface, used for connecting SCSI
peripherals, including magnetic and optical disks, and digital data storage
tape drives . S CSI is a very fast , industry-standard interface and an optional
product for most Series 300 systems.
• Centronics interface (parallel) is used by peripheral devices, such as printers.
Centronics is an eight-bit wide port capable of sending data through multiple
(eight) lines instead of one. This simultaneous transmission results in faster
data transfer than via the serial interface.
• High-speed HP-IB (IEEE-488) interface can be used to connect the System
Processor Unit (SPU) to disk drives. In some systems, a SCSI interface
replaces the high-speed HP-IB interface.
• RS-232C (serial) interface, used to connect a terminal, printer, plotter, or
other serial device to Series 300 systems .
• LAN interface, which enables you to hook up to a local area network.
• Audio output , which connects to an external speaker.
• HP-HIL (Human Interface Link) interface, used to hook up the keyboard and
other HP-HIL devices , such as a mouse, digitizer, and button box.
Other interfaces are available, including the standard-speed HP-IB interface,
used to connect numerous devices, including plotters, instruments, and tape
drives .

System Architectures 1 0-5

10

Graphics Interface

The video board provides the interface for the monitor. A number identifies
the video board, and should be used to set terminal type. Setting terminal
type incorrectly ( or too vaguely ) results in application products ( such as vi)
working less efficiently than intended.
The correct terminal type is complicated by the issue of how you will be using
your terminal. If you plan to work in raw mode, HP recommends that you
identify the actual product number of your display ( look for it near the HP
logo on the front or back panel ) or execute /etc/dmesg). The terminal type
( TERM ) corresponds to an entry in /usr/lib/terminfo, where the system looks
for information on terminal resolution, number of lines per screen, and other
monitor specifics .
To reference the proper terminal type, enter the following for Bourne or Korn
shells:
$ TERM=98550
$ export TERM
or for C shell:
Y. 1 : setenv TERM 98550
This applies the settings for the current terminal session, but to make them
permanent , you must add the same line to the prof ile file for Bourne and
.

Korn shells, or to login for C shell.

.

Note If running VUE or X Windows, use generic hp , hpterm, or

hp300h instead of the specific terminal type. This is because
VUE, X Windows , and SAM have requirements that overrule
specific terminal type.
Also, some terminal emulators ( such as LAN configurations )
require a different setting than indicated by the model number.
The documentation of the terminal emulator takes precedence
over these general guidelines.

A Graphics Accelerator, available on some color workstations ( including

340CHX, 360CHX, 370CHX), replaces the video interface and approximately
triples color graphics writing speed.

1 0-6 System Architectures

 10

Architecture of the Series 400 Bus

Like the HP 9000 Series 300 Workstations, the HP Apollo 9000 Series 400
Workstations run HP-UX. The Series 400 workstations and servers include
models 425e, 425t , and 425s, whose configurations include a variety of buses
(including DIO-II and E/ISA ) and networking capabilities , depending on
model. The following block diagram depicts the basic configuration:

_...,.,..
CPU

,.,,,, l...,..ll
11 -
J.J , 11
CPU Bus ________
rrtt@M@fitj�MMi��i���iiir@iiMW�iii=l��iiWtM �mt�iiiHi�::i:::::::::::::�W'ii'*MWMWf�:� rtwi:ti:

Bus Bus Bus

Converter Converter Converter

Built-in
---"'---
E/ISA
�� DI0-11
Graphics AT-Compatible
1/0 Card Cage
Card Cage

LG200 1 72_032a

Figure 1 0-3. HP Apollo 9000 Series 400 Workstations Block Diagram

For information on the HP Apollo 9000 Series 400 Workstations, consult the
installation guides , owner's manuals, and HP Apollo 9000 Series 400 and Series
700 Pricing and Configuration Guide.

System Architectures 1 0 -7
10

Addressing o n a Series 300/400

HP-UX finds devices connected to the bus by their select code and bus
address. The executable /etc/dmesg displays for the root user the system's
configured addresses, if run after boot-up .
Each select code and bus address uniquely identifies to the system an element
of the architecture, as shown in Figure 10-4:

DIO or 010-11 bus

Port
Bus address
Multiple
Devices (when required)
(Peripherals)

LG2001 72_031

Figure 1 0-4. Series 300/400 Addressing

Series 300/400 Select Codes

All extern al devices connect to an interface card, either standard or optional.
The interface card has a set of switches or jumpers , whose numeric value is
referred to as its select code. The select code value is used to specify the
port to which a device is connected, such as the LAN interface port shown
in Table 10- 1 . The select code uniquely identifies the interface occupying a
particular slot connected to the DIO or DI0-11 bus .

Note Every card in the system must have its own select code. Cards
cannot share select codes, even with other cards of the same
type.

1 0-8 System Architectures

 10

Table 1 0· 1 . Typical Series 300/400 Select Codes

Type of Interface Preset Select Code

Decimal Hexadecimal
High-speed HP-IB 1 14 OxOe
scs1 1 14 OxOe
Centronics 12 OxOc
Standard-speed HP-IB 7 Ox07
RS-232C (built-in) 9 Ox09
MUX (4-port multiplexer) 13 OxOd
MUX (8-port multiplexer) 28, 29 Oxlc, Oxl d
LAN 21 Ox15
HP-HIL 0 OxOO

1 H you have both high-speed HP-IB and SCSI interfaces,

the two must have different select codes.

In the simplest case, the select code uniquely identifies the interface on the
DIO or DI0-11 bus . In the case of HP-IB , SCSI, RS-232C, MUX, or HP-HIL,
when more than one peripheral can be connected to a single interface, the
select code is used to identify the interface and a bus address distinguishes the
specific device. In the case of multiple interfaces of the same type ( such as two
or more MUX cards ) , each card requires a unique select code.

System Architectures 1 0·9

10

Series 300/400 B u s Addresses

The interface card is placed in a slot in the backplane. The peripheral is
connected to the card via a port , the place of access through which I/O comes
into contact with the CPU. When an interface card is accommodating multiple
peripherals , a bus address is necessary to distinguish between those multiple
devices connected to a particular interface card.
Each device physically connected to an HP-IB , SCSI, or RS-232C MUX
interface must be identified uniquely by a bus address. Setting the bus address
may vary by device. For example, on the back of each HP-IB or S CSI device is
a series of DIP switches or a rotary switch, which can be set to distinguish bus
addresses between multiple devices .
Attaching to the MUX card is an automatic distribution panel with four or
eight ports (numbered 0-3 or 0- 7) into which a console or any serial device can
be connected. This port number is an alogous to a bus address and forms the
second set of digits in the MUX hardware address.
The address of each daisy-chained HP-HIL device is automatically determined
by its sequence, beginning from the HP-HIL port . Within the HP-HIL loop ,
many applications need the security ID module (part number 46084A) to run.
If you include the security ID module in the loop , be sure to place it between
the CPU and keyboard, and include it as one unit in the loop . Thus, for a
configuration consisting of an ID module, keyboard, and mouse, the ID module
is one, the keyboard address is two, and the mouse address is three.

Note Standard audio boxes do not have a bus address, even though
they are in the HP-HIL loop .

1 0- 1 O System Architectures
 10

Architecture of the Series 700 Bus

HP Apollo Series 700 workstations are designed around the PA-RISC 1 . 1 chip
set for demanding graphics and computation-intensive applications.
Figure 10-5 shows rear views of Models 720 and 730 ( left ) and 750 and 760
( right ) . Other models conform to the same basic design.

Graphics ------

SCSI
Parallel

LAN AUI
Graphics Thin LAN

RS-232

Keyboard
o+--tt-- Audio
a

o.

Figure 1 0-5. Models 720/730 and 750/760 Rear Panels, showing Ports

System Architectures 1 0- 1 1
10

8 M emory & g M ain

P rocessor System Bus Controller ....... . . . . . . . . . . . .

M emory
with Graphics Support

Standard Core l/0

EISA
Graphics
Adapter
Connect (SGC) 1 SCSI
2 LAN
3 HIL
4 Serial Port 1
5 Serial Port 2
6 Parallel

Voice-quality or
#
CD-quality Audio

# FOOi

# Fast/Wide SCSI

LG20021 1 _042

Figure 1 0-6. HP Apollo 9000 Series 700 Bus Architecture.

Figure 10-6 depicts the organization of the Series 700 bus architecture. Note
that not all elements shown in the diagram are available on all systems,
and addressing some elements of the core I/O varies by model. # symbolizes
optional interfaces whose function number varies by model. For example,
• EISA is optional on Model 720.
• The number of graphics connection varies by model.
• Voice- or CD-quality audio is available only on Models 710 and 760.

1 0- 1 2 System Architectures
 10

• Systems may be equipped with either LAN or FDDI, but not both.

Series 700 Modules

Each module depicted in Figure 10-6 has a system bus module number (shown
next to the bus); each interface of the core I/O also has a function number.
These bus module numbers and function numbers are used in device files and
minor numbers for the Series 700 . If you run /etc/ ioscan -f, you will see
these numbers displayed as part of the hardware path for each element of
the module. Chapter 1 1 , "System Configuration" discusses minor number
construction in detaH.
processor The processor module is implemented as a card with chips for
CPU, data and instruction caches, clocks, and floating point
coprocessor.
memory Depending on model, memory is implemented as SIMM cards
in 8MB , 16MB , 32MB , and 64MB increments on the processor
card, along with I/0-Memory control.
graphics The graphics adapter provides connectivity to color and
grayscale monitors, 24-plane and 3D graphics subsystem
capabilities. Voice- or CD-quality audio interface, plugged into
the SGC (System Graphics Connect) bus, is available on some
models .
core I/O Built-in I/O support is provided as a multi-functional card that
includes HP-HIL , audio, two asynchronous RS-232C ports,
IEEE 802.3 Ethernet LAN interface configured with ThinLAN ,
a Centronics parallel port , and an external SCSI-2 port . Each
function is identified by a separate function number. Possible
mass storage devices include 2 10MB , 420MB , 660MB , and
1 .3GB disks, CD-ROM, floppy disk drive, and DAT (DDS )
drive.
EISA An Extended Industry Standard Architecture (EISA) host
adapter provides connectivity to I/O and networking cards ,
including EISA S CSI-2 (differential) host adapter, EISA HP-IB
host adapter, EISA LAN/9000, FDDI/9000, X.25/9000, and
IEEE 802.5 Token Ring 9000 . The number of EISA slots varies
by Series 700 model. EISA devices are not configured using
SAM, but with the eisa_conf ig utility.

System Architectures 1 0· 1 3
10

Addressing o n a Series 700

During boot-up, the boot ROM identifies devices currently attached to the
system and their hardware paths. The boot ROM interface searches the
architecture for a device having a bootable file system. If it encounters a
cluster server, for example, it will give the LAN hardware address; if it finds a
bootable disk along the SCSI chain, it will report its address. When running
the boot ROM , if you press the (ESCaaj key, you get into the boot ROM
interface. ( See Chapter 2, "System Startup" for information.) You can see
the output of the boot ROM by looking at /etc/dmesg discussed later in this
chapter) .

Interpreting Series 700 Hardware Paths using ioscan

The ioscan ( lM) command is useful for visualizing the hardware configuration.
The numbered modules shown in Figure 10-6 correlate to hardware paths
displayed by ioscan -f . Here is a sample full listing on a Model 720:
Class H/V Path Driver H/V St atus S/V Status
=====================================================

scsi 2 .0.1 s imO ok(Ox7071 ) ok

disk 2 .0.1 .0.0 scsi ok(Ox5800202 ) ok
disk 2 .0. 1 . 2 .0 scsi ok (Ox101 ) ok
disk 2 .0.1 .4.0 scsi ok(Ox800100) ok
disk 2 .0 . 1 . 5 .0 scsi ok(Ox202) ok
lan 2 .0.2 lan01 ok(Ox7072) ok
hil 2 .0.3 hil ok(Ox7073) ok
serial 2.0.4 asioO ok(Ox7075) ok
serial 2 .0.5 asioO ok(Ox7075) ok
parallel 2 .0.6 parallel ok(Ox7074) ok
scsi 4.1.0 s imO ok (Ox7680) ok
disk 4.1 .0.6.0 scsi ok(Ox201 ) ok

Multiple disks attached to the SCSI interface ( hardware path 2 . 0 . 1 ) are

differentiated by the target number ( for example, the 4 in 2 . 0 . 1 . 4 . 0) , and
all the SCSI disks end in a terminating 0. The disk addressed 4 . 1 . 0 . 6 . 0 is
attached to an EISA host adapter at slot 1 .
Hardware-path numbering i s directly related to minor numbers ; compare the
information in "Minor Numbers on the Series 700," in Chapter 1 1 , "System
Configuration" .

1 0- 1 4 System Architectures
 10

Architecture of the Series 800 Bus

The Series 800 is built on a hierarchy of buses.
Buses are circuitry-the pattern of conductors (traces or wires) on a circuit
board-through which data travels between hardware modules (CPU, memory,
channel adapters, device adapters) within the system. Cards are plugged into
buses.
The number of conductors used for data-the bandwidth-equals the number
of bits of the bus. The fastest bus installed is used for data transfer between
the CPU and memory. In the Models 8x7, private buses separate CPU and
memory transactions from I/O , thus increasing system responsiveness, even
for I/0-intensive computation. Among the I/O buses, HP Precision Bus
(HP-PB) and CTB bus have bandwidths 32 bits wide; the CIO bus has a 16-bit
bandwidth. Typically too, interface cards for graphics , communication, and
factory-floor devices plug directly into the faster buses.
The Series 800 supports two overall bus architectures (HP-PB and CIO ) , with
several variations. The CIO architecture has two implementations: a two-tier
I/O bus structure (consisting of CIO and Mid-Bus) and a three-tier I/O bus
structure (consisting of CIO , Mid-Bus, and SMB ) . The HP-PB architecture is
used in numerous implementations, HP-PB alone and with private (processor)
buses, and HP-PB with PMB . Table 10-2 shows the buses present in the
various systems.

Table 1 0-2. Buses Supported on Series 800 Models

Acronym Bus Series 800

Models
HP-PB Hewlett-Packard Precision Bus 808, 815, 8x2, 8x7, 890
CIO Channel Input / Output 825 , 834, 835, 840, 845 , 850 , 855,
860 , 865 , 870
SMB System Main Bus 850 , 855 , 860 , 865, 870
CTB Mid-Bus or Central Bus 825 , 834, 835, 840 , 845 , 850, 855,
860, 865, 870
PMB Processor-Memory Bus 890

System Architectures 10·15

10

Depending on model, one or more of each bus i s present . Thus ,

• Models 808, 815, and 8x2 employ a single bus , HP-P B . The HP-PB provides
direct connection between processor and memory modules and I/0 cards .
• Models 8x7 employ the HP-PB bus for I/O and another set of busses for
processor and memory transactions. Processor and memory communicate via
subsidiary busses to a rnernory/I/O controller, which connects via the VSC
bus to HP-PB .
• Models 825, 834, 835, 845, and 840 use a two-tier I/O bus structure of CIO
bus and Mid-Bus. Models 825, 834, 835, and 845 can also be configured with
an HP-PB converter for MAP applications.
• Models 850, 855, and 870 use a three-tier 1/0 bus structure of CIO bus ,
Mid-Bus, and SMB .
• Model 890 uses a two-tier bus structure: PMB for processor(s) and memory,
and HP-PB for 1/0 .
Figures in corning pages illustrate these configurations.
Hardware modules-bus connectors, channel adapters, and device adapters­
connect to these buses, and external peripherals attach to the device adapters .
In the following section , each hardware module is discussed in turn, as well as
·

its addressing.

10-16 System Architectures

 10

Hardware Modules Connecting to the Buses

Hardware modules (bus converters, channel adapters, and device adapters)
are cards or boards, inserted into slots in the system's card cage to connect
to the bus architecture. The modules handle queueing, signal levels, and
protocol when transferring data between the CPU or memory modules and
specific devices (such as I/O and graphics) with distinctive signal needs. Once
configured, transfer of data among bus structures is transparent in almost
all respects. Each slot in the card cage is identified by a number or letter for
hardware addressing.
The following hardware modules serve to connect buses or buses and devices:
Bus converter A board that serves as an internal interface between higher­
and lower-speed buses, typically the SMB and Mid-Bus.
Because more than one bus converter can be connected to the
SMB, bus converters can increase the number of modules in a
system, thus enabling more devices to be connected.
The Series 890 is equipped with dual bus converters: The top
bus converter card resides in slots of the PMB (processor
memory bus) and is attached by ribbon cable to one or two
lower bus converters residing in HP-PB buses. (Each upper bus
converter can actually provide bus conversion to two HP-PB
buses.
Channel 1/0 A module that serves as an internal interface between the
adapter Mid-Bus and the lower-speed CIO bus. The channel adapter
also manages direct transfer of data between I/O modules and
main memory with minimal CPU intervention. More than
one channel adapter can be installed on most systems, thus
expanding the number of possible modules .
Device Interface cards such as HP-IB , HP-FL , SCSI, LAN , and MUX
adapter cards, installed in the I/O slots of the CIO or HP-PB bus .
Device adapters m ake the crucial link that enables the system
to communicate with external peripherals (such as terminals,
printers, disk drives) . Device adapters are also called I/O cards
and host adapter cards. CIO device adapters are also called
CIO cards.

System Architectures 10-17

10

Device adapters provide the data-transfer capabilities between

the CIO or HP-PB bus and external devices (peripherals) .
Because of this, device adapters are the hardware modules
most frequently configured by system administrators .
Common I/O interfaces include:
• HP-PB host adapters for HP-IB , HP-FL, and SCSI.
• HP-PB GPIO (general purpose parallel I/O ) , used for
interfacing 8 or 16-bit data buses to the system. (Not
supported on Models 8x7.)
• CIO multi-device interfaces for HP-IB , HP-FL, MUX, or
S CSI.
• CIO system-to-system communication, such as LAN
interfaces via IEEE 802.3 or Ethernet connection.
• I/O real-time interfaces , for faster response time than host
systems routinely provide.
• Mid-Bus graphics interfaces, for interfacing display
controllers to the system.
• Mid-Bus system-to-system communications interfaces, for
communicating with X.25 packet-switching networks, other
vendors' systems, or factory-floor devices.
For a complete listing of available interfaces , consult the
Configuration Guide for your series.
SCSI, HP-FL , and HP-IB interfaces accommodate multiple devices per
interface card; the devices are cabled together. With S C SI, you must cap the
final device with a terminator plug. SCSI devices require an additional layer of
addressing (described in the next section) . If only one SCSI device is attached,
the final character of the hardware address is 0.
When configuring HP-IB and other multiple-access interfaces, consider
electrical characteristics, as well as what priority peripherals should have for
accessing resources.
You need to understand major and minor numbers to configure HP-UX for
a device. Refer to Chapter 1 1 , "System Configuration" and to the System
Administration Tasks manual for your system. Refer to the Installing
Peripherals for complete installation and configuration instructions, including
device-driver information.

1 0- 1 8 System Architectures
 10

Addressing on Series 800

Data is transferred through a sequence of each bus architecture's addressable
elements, including bus location, bus converters, device adapters, and external
devices.
At boot-up, the system reads the hardware path of the root disk as a sequence
of addressable elements in the boot path from the CPU leading to the root disk
device.
The same principles apply for any configured peripheral: The system recognizes
the sequence of addressable elements in the I/O path from the CPU to the
device.
Each addressable element is represented by a module number in the hardware
path. The module number on both Mid-Bus and HP-PB bus is the actual slot
number on the bus into which the card or board is plugged, multiplied by four.
The multiplier of four is used to accommodate boards with multiple modules.
On the CIO bus, the module number is the slot number.
In subsequent sections, addressing examples are given with each sample
implementation. The addressable elements are symbolized as follows:
• b = bus converter address (PMB to HP-PB or SMB to Mid-Bus)
• m = module number (PMB , SMB , or Mid-Bus slot number times four; or
HP-PB slot number times four plus unit number)
• c = slot address of a CIO interface card
• d = device switch setting
• 1 = logical unit number

Note that S C SI requires an additional "logical unit number" of 0 appended

to the hardware address; if device is not SCSI, omit it . See insf ( lm) in the
HP- UX Reference.
The degree of addressing complexity depends on Series 800 model. The most
complex example follows the pattern b/m . c . d . 1 for the Model 850 family and
Model 890. HP-PB Models 8x2 and 8x7 are simpler to address, following the
pattern m . d or m . d . l .

System Architectures 10-19

10

HP-PB Implementations-Models 8x2

The HP Precision Bus (HP-PB ) enables data transfer through Models 808,
815, and 8x2 via a single bus . CPU board (including processor and optional
· :floating-point processor), memory modules , and all device adapters attach
directly to the HP-PB.
Figure 10-7 depicts the card-cage layout for a Model 8228 system, as seen from
the rear:

FRONT VIEW REAR VIEW

(Not User Accessible) (User Accessible)
....
�� ...l..!L.... MAP 1 (optional)
I ll MA P 0 (optional)
� rn LA N O (optional)
-
� � rn M UX O
�� � � � II � � rn M UX 1
� � � rn
i HP-J B (optional)

f ff
� �
0 0 N C") ..,. IO
0
.,... ..... .,... .,... .,...

z
m
a. le le le le

II � I
Memory

11 � I
Memory
H P-IB
(optional*)

�
Memory .....
(optional) .....
m � &I � :il
a.
Iii m
a. a. a.
.,...
m
a.
a. a. a.

*Standard On 832S

LG200 1 80_001J

Figure 1 0 -7. Sample Model 8x2S Card-Cage Layout

1 0·20 System Architectures

 10

Figure 10-8 depicts the bus architecture and hardware modules configured in
the Model 8x2S :

console other terminals printers

H P-I B LAN Map M ap

H P-I B
(opt) (opt) (opt) (opt)

Hard Disk Drives

DDS Tape Drive

LG2001 72_033a

Figure 1 0 -8. Model 8x2S bus architecture

Addressing on HP-PB Models

I/0 interface cards plug directly into the Precision Bus , making hardware
addressing relatively simple. The hardware address (m . d) consists of two
elements:

System Architectures 1 0-2 1

10

• m = module number (HP-PB slot number times four)

• d = device switch setting
The following are found in standard configurations:
• Slot 1 1 (module 44) is used for an embedded hard disk drive and DDS tape
drive. Thus, a DDS tape drive with switch set to 7, connected to the HP-IB
interface at slot 1 1 has the address 44 . 7 .
• Slot 12 ( module 48) i s used for attaching external HP-IB devices .
• The console must be attached to the first port on the first MUX (in this case
slot 4, module 16); each MUX can provide from five to sixteen ports and can
have other serial devices (such as terminals and printers) attached.
• If installed, a LAN card occupies slot 13 (module 52) by default .
• When present , Manufacturing Automation Protocol (MAP) cards are usually
installed in slots 14 and 15 (modules 56 and 60) for control of factory-floor
devices.
For more information about configuring HP-PB models, refer to the Installing
Peripherals and System Administration Tasks manuals.

1 0·22 System Architectures

 10

A Model 8x7 I mplementation

The Model 8x7 bus architecture consists of a bus architecture similar to the
Model 8x2 family, but with several refinements, including faster processor and
memory access. Figure 10-9 shows the backplane of a Model 8578/8778/8978 .

..-
..- (') in I'- en
..-
0 0 0 0 0 0
Ci5 .....
"O
Ci5 Ci5 Ci5 Ci5 Ci5 tll
(.)
� 2:-
::::> 0
'ffi a_ E
c: (.) Q)
0 C\I 0
C\I '<:!" co 00 ..- ..- �
.....
CJ)

0 0 0 0 0 0 Q)
a_
Ci5 Ci5 Ci5 Ci5 Ci5 Ci5

Ll320021 1 _037

Figure 1 0-9. A typical backplane of a high-end Model 8x7S

System Architectures 1 0-23

10

Figure 10-10 shows the bus architecture of a typical Model 8x7. Note that
processor and memory interactions are segregated from the HP-PB bus,
minimizing 1/0 contention.

8X7 Family

P rocessor .,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,.,. M�:�i'�(O .,.,.,. . . . .. . . :·:·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·=·= M emory

!) VSC Bus

� .-.-
I I I I I I I
1 1 '0 110 1 '0 110

LG20021 1 _0 1 4

Figure 1 0· 1 0. M odel 8x7S Bus Architecture

1 0-24 System Architectures

 10

In Figure 10- 1 1 , you can see the module numbers used to address each element
of the bus architecture.

Personality Card

_______,
_o 4--�
3--�
2--� 1 _____
a--�
5--� a 9 0�
..- 1_
..� 3-_
1 1__1�2_l_1� 1�l s ......
4-l_1_
H
:��::n
x
-0 ::>
w ::> P rivate Bus

(/) ::E 0
a.

Memory

H P-PB Module#=4 (slot #)

SCSI Address -

LG20021 1 _036

Figure 1 0 - 1 1 . Addresssing the Modules of a Model 8x7S

System Architectures 1 0-25

10

Addressing Models 8x7

Addressing the Model 8x7 is virtually identical to that of the Model 8x2, in
that I/O interface cards plug directly into the Precision Bus . The hardware
address (m . d) consists of two elements:
• m = module number (HP-PB slot number times four)
• d = device switch setting
If a S CSI device ( s ) is attached, append an extra 0 (m . d . s ) .
Models 8 x 7 feature multi-functional "personality" cards whose addressing
deserves some attention. For example, the standard personality card consists
of 8 to 16 (optional) RS-232 ports, console/access port , parallel/Centronics
port , and S CSI port . As shown in Figure 10- 1 1 , the personality card is plugged
into HP-PB slots 13 and 14 (modules 52 and 56) . Module number 52 has two
units-0 for SCSI and 1 for Centronics . Module number 56 is used for MUX
connections. The following addresses are derived:
52 . 6 . 0 SCSI root disk
53 . 0 Centronics printer (not supported on Model 807)
56 . 0 console

1 0-26 System Architectures

 10

A Mid-Bus/CIO Implementation-Model 845SE

The two-tier bus architecture of the Model 845 (Figure 10- 12 and Figure 10- 13)
is representative of a family of systems that include Models 825S, 835S/SE, and
845S/SE.

Addressing o n Mid-Bus/CIO Models

Hardware addressing on Mid-Bus models typically requires a three-part

hardware address (m . c . d) to take into account multiple bus structures of
Mid-Bus and CIO bus.
• m = module number (Mid-Bus slot number times four) of the channel adapter
• c = slot address of the interface card in the CIO bus
• d = device switch setting
If a SCSI device ( s ) is attached, append an extra 0 (m . c . d . s) .
Boards on the Model 845 plugged directly into the Mid-Bus are identified by
module number, Mid-Bus slot times four. Thus, a memory board connected at
slot 8 (default position) has a module number of 32 .
Addressing the processor-dependent hardware (PDH) is slightly different , for
two reasons: the board occupies two mid-bus slots (2 and 10) , and has two
channel adapters (one for each CIO bus-one internal and one external) . Each
channel adapter is addressed separately and may have module number 8 or 40,
depending to which CIO bus the channel adapter refers.
Another addressing variation occurs on Model 835, where the channel adapter
chip set is located directly on the system card, plugged into Mid-Bus slot 1. In
this case, the channel adapter chip set has a module number 4.
To address peripherals (attached to CIO interface cards) requires addressing
the channel adapter module, the interface card at its CIO slot , and the device
setting itself. In Figure 10- 13 , a disk connected at the first port (port 0) of the
HP-IB at CIO slot 1 is addressed 8 . 1 . 0 .

System Architectures 1 0-27

10

Factory-Floor Communication
An alternative configuration enables Mid-Bus models to handle computer-aided
manufacturing. The HP Precision Bus ( HP-PB ) is added to a Mid-Bus system
through an HP-PB bus converter. This enables installation of Manufacturing
Automation Protocol ( MAP ) cards (IEEE 802.4) for control of factory-floor
devices . Figure 10- 1 2 depicts the card-cage layout of a Model 845SE, with

[J
attached CIO expander.

0�
Access Port

:�:��::
L
Ope
Open for H P-18
ao
��w 32 MB Memory Board
Open for memory expansion

... - - - -
Mid-bus slots

�I
0 1 im== - - ·- -
.
8 32
1 :m:= 7
�
- 28
2 � 6 24
3 5 20
4 4 16 Module#
5 3 12
2/1 0 8140
mm rmlll
� 1 4
-
0 0
- ....._
- -

Power Supply
- -

Open for LAN - - PDH wilh lntemal & extemal

channel adapter

[j
Open for MUX- - Processor Board

Avabble CIO ....

Buffer Card
CIO slot #
-
0
1 �
2 ..
3 ..
4
5
6
To CIO Expander -
7
8 -

Power Supply
I

LG200 1 72_036

1 0-28 System �fle!'lo?� Sample Model 845 card-cage layout

 10

Figure 10- 13 depicts the bus architecture and hardware modules configured in
the Model 845 system:

open for
memory expansion
...-���....,A-���-
( \

Channel
Adapter

CIC-Bus

HP-B
Access
or HP-IB MUX LAN MUX
Port
HP-FL

Console avai lable 1/0

slots

LG2001 72_037

Figure 1 0- 1 3. Model 845 Bus Architecture

System Architectures 1 0-29

10

A n SMB Implementation-Model 850S

Models 850S , 855S , 860 S , 865S, and 870S are large systems built on a
hierarchy of three buses-SMB , Mid-Bus, and CIO . This arrangement enables
high-capacity operation via numerous devices.
At 9.0, Model 870S can be equipped with up to four processors .
Figure 10- 14 depicts the card-cage layout of a Model 850S .

Addressing on SMB Models

Hardware addressing on SMB models must descend through all the elements in
the architectural hierarchy. Thus,
• Devices attached to the Mid-Bus are addressed by specifying the bus
converter between SMB and Mid-Bus and module into Mid-Bus.
• Devices attached to the CIO bus are addressed by specifying the bus
converter between SMB and Mid-Bus, channel adapter between Mid-Bus and
CIO , CIO card slot , and device setting.
The hardware address of a device configured into an SMB system follows the
pattern b/m . c . d, which corresponds to:
• b = bus converter address ( SMB to Mid-Bus-either 2 for bus converter 0 or
6 for bus converter 1 )
• m = module number ( Mid-Bus slot number times four ) of the bus adapter
• c = interface card ( the actual CIO slot number )
• d = device switch ( set on the device itself )
Figure 10- 1 5 depicts the bus architecture and hardware modules configured in
the Model 850S system. For example, the tape drive connected to an HP-IB
port has the address 2/4 . 2 . 3 .

1 0·30 System Architectures

 10

M odel 850S Famlly Computers

Channel Adapter . . . . . Chan nel Adapter

for CIO Slots 0-4 l ff I H for Cl 0 Slots 0-4
(Upper Left-Hand (Upper Right­
Card Cage)

11 11111
Hand Card Cage)

Chan nel Adapter

Channel Adapter for Cl 0 Slots 0-4
for CIO Slots 0-4 (Lower Right­
(Lower Left-Hand Hand Card Cage)
Card Cage) 01 2 3 4 01 234
.
UH H P-IB
U UU 1 1 (Root Disk)

�� 1 1 L
or H P-FL

L
MUX

•r[mL
(Console)

�
u I I

01234 •
. . .
1 234 5 6
� �
L
- '"-
-
. .
6 5 43 21
01 234
ii
H P-I B
LAN
Access
Port
Hardware Path :6 - H
ardware Path = 2
Memory Controller

CIO (Top) CIO (Top)

Slot 0 = 6/8.0* Slot 0 = 218.0
Slot 1 = 6/8. 1 Slot 1 = 2/8.1
Slot 2 = 6/8.2 Slot 2 = 2/8.2
Slot 3 = 6/8.3 Slot 3 = 2/8.3
Slot 4 = 6/8.4 Slot 4 = 2/8.4

CIO (Bottom) Mldbus 1 System Mldbua o CIO (Bottom)

Slot O = 6/4.0 Slot 1 = 6/4 MalnBus Slot 1 = 2/4 Slot 0 = 2/4.0
Slot 1 = 6/4. 1 Slot 2 = 618 Slot 2 = 2/8 Slot 1 = 2/4.1
Slot 2 = 6/4.2 Slot 3 = 6/1 2 Slot 3 = 2/1 2 Slot 2 = 2/4 .2
Slot 3 = 6/4.3 Slot 4 = 6/1 6 Slot 4 = 2/1 6 Slot 3 = 2/4.3
Slot 4 = 6/4.4 Slot 5 • 6/20 Slot 5 = 2/20 Slot 4 = 2/4.4

Hardware Paths

*Slot # = Hardware Path

LG2001 72_038

Figure 1 0- 1 4. Sample Model 850S Card-Cage Layout

System Architectures 1 0-3 1

10

====� (cont'd
===�\
� below)

15
· · · �« . . ':!:'
. .

• • •
80

slot # slot #
--
CIO CIO
8/4.0 6/8.0
1hru 8/4.4 1hru 6/8.4

dlsk drlw

�
tape drlw

HP-18 address
2/4.0.0
HP-t8 addres
2/4.0. 1
�
HP-tB add res
2/4.2.3
HP-18 address
2/4.2.1
LG2001 72_039

Figure 1 0 - 1 5. Model 850 bus architecture

1 0-32 System Architectures

 10

Processor-Memory Bus (PMB) Implementation-Model 890

The Series 9000 Model 890 Corporate Business Server is the high-end HP-UX
computer. Architecturally, its organization is hierarchical, like SMB systems,
but using the HP-PB (instead of CIO ) bus.
• Like the Models 870 and 880, the cardcage of the Model 890 can be outfitted
with up to four processors , to perform multi-processing. Configuration of the
multiple processors should be handled by an HP customer engineer.
• Like the Models 8x7, memory communicates with the processor( s) via a bus
structure separate from I/ 0. This ensures that extremely fast processing
time, even on 1/0-intensive systems .

Physical Layout of the Model 890

As shown in its simplest form, in Figure 10- 16, a Model 890 consists of two
card cages (PMB and HP-PB) stacked atop one another. The upper PMB
cardcage is further divided into front and back by a backplane. Processor
and memory boards are plugged into the front area; additional processor and
memory boards can be plugged into the back if necessary.
The lower card cage is an HP-PB bus, which can accommodate the family of
HP- PB devices . Device adapters might be half-height or full-height cards,
each of which can contain up to four modules . A minimal configuration might
consist of a SCSI or HP-FL device adapter for disks and tapes, MUX cards ,
and data communication boards. Peripherals can be plugged to each board, as
with any HP-PB system.

System Architectures 1 0-33

10

F ront Rear

1 0 11 12 1314 15 SP 0 1 2 3 4 5 6 7 8 9

Jj
�
....
0
�
Ill

8� 0
Power
-� � E
Gl
ll..
Supply JJ 0 ::il

Fans Modules
14 12 10 8 6 4 2 0

1�
....
Gl .lo:

i
E
8

�
� f/J C\I
a ...!.
�
.9 - �
� f/J ::il

1 5 13 11 9 7 5 3 1

Fans

LG20021 1 _039

Figure 1 0 - 1 6. Model 890 Card Cage

Figure 10- 1 7 shows the bus architecture of the Model 890.

1 0-34 System Architectures

 10

Peri pherals

Service
Processor
To HP-PB
Cardcages
Upper BCs in
L.. L.. .J Expansion
Cabinets

- - - - 1 , - - - - - - - 1
I
I I
I

I
I I
I
I
I
I
==���==s=�P=' I
I Lowerl
I
I BC I
I
I I
I :;:::;;:::
:om �
::::; t=;;�

:G G G!
I

L _ _ _ _ _ _ _ _ _J

H P-PB Cardcage

SPU Cabi net Expansion Cabinet

LG200204_01 1 a

Figure 1 0- 1 7. Bus Architecture of the Model 890

Communication between the PMB and HP-PB buses is handled by Dual Bus
Converter boards, which occupy back slots of the upper cardcage. Model 890
bus converters differ from other systems: two boards perform the conversion
from SPU bus to HP-PB bus : a "Du al Bus Converter" board, plugged into
the SPU bus for communication with processors and memory, and a "Lower

System Architectures 1 0-35

10

Bus Converter" board, plugged into the HP-PB bus, for communication with
each I/O card cage. As the name implies, each Dual Bus Converter board
actually consists of two upper bus converters , which are addressed separately
and can provide communication to two Lower Bus Converter boards (that is,
two HP-PB card cages) . Each bus converter set connects between the PMB
and HP-PB buses, using two ribbon cables to accommodate high and low
differential voltage.
In addition to the basic HP-PB bus, ribbon cables can connect from Dual Bus
Converter boards to external (expander bay) HP-PB card cages .

Addressing on a Model 890

Addressing on a Model 890 is consistent with other HP-PB systems. Bus
converters and HP-PB cards are numbered as modules of the slot into which
they are plugged (that is, slot times four) . Overall, every hardware path for
components on the Model 890 conforms to the following convention:
bus_converter I HP-PB_card . port _number
Because the Dual Bus Converter can provide connectivity from PMB to two
HP-PB buses, the lower Dual Bus Converter board is numbered:
slot number t imes four
The upper Dual Bus Converter boards is numbered with an offset :
slot number times four plus two
HP-PB cards on the Model 890 are addressed much like Models 8x2 or 8x7.
Because every slot might accommodate a card having up to four functions , the
address of a card in a slot is numbered:
slot number t imes four plus off s et
The offset is determined by the function and should be documented in the
installation guide provided with the card.
Finally, the port number of the HP-PB card is based on a setting on the device
itself.
If a S CSI device is attached, a zero is appended to the address. Similarly,
HP-FL devices may have an additional number, based on

1 0-36 System Architectures

 10

Model 890 Guidelines

• In powering on the system, the Model 890 requires that you power on all
external I/O bays before the SPU (PMB cardcage) . There are two reasons
for this:
o so that the SPU can find all devices when it configures the system, and
o to avoid a race condition that would occur if this powering-on order is not
observed.
• The Model 890 features new interfaces that can be accessed from various
system states. These include:
o Service Processor ( S P ) . Unique to this system.
o Processor Dependent Code (PD C ) .
o Initial System Loader (ISL).
o Access Port ( AP) .

In addition, new key sequences have been implemented for TC (transfer of

control) and RS (reset) .
For information on these interfaces, see HP3000 Corporate B,usiness Systems
and HP9000 Corporate Business Server Operator's Guide: Series 990/992
and Model 890, HP part number A1809-90009.

System Architectures 1 0-37

10

Finding Out About the Current System Configuration

Two tools are available to find out what hardware is configured into your
system: I etc/ dmesg and I etc/ ioscan.

Viewing Hardware Configuration Using /etc/dmesg

Root users can use the command I etc/ dmesg to see what is configured on the
system. /etc/dmesg displays the contents of an error-message buffer printed by
the kernel when it begins to boot-up . While loading, the kernel again checks
the buffer for existing hardware. The buffer is filled "FIFO" ( first in first out ) ;
thus , later messages can overwrite earlier messages.
In the Series 300 and 400, /etc/dmesg shows what cards are present, identified
by select code. For example:
Init ializing hidden mode int er.faces
Hidden mode init ializat ion complete

real mem = 8376320

us ing 166 bu££ers containing 679936 byt es 0£ memory
Root device maj or is 0 , minor is OxeOOOO , root s it e is 0
-- BATTERY BACKED REAL TIME CLOCK
Swap device t able : (start t size given in 512-byt e unit s )
entry 0 - aut o-con£igured o n root device ; start = 528444 , size = 65772
avail mem = 5177344
lockable mem = 983040

1 0-38 System Architectures

 10

In the Series 700, I etc/ dmesg shows system configuration by select code and
function number. It also shows information on the memory subsystem and
LAN.
Kar 28 1 5 : 32
Beginn ing I/D System Con£igurat ion .
Block TLB ent #5 £rom 0Jd:9000000 t o Ox£9££££££ a11ocated .

CDISDLE is on the ITE

I/D System Con£igurat ion complet e .
Con£igure called
Root device maj or is 7 , minor is Ox201000 , root site is 0
Svap device table : (start i size given in 512-byt e units)
entry 0 - aut o-con£igured on root device ;
start = 1218000 , size = 84670
Warning : unable to con£igure d'Dlllp device . . . using primary svap instead .
core image 0 £ 4096 pages vill b e saved at block 634951 on device Ox7201000
@ (#) B2352A HP-UX (sys . A . 08 . 05J) #1 : Wed Kar 27 23 : 09 : 07 KST 1991
physical page size = 4lb
real mem = 16777216
lockable mem 7806976
=

avail mem 9359360

=

using 404 bu££ers containing 3313664 byt es 0£ memory

lanO : AppleTalk node 9

System Architectures 1 0-39

10

I n the Series 800 , / etc/dmesg shows system configuration b y hardware path

and device driver. It also shows what subsystems and which type of file
systems are initialized. For example:

I/O Syst em Configurat ion complet e .

Configure called
Beginn ing Subsyst em Init ializat ion
nipc init ialized
dskless init ialized
inet init ialized
lan init ialized
uipc init ialized
Subsystem Init ializat ion Complete
Beginning Filesyst em Init ializat ion
ufs init ialized
cdfs init ialized
nfs init ialized
Filesyst em Init ializat ion Complete
Root device maj or is 10 , minor is Ox10d , root site is 0
real mem = 16777216
lockable mem 7342080
=

avail mem 9807872

=

using 609 buf£ers cont aining 331 571 2 byt es 0£ memory

For more information, see dmesg(lm) in the HP- UX Reference .

1 0-40 System Architectures

 10

Interpreting Hardware Paths on a Series 800

As you can see from Series 800 /etc/dmesg output , a hardware path is
a numerical sequence of hardware addresses from bus to external device.
(The same principle is true for Series 700 , although not apparent from its
/etc/dmesg output .) Data reaches the external device via a software module
path-a sequence of programs corresponding to the hardware path and leading
to a device driver that controls the external device.
The following table gives sample hardware paths and their corresponding
software module paths:

Table 1 0-3. Sample Hardware Paths

Hardware Software Module Path Description

Path to Device Driver

16.0 hpib1 . disc1 an HP-IB disk with bus address 0

attached to slot 4 of a precision bus on

a Model 808.

4.4 cio_caO . lanO a LAN interface card attached to a

CIO bus via a channel adapter
( cio_caO ) on a Model 825.

2/4.0.1 bus_converter/cio_ca0 . hp£10 . disc2 a disk with bus address 1 attached to

an HP-FL interface card plugged into
slot 0 of a CIO channel adapter,
connected to Mid-Bus slot 1 of a bus
converter on a Model 850.

Viewing Hardware Configuration Using ioscan

The command /etc/ioscan displays devices on your system by the following
major characteristics:
• Hardware path
• Device drivers
• Class of device
• Logical unit number (Series 800 only)
A sample full I/O listing ( ioscan -fu) of a Model 890 shows the chain of
device drivers that correspond to the hardware-path elements:

System Architectures 1 0·4 1

10

Class LU H/W Path Driver

==============================================================

tty 0 0/28 bus_converter .mux2

disk 6 0/32 . 0 . 0 bus_convert er . hp£11 . t arget . disc4
disk 7 0/32 . 0 . 1 bus_convert er . hp£11 . t arget . disc4
tty 3 0/40 . 0 bus_converter . 18Illlul x0 . mux4
lan 0 0/40 . 1 bus_converter . 18Illlul x0 . lan3
lantty 0 0/40 . 2 bus_converter . 18Illlu
l x0 . lantty0
tape_drive 0 0/52 . 1 . 0 bus_converter . scsi1 . t arget . t ape2
disk 8 0/52 . 2 . 0 bus_converter . scsi1 . t arget . disc3
printer 1 0/54 bus_convert er . lpr2
disk 12 2/0 . 0 bus_convert er . hpib1 . disc1
disk 13 2/0 . 1 bus_convert er . hpib1 . disc1
disk 14 2/0 . 2 bus_convert er . hpib1 . disc1
disk 15 2/0 . 3 bus_converter . hpib1 . disc1
printer 0 2/1 bus_convert er . lpr2

Note some characteristic differences among device-driver interfaces, which show

up in the ioscan output:
• HP-IB device and unit have a 1 : 1 correspondence.
• SCSI and HP-FL device and units have a 1 : many correspondence. For
example, there are two disks in the disk array addressed as 0/32 . 0 . 0 and
0/32 . 0 . 1 .
ioscan has numerous output options; Series 300/400/700 output is more
limited than that of Series 800. See ioscan(lM ) in the HP- UX Reference for
full specifications .

1 0·42 System Architectures

 11

11
System Configuration

What Does Configuring Mean?

When you configure your system, you inform the operating system what
hardware is present by associating the hardware with necessary software­
device drivers and device files. You might also have to set values or limits­
configurable operating-system parameters-for optimal system performance.
Configuring your system is important when you initially install it, when you
make certain major changes to HP-UX, or when you add a device and the
driver is not present in the currently running kernel. For example, when you
add a printer, not only must the peripheral be physically attached, but the
executable code (the device driver, such as lpr1 ) and the printer device file
(such as /dev/printer) must also be present . Or, a new application might
need more data space, which would require adjusting parameters.
Once the necessary files are created or included in the configuration file, you
must regenerate the HP-UX kernel to activate the change. Much configuration
occurs automatically, when you reboot the system. Sometimes, however,
you must override the defaults or attach a peripheral device that cannot be
configured automatically . For those instances, you use SAM menus (or HP-UX
commands) and reboot the system to enable the configuration.
This chapter describes the software elements used by the operating system to
activate your system configuration.

System Connguration 1 1-1

11

Understanding the Role of the Kernel

The kernel is the software core of HP- UX and is the part of the operating
system that deals directly with the hardware. The kernel also manages the
low-level functions for the operating system and insulates the operating system
from the details of the hardware.
As a system administrator, you identify devices on a hardware level by their
physical address. The kernel recognizes any hardware by two elements-device
drivers and information found in the device files. For the kernel to find and use
any class of hardware, the appropriate device drivers �ust be included in the
kernel and the appropriate device files must be installed in the file system.
Figure 1 1- 1 shows the kernel in relationship to the entire HP-UX environment :

user environment
(device Independent)

HP-UX Kemel
(Including device files
& device drivers)

s>:
dr vers

1/0 Devices

LG200172_040

Figure 1 1-1. Kernel in Relation to the User Environment and 1/0 Devices.

1 1 -2 System Configuration
 11
Kernel Configuration Files
The df ile for Series 300/400/ 700 or S800 for Series 800 is the principal vehicle
from which the kernel derives its image of your system. df ile and S800 are
termed kernel configuration files . Although they differ in name and format,
df ile and S800 are both used to create the kernel for their respective systems;
both contain information vital to the kernel about device drivers , subsystems ,
swap configuration, and configurable parameters .
Every time you change the kernel configuration file, you have to generate a new
kernel and reboot the system to activate the changes in that file. Figure 1 1-2
compares what happens when you generate a new kernel on Series 300/400/700
and Series 800.

H I
/etc/m aster system libraries
t
Series
clfile config
conf. c � m ake h p-ux
300/400/700 ..._____,
config. m k ---+! ---

conf. c
config.h

:=-8
Serles 800 SBOO uxgen

�
libs file

� conf. o
mak file

LG20021 1 _041

Figure 11 · 2 Kernel Configuration Files.

.

System Configuration 11 · 3
11

dfile-Series 300/400/700 Kernel Configuration File

The df ile is the Series 300/400/700 configuration description file, used
to create a new or customized kernel with either config( lM) or SAM. By
convention, df ile is located in the /etc/conf directory. Several templates
reside in /etc/conf for both server and client configurations. Use the df ile to
specify configurable parameters and device drivers .
The df ile shown on the next page is written for both bit-mapped and remote
terminal consoles, and includes the cs80 , 98624, 98625, 98628, and 98642
drivers. It also specifies the default kernel for a diskless cluster node being
added to an HP-UX cluster, and is the template used by SAM if you specify
the option "Create a minimally loaded operating system."
The config( lM) manual page in the HP- UX Reference provides detailed
information on the component parts of the df ile and how to customize it for
your own system needs.
Note that the notation in the df ile beginning with an asterisk ( * ) or /* are
commentary, ignored by the kernel generation process.
For explicit procedures for generating a new kernel, refer to the System
Administration Tasks manual.

1 1 -4 System Configuration
 11
Here is a sample df ile for the Series 300/400:
* DEVICE DRIVERS
cs80 I• disk and tape drivers •/
scsi
tape
st ape

printer /• printer drivers •/

* pseudo t erminal drivers
ptymas
ptyslv

srm /• shared resource management driver •/

rj e /• remot e j ob entry •/
dskless /• Diskless code pseudo-driver •/

* Netvork File System (NFS ) and netvorking drivers

nfs
uipc
nipc
inet
netman
lla
lan01
ni
netdiag1
•cdfs /• CD File Syst em (CDFS ) •/
meas_sys I * Measurement Syst em driver •/

* CARDS
98624 /• HP-IB int erface •/
98626 /* high-speed HP-IB interface •/
98626 /• RS-232C serial int erface •/
98628 /* RS-232C datacomm interface •/
98642 /• RS-232C mult iplexer •/
98668 /• DID SCSI interface •/
98266 /• SCSI int erface •/

* SWAP CONFIGURATION
* CONFIGURABLE PARAMETERS
dskless_node 1
ndilbuff ers 1
npty 30
* Disable support for HP98248A (Float ing Po int Accelerator)
* by changing the 1 to a 0 in the folloving line .
fpa 1

System Configuration 1 1 ·5
11
Here is a sample df ile for a Series 700 system:
* Installed softvare drivers and I/O int erface cards
aut ox I • Magneto-Opt ical aut ochanger •/
aut och

* SCSI disk and DAT device drivers

scsi
scsit ape
scsifloppy
s imO

parallel !• Parallel int erface driver •/

* Netvork management
netman
nipc
uipc
inet
lan01
asioO
nfs /• Netvork File System (NFS) •/
cdf s /* CD File System (CDFS ) •/

* PTY pseudo drivers

ptyO
pty1

dconfig /• libIO Kernel int erface •/

dmem /• Memory Diagnostics •/

dskless /• Diskless cluster code pseudo driver •/

rdu /• Remote Svap for diskless •/

• I/O drivers
eisa
CharDrv
hpib
cs80
hshpib

• Tunable parameters
dskless_node 1
server_node 1

• Svap info
•svap auto /• Svap after fs on root device default . • /
•svap scsi 201 500 -1 /• Svap after fs . •/
•svap scsi 201400 0 /• Use the vhole disk for svap . •/

1 1 ·6 System Configuration
 11

SSOO file-Series 800 Kernel Configuration File

The S800 file (typically, /etc/ conf/gen/S800) is used by uxgen ( lM) or SAM
to generate a new kernel for Series 800 computers . Templates for each model
reside in / etc/conf/gen/t emplat es, with executable scripts used by regen
to customize a configuration file appropriate to your system. The S800 file,
structured differently than the df ile, contains :
• An include statement for the I etc/mast er file, which the kernel uses for
pointers to modules, device-driver implementation, subsystems , and default
operating-system parameters.
• Device drivers that govern 1/0 for subsystems and peripherals.
• Operating-system parameter values (see Appendix A of System
Administration Tasks manu al)
• Addresses of kernel devices (such as console and swap ) , which can be
specified as default , as sections , or as logical volumes. (The S800 example on
the next page shows an LVM implementation. a standard implementation
might show kernel devices as follows:
console on default ;
root on default sect ion 1 0 ;
svap on default sect ion 1 ;
d1D11p s on default sect ion 1 ;
• io statement , used to configure devices that cannot be bound automatically,
overrides the system's automatic configuration capabilities. (The io
statement is discussed in more detail later in this chapter.)
Refer to uxgen ( lM) in the HP- UX Reference for information on the component
parts of the S800 file and how to customize it for your own system needs. See
System Administration Tasks manual for procedures on generating a kernel.

System Configuration 1 1 -7
11
Here is a sample S800 file on a Model 837 running LVM:
#include "/etc/mast er"
include mirror ; /• subsyst em and driver include stat ements •/
include s1litch ;
include nipc ;
include lv ;
include lvm ;
include nfs ;
include lan ;
include ni ;
include nm '·
include inet ;
include uipc ;
include t arget ;
include t ape2 ;
include scsi1 ;
include mux2 ;
include memory ;
include lpr2 ;
include lan3 ;
include disc3 ;
nclude hpib1 ;

t imezone 480 ; I• set s system clock to minut es from GMT •/

console on mux2 at 56 ; /* kernel devices •/

root on lvol ;
dumps on lvol ;
s11ap on lvol ;

io {} /• io stat ement •/

1 1 -8 System Configuration
 11
The io Statement of the seoo File

Series 800 systems automatically bind device drivers included in the S800 file
to their appropriate device. This "autoconfiguration" applies to most device
drivers. You must use the io statement to configure drivers that cannot be
bound automatically into the kernel using SAM. This section describes the io
statement of the S800 file.
The syntax of the io statement is:
io {
addr�ss_specifi cation

address_specification is a set of one or more nested device-driver and address

specifications for the installed non-autoconfigurable buses, bus converters,
channel adapters, 1/0 cards, and peripheral devices.
The io statement contains one or more nested address_specification for each
level of:
• Bus converter (for example, plugged into SMB and Mid-Bus )
• Channel adapter ( plugged into the Mid-Bus on CIO machines or optionally
into the HP-PB bus )
• Device adapter manager
• Device manager
Hardware paths can be derived from information in the io statement of the
S800 file:
bus_convert er address 2 {
c io_caO address 4 {
hpibO address 2 {
instrO address O ;

In this example, the hardware path fo r an instrO device i s 2/4 . 2 . 0 , derived

by reading the number following "address" in each line of the nested io
statement leading to and including instrO.

System Configuration 1 1 ·9
11
The io Statement of a Mid-Bus/CIO Configuration.
Suppose you have a Model 835 containing:
• two HP-IB disk interface cards , each with four disks connected
• two MUX cards
• another HP-IB interface card with:
o two line printers
o two tape drives
o a plotter
• a LAN card
• a GPIO card
Since most devices listed are configured automatically, their device drivers are
referenced as include statements .
include cio_caO ;
include hpibO ;
include discl ;
include muxO ;
include lprO ;
include t apel ;
include instrO ;
include lanO ;
include gpioO ;

Two devices listed-plotter and GPIO card-cannot be configured

automatically ( see Table 1 1- 1 ) , and therefore must be specified in the io
statement , by device driver and address. The device drivers are defined
in I etc/master; to determine which driver to use, consult the Installing
Peripherals . The address is based on the module number of the channel
adapter board on the system bus or the slot number of the device adapter
board.

1 1 · 1 O System Configuration
 11
Here is an io statement for the plotter and GPIO card:
io {
c io_caO address 4 { address of CIO chann el adapter
hpibO ad.dress 2 { address of HP-IB card for tapes & printers
instrO address 7; address of p lotter
}
gpioO address 5; address of GPIO card
}
}

The first level of nesting in this io statement shows the CIO channel adapter,
which is used by both devices to communicate with the system bus and
processor:
cio_caO ad.dress 4 {
ifddress_specifi cations

cio_ caO The name of the controlling CAM ( channel adapter manager,
explained in the section, "Device Drivers" ) .
address 4 Physical location of the channel adapter module in slot 1 of
the Mid-Bus . Module boards may contain multiple modules ;
the address is set as slot number multiplied by 4 so that other
modules on the board can be addressed by offsets.
The addresses specified after the CIO channel adapter include all device
adapters inserted in the CIO Bus card cage that cannot be configured
automatically. The same pattern of driver name and address is used:
hp ibO The device adapter manager that controls the HP-IB card.
address 2 The physical slot of the HP-IB card in the CIO-Bus card cage.
instrO The device driver that controls the plotter. Because the plotter
is attached to the HP-IB card, instrO is specified within the
nested address of hpibO .
address 7 The bus address of the plotter card ( in this case, an HP-IB
address switch located on the plotter ) .
gp ioO The device driver that controls the general-purpose 1/0
( GPIO ) card.
address 5 The physical slot of the GPIO card in the CIO-Bus card cage.

System Configuration 1 1-1 1

11
The io Statement of an HP-PB Configuration.
To configure a plotter and GPIO card to a Model 857, your io statement might
look as follows:
io{
hpib1 address 44{
instrO address 31 ;
}
gpio1 address 48 ;
}

As in the CIO example, the HP-PB io statement consists of a nesting of device

drivers and addresses for each level of architecture, from the Precision Bus to
the peripheral device.
The module address 44 locates the HP-IB board at slot number 1 1 of the
Precision Bus; module address 48 locates the GPIO board at slot 12. As in the
CIO example, the plotter uses the instrO device driver. The plotter's HP-IB
address is set to 3 1 .

1 1·12 System Configuration

 11
When Do You Reconfigure the Kernel?
For most purposes , the kernel generated when you bring up the system will
serve satisfactorily. However, some circumstances might require you to modify
the kernel.
Simple changes are performed most easily using the SAM utility. ( See the
System Administration Tasks manuals for your specific system for instructions
on modifying the system. )
The following tasks require kernel configuration:
• Change the configuration of 1/0 cards and device drivers.
• Add or remove a type of peripheral device, such as a disk, tape drive, or
CD-ROM.
• Improve system performance.
• Change the swap configuration.
• Add or remove optional subsystems such as NFS , LAN , FDDI, or
diagnostics .
• Remove unused 1/0 drivers or subsystems to reduce kernel size.
• Modify tunable operating-system parameters.
To accomplish these tasks, you modify the kernel to include a necessary device
driver or subsystem, or change configurable operating-system parameters.
Modifying the kernel involves changing the df ile or 5800 file. The next
sections will examine the kinds of things you change when you reconfigure the
kernel.

System Configuration 1 1 - 1 3
11

Statements to Include Device Drivers

In both df ile and S800 file, you include statements for device drivers . In the
df ile , device drivers are simply listed. In the S800 file, the device drivers are
listed as include statements. Each listing corresponds to a device driver for
subsystem, I/O card, or peripheral device found on the system that can be
automatically configured. The kernel finds the source code for all included
device drivers by referencing through the /etc/master file. (The /etc/master
file is discussed later in this chapter.)

Kernel Devices
In the Series 300/400/ 700, kernel devices are found via the /etc/master file.
In the Series 800, kernel devices-console , root , swap , and dumps-are
cited separately in the S800 file to allow the user to modify them. Whenever
possible, they are liste_d with default hardware addresses. Defaults are derived
from processor-dependent code, as follows :
• The default console path is the address displayed during the boot process,
before you get an ISL prompt.
• The default for the root device is the place from which the kernel was loaded,
usually indicated by the primary boot path. Normally, the root disk defaults
to section 13 of the disk from which you loaded the kernel. If you are using
the Logical Volume Manager (LVM) , your root logical volume is in your
root volume group . (For information on implementing LVM , see Chapter 9,
"Logical Volume Manager" in this manual and "Managing Logical Volumes"
in System Administration Tasks manual.)
• Swap defaults to section 15 of the root disk or a logical volume in the root
volume group , if you are implementing LVM.
• The dump device defaults to primary swap .
Defaults can be changed by executing the hpux command with options at the
ISL prompt during installation. See hpuxboot(lM) in the HP- UX Reference
and Installing and Updating HP- UX.

1 1·14 System Configuration

 11
Kernel Dump Device
Your HP-UX system is very robust ; however, certain situations ( such as data
segmentation fault or illegal instructions ) can cause the system to panic. And
as the system goes down, the kernel writes ( dumps ) an image of the memory
core to the dump device ( s ) and reboots. (If the system hangs, you can generate
a core dump from the console, by entering "console mode" and using ( control-B )
and the tc, "Transfer of Control" command. )
During the reboot, the savecore code is executed by I et c/re . savecore
locates the core image and deposits a core file in the file system ( typically in
/tmp/ syscore ) . Once reboot is completed, the core file can be retrieved and a
programmer can debug the program using standard debugging tools , such as
adb(l).
Typically, core dumps write to primary swap . On workstations, this i s true
by default; on Series 800 systems the S800 file designation dumps on default
causes core dumps be written to primary swap .
This default is adequate, provided primary swap is large enough to
accommodate a "snapshot" of the entire real memory. If primary swap is
smaller than real memory and no other dump device is designated, only a
partial core image is saved when a system crashes.
If your system is configured with memory that exceeds the size of primary
swap , you can configure other dump options to ensure that a total core image
is saved if the system crashes.
On Series 300/400/700, you can use entire devices to share swap and dump
purposes. Use same syntax for dump as you use optionally for designating swap
in the df ile. Note, do not have a file system on this device, because there
is a small chance you might lose data. Also, you must have swap on the root
device, since /hp-we requires it to boot .
On the Series 800, you can specify multiple dumps devices in the S800 file.
These devices can supplement primary swap or substitute for primary swap . If
supplementing primary swap , the primary swap device must be the first dumps
device listed, or wcgen will fail. Both swaps and dumps devices can be specified
as disk sections, logical volumes, or even dedicated disks. On Series 800, the
dumps designation in the S800 file tells you the location of the core file.

System Configuration 1 1-15

11
For complete information about setting up dumps areas , consult Chapter 7,
"Managing Swap Space" of the System Administration Tasks manual for your
system. Also see savecore(lM) , config(lM) , swapon(lM), and for systems
implementing LVM, lvlnboot( lM), in the HP- UX Reference Manual.

Operating-System Parameters
Configurable operating-system parameters are values set in the df ile or S800
file to customize kernel characteristics . Knowledgeably setting them can
optimize performance, but proceed cautiously, because many parameters are
interrelated and imprudent changes might cause unpredictable results!
Operating-system parameters affect the following areas :
• inter-process communication
• swap space allocation
• memory management
• cluster system management
• process memory limits
• message mapping, queueing, and size characteristics
• file system buffer cache
• networking process scheduling and memory allocation
• maximum number of system-wide open files, file locks, text descriptors
Operating-system parameters and their dependencies are listed in Appendix A
of the System Administration Tasks manual.

1 1-16 System Configuration

 11

What Happens when you Regenerate the Kernel?

Series 300/400/700 systems execute config( lM) using information in df ile;
Series 800 systems execute uxgen( lM) using information in saoo .
For HP-UX to use an I/O device, the device must be physically connected
and turned on. As a system administrator, you identify I/ 0 devices by their
physical address, but the kernel identifies them by device driver and device
files. Then, to actually communicate with the I/O device, certain logical
( software ) associations must exist , including device driver, hardware address,
logi cal unit number, and device file.

Kernel Regeneration-the Series 800 Case

During the boot process, the kernel reads the saoo file to include device
drivers for all hardware components, to mount the root device, to open swap
devices, and to establish the dump device ( s ) . The kernel launches init , the
initialization process, which reads /etc/ inittab , performs any required
machine-dependent initialization, spawns processes, and calls ioinit . ioinit
reads a reference file called / etc/ ioconf ig, which maintains mapping
information of device driver and logical unit number to the hardware addresses
of devices. Figure 1 1-3 shows this sequence of events.
As system administrator, you cannot directly access /etc/ioconf ig , whcih
is maintained by the kernel. When you change the system configuration and
reboot, the system probes the hardware for all devices attached, and maps the
information into / etc/ ioconf ig.
Changing your system's configuration causes the kernel to probe the system
twice: The first time, the kernel scans the backplane, identifies all cards it
finds, initializes global interrupt-type counts, and determines the size of the
physical address space to be mapped into virtual space. Data structures inside
the kernel probe all hardware to determine what is hooked up ( configured ) into
the system.

System Configuration 1 1-17

11

System Boot

loads
(system i nitialization)
_. probe hardware and match drivers
to hardware
i nvokes

/etc/inittab

transfers data from /etc/ioconfig

Other system ioinit Kernel data
startup processes structures
(io-tree)
If
/
/
/
/
/ data
/
/
/
LU num bers /etc/ioconfig
(for new devices) creates (contains mapping of
device files to hardware)
device files
(for new devices)

LG20021 1 _040

Figure 1 1 -3. System Initialization by the Kernel

The kernel detects all buses, channel adapters , device adapters, and external
devices and binds all hardware address space pointers to their corresponding
software modules ( device drivers ) and device special files. This is done
by writing to the kernel's internal data structure, io_ tree, as well as to
/etc/ioconfig. Then, ioinit calls insf to assign logical unit numbers to
all newly configured devices and downloads the lu numbers into the kernel.
insf then creates device files for any new devices it finds, then updates
I etc/ ioconf ig.
Having completed the 1/0 mapping, init then creates a process for each
terminal on the system, thus establishing a multi-user environment .

1 1 · 1 8 System Configuration
 11
Figure 1 1-4 illustrates the auto-configuration process and shows user access to
the configuration after the boot process.

1/0 Configuration
/etc/ioconfig
During the Boot
Holds LU: hardware
Process
mapping
(user level)

Downloads mapped
LUs to /dev
- -
files Calls /etc/insf to install special files.

- /
/etc/insf can modify /etc/ioconfig

1 �1

Where the 1/0 Probes and binds drivers,

Configuration Is assigns manager indices, records them in
Maintained
(kernel level)

User Uccess to the

1/0 Configuration
After Boot Process
(user level)

Commands that you can use to affect or see the kernel configuration.
.-�����--. ....�����--.
.
/etc/ioscan /etc/rmsf /etc/lsdev /etc/insf
Reads data struc- Removes device /etc/lssf Makes new
tures and displays special files Lists device device special
the 1/0 and drivers or device files.
configuration. LU assignments. special files. Assigns new LUs.

/etc/ioconfig
Holds current LU:
hardware mapping

LG2001 72_043a

Figure 1 1 -4. Series 800 1/0 Auto-configuration

System Configuration 1 1-19
11
Device Drivers
Compiled inside the kernel for all HP-UX systems are software modules known
as device drivers that control 1/0 for a particular device or class of devices.
All drivers shipped are by convention located in a kernel archive library
directory, / et c/ conf/lib .
The kernel identifies the device drivers i t needs via its kernel configuration file
and finds the device driver via the I etc/master file.
Although the term is used generically throughout this text, "device driver"
actually summarizes a hierarchy of programs known collectively as 1/0
managers . 1/0 managers control the various levels of hardware in the bus
architecture. Note evidence of this hierarchy in both the previous S800 file (the
indented lines in the io statement , discussed earlier in this chapter) .
Each kind of I/ 0 manager manages a specific level (or levels) of architecture.
In the Series 300/400/700:
Interface Controls the interface card for handling memory-mapped and
driver processor 1/0 .
Device driver Controls the device for kernel data structures to communicate
from the interface card to the external device itself.
In the Series 800, 1/0 managers are somewhat more complex, but in the
simplest case, might be defined as follows:
CAM The channel adapter manager manipulates data between the ·

Mid-Bus and CIO buses; for example, cio_caO .

DAM The device adapter manager controls the communication
between the 1/0 bus (such as CIO) and interface card. For
example , hp ibO controls HP-IB data protocol transfer for the
CIO bus and hp ib 1 controls HP-IB data protocol transfer for
the HP-PB bus .
DM The device manager controls device-specific communication; for
example, lpr10 for a ciper printer.
LDM The logical device manager provides the software interface
between the user and the operating system, is used as an
interface to a class of D M , and handles generic requests (such
as buffering and queueing) ; for example, lprO .

1 1 ·20 System Configuration

 11
Generally, each 1/0 manager corresponds to a layer of hardware, but many
Series 800 drivers are written to encompass more than one layer. These are
known as monolithic drivers. For example, muxO is a monolithic driver that
combines LDM, DM, and DAM , and transfers requests between the HP-UX
kernel and the channel adapter manager, cio_caO .
In the Series 800, each device driver ( D M and/or LDM) corresponds t o a
class of hardware, such as dis c 1 for all HP-IB disks, disc2 for HP-FL disks
connected through CIO , disc3 for all S C SI disks in a CIO or HP-PB system,
or disc4 for HP-FL disks in an HP-PB system.
The device driver enables the kernel to communicate with the device. The
kernel communicates with the hardware by associating the device-driver name
with a hardware address. The kernel derives its information about device
drivers and their corresponding hardware addresses from the df ile for Series
300/400/700. On Series 800 systems, the kernel scans the hardware to derive
information about device drivers and their corresponding hardware addresses ,
as shown in Figure 11-2.
In the Series 300/400/700, device drivers are listed in the df ile and the
conf ig program derives device-driver definitions from the /etc/master file.
In the Series 800, device drivers are specified in several areas of the S800 file:
I etc/master, include statements, and io statement.

System Configuration 1 1 -2 1
11
The /etc/master File
All HP-UX computers use a /etc/master file, which contains critical device
driver definitions in terms of file type (block or character) , major number, and
I/O functions. These definitions are used by config( lm) (on Series 300, 400,
and 700) and uxgen(lm) (on Series 800) .
O n the Series 300/400/ 700, the /etc/master file consists of lists of devices
and cards , their associated interface or device driver, major numbers by block
or character type. On the Series 700 , the I etc/master file also lists the
subsystems for which library a given device driver is dependent .
On the Series 800, the I et c/master file defines device drivers in a format
resembling C syntax. The first entry in the S800 file includes the I etc/master
file.
Because the bus architectures of Series 800 computers are hierarchical, the
I etc/mast er file also reflects the hierarchical I/ 0 manager relationships of
parent and driver. For example, the following excerpt identifies c io_caO as the
parent driver for the hpflO, the device driver that controls a CIO fiber-link
card:
module hpflO { /* CID HP-FL card 27 1 1 1A */
parent is cio_ caO ;
driver hpflO ;
}

Note Since /etc/master contains critical information used by

config( lM) (for Series 300/400/700) and uxgen(lM) (for Series
800) , do not alter it.

1 1 -22 System Configuration

 11
Device Driver Configuration
You add or delete device drivers when inst alling or updating the system, or
adding a peripheral device to a system that lacks the associated device driver.
Use the procedures in System Administration Tasks manual and Installing
Peripherals manuals for your particular system.
When configuring a new device into the system, Series 300/400/700 and 800
perform somewhat differently.
• In the Series 300/400/700, you invoke SAM, reconfigure the kernel, and
reboot the system to inform the kernel of all configured devices.
On Series 400 and 700 systems equipped with an EISA bus , when the system
boots, the eisa_conf ig utility runs in automatic mode to ensure that all
resources on the EISA bus are properly provided for.
However, kernel drivers cs80 (Series 300/400/700) and hshpib (Series
700 only) are not included in the default configuration. These drivers are
needed for the optional EISA HP-IB card and its peripherals (for example,
to communicate with an HP-IB DDS-format drive). To configure them,
you must invoke SAM to add the cs80 device driver, which also configures
hshp ib , a dependent driver. When the system is rebooted, SAM writes the
drivers to the df ile . sam file. If you build the kernel manually, you must add
the device drivers to your df ile . To view a list of configured device drivers
and their major numbers , run lsdev (see "Major Numbers and lsdev" later
in this chapter).
• In the Series 800, if the necessary driver is in the kernel, you do not have
to reconfigure the kernel. Using ioscan and insf , the system locates , the
devices and binds their necessary device driver. (You can run ioscan and
insf manually; if you reboot , the system runs them automatically.) If the
driver is not in the kernel, use SAM to add the driver or to configure a device
that is not automatically configurable. For the few drivers not automatically
configurable even by SAM, you must edit the io statement of the S800 file,
then regenerate the kernel. When you reboot , the system finds the devices
and binds their necessary device drivers.

System Configuration 1 1 -23

11
Table 1 1- 1 lists device drivers and whether they can be automatically bound
into the Series 800 kernel. ( Note that the driver must be added, usually by
SAM, if it is not included in the S800 file. )

Table 1 1 · 1 . Device-Driver Capabilities (Series 800)

Device D river Description Bound
Automatically?

aut o:x:O Cartridge tape management yes

centO parallel SCSI interface yes
cio_caO Channel adapter yes
disc1 , disc2 Disk-drive management yes
disc3 , disc4 Disk-drive management yes
displayO , graphO , graph2 Graphics management yes
:fddi FDDI networking protocol yes
gpioO (
AFI Asynchronous FIFO Interface ) no
gpio1 General-purpose parallel interface no
hp:flO , hp:f11 Fiber-link interface yes
hpibO , hpib1 HP-IB devices yes
instrO Plotter and HP-IB instrumentation no
lanO , lan1 , lan3 (
LAN networking ) yes
lprO , lpr1 , lpr2 serial and parallel printers yes
mu:x:O , mux1 , mux4 terminals, consoles, keyboards yes
osiO LAN card for MAP networking yes
psiO (
Programmable serial interface networking ) no
pdnO Public data network no
rt iO Real-time interface no
sccl , lanttyO terminals, consoles, keyboards yes
scsil , scsi2 SCSI-interface management yes
t apel , tape2 tape drives yes
t oken2 (
Token Ring IEEE 802 .5 networking ) yes

1 1 ·24 System Configuration

 11
Pseudo-Drivers

Pseudo-drivers are another kind of device driver found in the configuration

description file.
A pseudo-driver ( such as /dev/conf ig) is a piece of code in the kernel, like any
other driver, but not actually associated with a device. Instead, pseudo-drivers
are used by other logical device files such as /dev/null ( the "bit bucket" ) and
/dev/kmem ( kernel memory) and enable them to open, close, read, write, and
perform 1/0 control.
A pty, meaning pseudo-terminal, is another example of a pseudo-driver. The
pty has two elements-master and slave-both of which are necessary for
emulating a terminal's functions. These elements are used by programs that
must address another piece of code capable only of communicating with a
terminal, such as a test program that interacts with vi.

Syste m Configuration 1 1 ·25

11

Device Fi les
HP-UX communicates with peripherals through files called device :files , which
are kept in /dev or one of its subdirectories . Device special files contain no
actu al data, but exist solely as the way to write to or read from a device.
Device files point to the logical address for peripherals and other devices.
Device files contain information encoded in major and minor numbers to
identify device drivers , hardware addresses , and other characteristics. HP-UX
uses this device-driver and address information to manage all data transfers to
and from the device.
In the Series 300/400/700, the device-file name encodes hardware-address
information directly. In Series 800, the hardware information must be inferred
from the minor numbers, found either by running 11 on the device file or (for
Series 800 only) by running the /etc/lssf or the /etc/ios can command.
Device files link devices to the kernel and enable I/O between the kernel
and devices . HP-UX uses device files to determine which driver to use when
communicating with a device and the device's hardware address . (Valid drivers
must be compiled into the running kernel.)
The kernel treats I/O to a device much like I/O to a file, by performing I/O
operations as if through the file system. For example, each terminal has its own
device file through which HP-UX writes data (which appears on the termin al
screen) and reads data (typed by the user at the terminal keyboard) .
Typically, two device files-a block and a character device file-are created for
each I/O device in the system. (These are described shortly.)
Individual device files can be created by the following commands:
• insf ( lm )-Series 800 only.
• mksf ( lm)-Series 800 only.
• mknod( lm)-Series 300/400/700/800.
On the Series 800, insf is the easiest , fastest way to handle special files . insf
can be used any time. It is not actually run when the kernel is regenerated,
but rather when the new kernel is booted. insf assigns logical unit numbers ,
needed for the device before you can use either mksf or mknod to create a
device file. mksf and mknod should be used only for special cases , as they are
more difficult to use than insf .

1 1 ·26 System Configuration

 11
Characteristics of Device Files
As a system administrator, you need to understand the following characteristics
of device files to interpret output from a system configuration or to configure
devices on a system:
• device file directories
• I/ 0 type (block and character device files)
• device file listings
• device file naming conventions·

• major and minor numbers

Organization of Device File Directories

All device files should be located in the I dev directory. Many commands
expect to find device files in I dev and fail if the required device file is not
there. Peripherals are associated with block and character (raw) special files
found in the /dev subdirectories, as listed in the following table:

Type of peripheral Block Mode Character Mode

Disk drives dsk rdsk
Cartridge-tape drives ct rct
Magnetic and DDS-format tape drives mt rmt
Graphics framebuffer devices crt
Floppy disk drives floppy rfloppy
Autochangers ac rac

Device File Directories in a Cluster (Series 300/400/700 only)

If your system is an HP- UX cluster, the I dev directory is a context-dependent
file (CDF).
Log into the computer to whkh the peripheral is a.tta.ched when adding or
changing device information in a cluster. See Managing Clusters of HP 9000
Computers: Sharing the HP- UX File System.

System Configuration 1 1 -27

11
Block versus Character Device Files
There are two main types of device files: block and character, represented by
the characters b and c, respectively. All 1/0 devices can be classified as block
or character.
Block device files transfer data using the system buffers to speed up 1/0
transfer.
Storage devices such as tape drives, hard and floppy disks, and magneto-optic
drives can use block device files.
Character device files transfer data blocks of varying sizes one character at a
time without using the system's 1/0 buffers. The user program does its own
buffering. Typically, the following devices use character device files: terminals,
printers, plotters, digitizers, magnetic tape drives, cartridge tape drives, and
disk mass storage devices. Character 1/0 transfer is also called raw 1/0
transfer, and sometimes character devices are called raw devices.
Disk drives generally need both block and character device files. The file
system accesses ordinary files and directories with block device files, using the
buffer cache for speed. A database can use a character device file to access
an area of the disk that the file system does not use. File system integrity
is checked using fsck through the character device file, since f s ck requires
low-level control of the device.
See the HP- UX Reference for manual page specifications on f s ck and newf s .

Device File Listing Format

When you execute an 11 command from the /dev directories , you see a
device-file entry with different information than other long file listings:

brw-rw-rw- 1 b in bin 0 Ox400222 Apr 5 1990 /dev/ct/ c2d1s2

The first character in the first field identifies the file type; that is, whether it
is a character ( c) or block (b) device. ( Remaining information in the first and
next three fields resemble any file listing, with file access permissions, linkage,
owner, and group . )
Then, instead of a field showing the size of a file, device files feature two unique
fields : major number and minor number ( in this case 0 and Ox400222) . Major

1 1 -28 System Configuration

 11
and minor numbers derive from hardware-path information given to commands
for creating the device file. (See "Overview of Major and Minor Numbers" for
further discussion.)
If the file is a symbolic link, you will derive more accurate output by using 11
-L, which lists the file or directory to which the link refers.
In the Series 800, detailed information about a device file can be displayed by
using the I etc/lssf command. For each device file specified, ls sf displays the
major number, file type, characteristics encoded into the minor number, and
logical unit specification (if present) .
lssf can b e thought of as a corollary t o mksf(lm). For example, a device file
created with the command
mksf -d disc2 -1 1 -u 0 -s 3 /dev/dsk/c 1d0s3
would be decoded:
# lssf /dev/dsk/c 1 d0s3
disc2 lu 1 unit 0 sect ion 3 address 4 . 2 . 0 /dev/dsk/ c 1d0s3
Another tool, ioscan( lM), displays the following output for the same
information:
# ioscan -£ /dev/dsk/c1d0s3

Class lu H/V Path Driver H/V St atus S/V Status

=================================================================

disk 1 4.2.0 cio_ca0 . hp£10 . disc2 ok (Ox2) ok

See lssf ( lM), mksf ( lM), and ioscan(lM) in the HP- UX Reference .

System Configuration 1 1 -29

11

Overview of Major and Minor Numbers

The kernel recognizes a device and its device driver by numbers associated with
device files called major and minor numbers .

Device Fiie

Major number Minor number

(8 bits) (24 bits)

I
Which driver
\
Where is the device the driver
do I taik to? needs to access?
What are the device file's special
access modes (e.g. tape density)?

LG2001 72_042

Figure 1 1 -5. Device Files Contain Major and Minor Numbers.

Each device driver in the system is assigned a major number, which the kernel
uses to locate the driver routine to service an 1/ 0 request .
The driver uses the minor number to get to the specific device and for
information regarding how to handle data.

1 1 -30 System Configuration

 11

Major Numbers
The major number is an index for the device driver into one of two kernel
tables-bdevsw, the block device switch table and cdevsw, the character device
switch table. These device switch tables list driver entry points.
Drivers that support both block and character I/O (such as scsi disk driver
and optical autochanger) have two major numbers-a block major number and
a character major number-each of which can be referenced in their respective
switch tables , depending on how the device is used . Or, explained from another
perspective, the character device switch table finds all the functions that can be
called by a character device driver (such as the scsi driver) in character mode.
Major numbers are listed in the I etc/master file, discussed earlier in this
chapter.

Major Numbers and lsdev

You can execute the lsdev( lm) command to get the major number of the
devices configured on your system.
The - 1 designation in either character or block major number field can have
several meanings:
• The device might not require that kind of major number; some devices are
accessable only in block or character mode.
• The driver is not configured into the kernel.
• The user may not access the driver directly. For example, the 98625 driver
handles access to the high-speed HP-IB I/O card. The user may not access it
directly, but must configure the cs80 driver, which can access 98625 .

System Configuration 1 1 -3 1
11
Here is sample' output from the lsdev command issued on a Series 300, 400, or
700:
Charact er Block Driver
0 -1 /dev/console
1 -1 BP98628 , BP98626 , and BP98642
2 -1 /dev/tty
3 -1 /dev/mem , /dev/kmem , and /dev/null
4 0 CS80 disk
5 -1 BP7970/BP7971 nine-track magnetic tape
6 1 HP9826/BP9836 internal flex disk
7 -1 BP-UI print er
8 -1 /dev/svap
9 -1 BP7974/BP7978 nine-track magnet ic tape
11 2 Amigo disk
13 -1 SRM opt ion
16 -1 Master pty
17 -1 Slave pty
18 -1 IEEE 802 device
19 -1 Ethernet device
21 -1 BP-IB DIL
22 -1 GPIO DIL
23 -1 raw 8042 BIL
24 -1 BIL
25 -1 BIL cooked keyboards
26 -1 c iper printer
47 7 SCSI disk
54 -1 SCSI t ape
55 10 opt ical aut ochanger

1 1 -32 System Configuration

 11
Here is sample output from lsdev on a Series 800:
Character Block Driver Class
-1 -1 processor processor
-1 -1 memory memory
-1 -1 bus_ converter be
51 -1 lanl lan
-1 -1 c io_caO cio
7 8 disc1 disk
12 10 disc2 disk
13 7 disc3 disk (SCSI)
-1 -1 hpibO hpib
12 10 hpflO hpfl
50 -1 lanO lan
1 -1 muxO tty
5 5 t apel tape_drive
57 11 rdu pseudo
34 -1 ktestio pseudo
29 -1 kt est pseudo
60 -1 nm pseudo
56 -1 ni pseudo
46 -1 netdiagl pseudo
48 -1 mirconfig pseudo
31 -1 sy pseudo
-1 13 slfl pseudo
8 3 Sti' pseudo
17 -1 ptyl pseudo
16 -1 ptyO pseudo
3 -1 mm pseudo
41 -1 meas_drivr pseudo
27 -1 dmem pseudo
28 -1 diagO pseudo
69 -1 devconf ig pseudo
0 -1 en pseudo

For more information, consult the lsdev ( lm) and lssf ( lm) manual pages in the
HP- UX Reference .

System Configuration 1 1 -3 3
11
Minor Numbers
The minor number is a six-digit hexadecimal encryption of address and
characteristic hardware information, stored in the inode of the device's speci al
file. The minor number gives the device driver the channel or connection
through which to perform 1/0 .
Minor numbers are used to distinguish among devices whose major numbers
are identical ( that is, use the same driver ) and are connected to the same
system processor unit ( SPU ) . The minor number typically defines one or both
of the following:
• The device's physical location ( hardware address or logical unit number ) and
switch settings, to enable the system to locate the peripheral.
• Behavioral information. The minor number's content depends on the type of
device; for example, the minor number of a magnetic tape drive encodes tape
density and other behavioral information.
Minor numbers also provide some flexibility in your system configuration. On
Series 300, 400, and 700 systems , the minor number can be decoded by select
code or function number to a specific function on a multi-function card.
Device drivers can use minor numbers to select a communication line ( such
as a specific terminal on a multiplexer ) or operating mode ( such as magnetic
tape drive with different minor numbers to enable different formats ) . In the
following long listing of two device :files of a medium-density tape drive file on a
Series 800,

crw-rw-rw- 1 b in bin 5 Ox020000 O ct 1 1 1 : 04 /dev/rmt /Om

crw-rw-rw- 1 bin bin 5 OxOaOOOO Nov 8 1989 /dev/rmt/Omn
Ox020000 and OxOaOOOO are both minor numbers that can be used to
access the tape drive, but as you can see from the output of I et c/ls sf
I dev /rmt /Om* ,
tap e 1 lu 0 bpi 1600 at&t address 4 . 2 . 3 /dev/rmt /Om
tape1 lu 0 bpi 1600 no_rewind at&t address 4 . 2 . 3 /dev/rmt/Omn
/dev/rmt /Om rewinds the tape when the programs are :finished with the device,
whereas /dev/rmt/Omn does not . Thus , if you specify /dev/rmt/Om in the

1 1 -34 System Configuration

 11
fbackup ( lm) command, the system will rewind the tape at the end of the
backup; if you specify /dev/rmt/Omn, it won't .
Since minor numbers give device-driver specific information, they differ by
system architecture.

Minor Numbers on Series 300/400

For the Series 300/400 systems, the digits after Ox identify the select code and
other device-specific information. ( See discussion of select codes earlier in this
chapter. )
The select code is preset for all device interfaces on your system. Each device
interface must have a unique select code. You might need to change the select
code of a device interface if the preset value is the same as another on the
system. Refer to Installing Peripherals, Volume One , chapter 2, "Installing
Interface Accessary Cards," for instructions to change the select code for a
particular interface.
Other fields in the minor number are device-driver dependent . After the select
code is information identifying bus or port address. The Installing Peripherals
manual gives specific instructions for setting up devices and interface cards
to a specific address, using SAM ( recommended method ) or manually ( using
/etc/mknod).
Table 1 1-2 summarizes minor number formats for Series 300/400 HP-UX
peripheral devices. In all cases, Ox indicates hexadecimal format . Unless
otherwise noted, Sc indicates the select code of the interface, an 8-bit value.

System Configuration 1 1 ·35

11
Table 1 1 -2. Summary of Series 300/400 Minor Number Formats

Peripheral Type Format Syntax

Disk drive OxScBaUV Ba = HP-IB or SCSI bus address (4-bit value)

U = unit number.
V = volume number (0 for single file systems) .

Magneto-optical disk OxScBOSur B = SCSI bus address (4-bit value) .

drive library o = Reserved (always o)
Sur = surface (16-bit value)
(See au.tochang er(7) for surface values.)

Nine-track magnetic OxScBaUV Ba = bus address (from switch settings on the tape drive).
tape\ DDS-Format tape U = unit number (O=SCSI drive, 4=HP-IB drive; 4-bits) .
drive V = volume number (4-bit value) , as follows:
bit 3--always 0
bit 2-immediate report (O=on,l=off)
bit 1-close style (O=AT&T,l=Berkeley)
bit 0-rewind on close (O=yes,l=no)

Cartridge tape drive OxScBaUV Ba = HP-IB bus address (from tape-drive swit ch settings) .
U = unit number.
V = volume number ( 0 for single file systems).

Terminal and modem OxScPaOI Pa = Port address (swit ch settings on the device)
0 = 4 bits of 0.
I = hexadecimal 4-bit binary number, defined as follows:
bit 3--always 0
bit 2-0=modem; l=direct connect
bit 1-protocol (O=Simple/USA; l=CCITT/Europe)
bit 0-modem (O=dial-in; l=dialout)

Pseudo-terminal device OxOOYYYY YYYY = unique hexadecimal value (0 to npty -1) ,

file to identify relationship between master and slave.

HP-IB plotter and OxScBaOO Ba = HP-IB bus address (4-bit value) .

digitizer 00 = always 0

1 1 -36 System Configuration

 11
Table 1 1 -2.
Summary of Series 300/400 Minor Number Formats (continued)

Peripheral Type Format Syntax

RS-232-C plotter O:z:ScPoOA Po = Port number, values 0-3 (8-bit value) .

0 = always 0 (4-bit value)
A = access type, defined as follows:
bit 3-always 0
bit 2-0=modem; l=direct connect
bit 1-protocol (O=Simple/USA; l=CCITT/Europe)
bit 0-modem (O=dial-in; l=dialout)

HP-IB printer O:z:ScBaOll Ba = HP-IB bus address (4 bit value) .

O = always 0 ( 4 bit value) .
II = mode of operation, defined as follows:
bit 3 (Auto FF)-l =NO-EJECT
bit 2 (Case Fold)-l =Upper
bit 1 (Overprint)-l=NOCR
bit 0 (Protocol)-O=block (Amigo); l =character

RS-232-C printer O:z:ScPoOA Po = Port number, values 0-3 (8-bit value) .

0 = always 0 (4-bit value)
A = access type, defined as follows:
bit 3-always 0
bit 2-0=modem; l =direct connect
bit 1-protocol (O=Simple/USA; l=CCITT/Europe)
bit 0-modem (O=dial-in; l=dialout)

Parallel printer O:z:ScPoOOO Po = Port number, values 0-3 (8-bit value) .

000 = always 0 (12-bit value)

Graphic device O:z:STXIIX S = Select code (4-bit value) .

T = Device resolution and physical address (12-bit value) .
IIXI = Device-driver information

HP-HIL device O:z:OOOOBO 0000 = always 0 (16-bit value) .

B = Position on HP-HIL bus (4-bit value) .
0 = always 0 (4-bit value) .

GPIO device O:z:ScOOOO Sc = Select code (8-bit value) .

0000 = always 0 (16-bit value) .

System Configuration 1 1 -37

11
Minor Numbers on the Series 700
The minor number, used to describe the hardware path to the device, is a
24-bit value with the following format :

4 bits 4 bits 4 bits 12 bits

SMB # Slot # Fu nction # Driver Specific I nformat i o n

Figure 1 1 -6. Series 700 Minor Number Format

where:
SMB # is a four-bit single hexadecimal digit representing the System
Bus Module number with the following decim al values:
0 Standard Graphics Connect (SGC) slot 1
1 SGC slot 2
2 Core 1/0
4 EISA adapter
8 Processor
9 Memory
Slot # is a four-bit single hexadecimal digit representing the EISA
interface card slot number.
Function # is the four-bit value specifying a particular function if the
interface card is a multifunction al card. The basic function
numbers of the Core 1/0 are listed:
1 S CSI
2 LAN
3 HIL
4 Serial port A
5 Serial port B
6 Parallel
If the interface is not a multifunction card, the function
number contains driver-specific information. (Additional
function numbers may be model-specific.)
Driver Info 12-bit (three-hexadecimal digit) value that can be found in the
HP- UX Reference, section 7 for the particular driver you are
using.

1 1 -38 System Configuration

 11
Decoding a Minor Number for a Series 700 SCSI Device.
Table 1 1-3 may be useful for interpreting the information that follows.

Table 1 1 -3. Decimal, Binary, and Hexadecimal Equivalents

Decimal Binary Hexadecimal Decimal Binary Hexadecimal

0 0000 0 8 1000 8
1 0001 1 9 1001 9
2 0010 2 10 1010 A
3 001 1 3 11 1011 B
4 0100 4 12 1 100 c

5 0101 5 13 1101 D
6 0110 6 14 1110 E
7 0111 7 15 1111 F

The hexadecimal notation for the Series 700 minor number follows the format
OxSSFDDD , which encodes S MB , slot number, function, and device-specific
information. The minor number for a device connected to the Core I/O SCSI
interface with a driver specific value of Ox333 is:
Ox20 1333
The corresponding binary minor number value is:
I 0010 I 0000 I 0001 I 001 1 0011 001 1 I
These hexadecimal values can be interpreted as follows:
0010 The core I/O card occupies SMB slot is 2.
0000 The core I/O card has no slots .
0001 The S CSI function is number 1 on the core I/O card.
001 1 001 1 001 1 The driver-specific information i s Ox333.
information is Ox333.

System Configuration 1 1 -39

11
Decoding a Minor Number for a Series 700 EISA SCSI Device.
The minor number for a device connected to an EISA S CSI interface in slot
one with a driver specific value of Ox333 is:
Ox410333
The corresponding binary minor-number value is:
I 0 100 I 0001 I 0000 I 0011 001 1 001 1 I
where the S MB # is 4, the slot number is 1 , and the driver-specific information
is Ox333.

Note Function numbers only apply to the Core I/O interfaces and
slot numbers only apply to the EISA interfaces .

Minor numbers might vary from the formula shown, particularly in the case of
pty drivers, which have unique requirements.

1 1 -40 System Configuration

 11
Table 1 1-4 summarizes minor number formats for Series 700 HP-UX peripheral
devices. In all cases, Ox indicates hexadecimal format .

Table 1 1 -4. Summary of Series 700 Minor Number Formats

Peripheral Type Format Syntax

SCSI disk drives OxBMFTOO BmF-Bus module and function number of controller,
asfollows:
201 = SCSI on core I/O
4SO = SCSI on EISA (s = slot number on EISA bus)
T-Target number (0-6 = bus address setting on device)
00-Place holders for disks

DDS-format tape drive OxBMFTDO BmF-Bus module and function number of controller,
as follows:
201 = SCSI on core I/O
4SO = SCSI on EISA (S = slot number on EISA bus)
T-Target number (0-6 = bus address setting on device)
D-Device compression, partition selection bits
0-0peration bits (e.g. immediate reporting, rewind)
(See Installing Peripherals for DDS
tape D and 0 bit definitions.)

Magneto-optical OxBMFTSr BmF-Bus module and function number of controller,

autochanger disk-drive as follows:
library 201 = SCSI on core I/O
4SO = SCSI on EISA (s = slot number on EISA bus)
T-Target number (0-6 = bus address setting on device)
Sr-Surface for media
(See Installing Peripherals and autochanger (7)
for defining media surfaces.)

Parallel printer OxBMFTOOA BmF-Bus module and function number of controller,

as follows:
201 = SCSI on core 1/0
4SO = SCSI on EISA (S = slot number on EISA bus)
T-Target number (0-6 = bus address setting on device)
00-always 00 (1 2-bit value)
(See Installing Peripherals and cent(7)
for handshake modes.)

System Configuration 1 1 -4 1
11
Minor Numbers on Series 800

The assignment of minor numbers on Series 800 is virtually foolproof: if you

use insf -d driver _name , insf installs device files in the current directory
with properly encoded minor numbers and logical unit numbers. insf
is also run automatically at boot-up . The Installing Peripherals manual
has device-specific information on minor numbers throughout its text and
Appendix A provides minor number formats for all Series 800 device drivers.
The minor number on Series 800 is a six-digit hexadecimal field containing such
information as:
• logical unit number (all physical devices)
• section number (disks)
• port number (terminals)
• SCSI target number (disk, tape)
• HP-IB address (plotter, disk, tape)
• device options (most devices)
As with the Series 300/400/700, Series 800 minor numbers are constructed
differently for each kind of device. For example,
• The logical unit of each component daisy-chained to an HP-IB , HP-FL ,
MUX, SCSI, or HP-HIL interface is encoded into the peripheral's minor
number.
• The minor number for a tape device specifies the density at which the tape
drive is to operate and whether it should rewind at the end of a file.
Refer to the Installing Peripherals , device documentation, and HP- UX
Reference Manual for information on the specific device in question.

1 1 -42 System Configuration

 11
Logical Unit Number.
The lu is the logical unit specification that allows one device driver to handle
many devices. (The lu is necessary because the complexity of the Series 800
bus structure makes it impossible to store all hardware addressing information
in the minor number.)
The lu is assigned by insf ( lM). The lu is mapped into the I/ O tree, which the
kernel then consults to determine which specific hardware address the driver
wants to access.
The lu allows the system to keep track of all configured devices belonging to
the same class of device.
For example, consider a system configured with HP-IB , HP-FL , and S CSI disk
drives, which are run by device drivers disc 1 , disc2 ( disc4 for HP- P B ) , and
disc3 respectively. If a system has 10 disk drives, each one is assigned an lu
number from 0 to 9. All three kinds of disk drive (HP-IB , HP-FL , and S CSI)
share the same pool of lu numbers, because all are of the same class of devices
(disk).
lu numbers are assigned in chronological order, according to when the device
was added to the system rather than by architectural proximity, although the
system's automatic configuration capabilities associate the device lu with a
low-to-high ordering of slot numbers in hardware.
The lu is associated with hardware paths in the I/O tree and in
/ etc/ ioconf ig . /etc/ioconf ig is used to maintain logical unit number to
hardware path mappings across system boots.
The Installing Peripherals (Series 800) manual contains information about
logical unit numbers.

System Configuration 1 1 -43

11
Sources for Minor Number and other Device Information
Other bits in Series 800 minor numbers provide behavioral information or
location about the device, including section number, MUX port , device options,
and other flags. Since each is directly related to the kind of device accessed,
consult the device's manual page in the HP- UX Reference for specific details.
Sometimes you might need to make device files for individual devices or classes
of devices .
The following manual pages in the HP- UX Reference give minor-number and
device-file conventions for specific classes of devices:
• intro(7)-nine-track tape and disk devices
• autochanger(7)-SCSI optical mass-storage devices (Series 300/400/700 only)
• cent(7)-parallel printers (Series 300/400/700 only)
• ct(7)-cartridge-tap e devices
• disk ( 7 )-disk devices
• flopp y ( 7 )-direct flexible disk access (Series 700 only)
• graphics(7)-CRT graphics devices (Series 300/400/700 only)
• hil(7)-HP-HIL device driver
• hpib(7)-HP Interface Bus driver (Series 800 only)
• iomap(7)-physical address mapping (Series 300/400 only)
• lp ( 7 )-line printer (Series 300/400/800 only)
• mt(7)-magnetic tape, DDS-cartridge tape, and QIC-cartridge tape interface
and controls
• scsi(7)-Small Computer System Interface device drivers
• scsi_changer(7)-SCSI media changer device driver (Series 700 only)
• scsi_ctl(7)-SCSI device control disk driver (Series 700 only)
• scsi_disk(7)-SCSI direct access disk device driver (Series 700 only)
• scsi_tape(7)-SCSI sequential access tape device driver (Series 700 only)

1 1 -44 System Configuration

 11

Device File Name Conventions

The intro(7) and manual pages of the HP- UX Reference describe naming
conventions for various types of device files, as they are shipped on your
system. Using the naming conventions makes your system easier to support
and maintain. Also, some commands (such as the insf ( lm) command on the
Series 800 and SAM) expect the recommended path names and file names.
Critical elements of a device file are its major and minor numbers. Particularly
on a small system, you can name device files as you wish. If you do so, keep
the names of your /dev device files intuitive. For example, printer device files
are easily identified when named by their product , such as /dev/ptr2567C
for an HP 2567C printer. Similarly, the subdirectory name /dev/ac or
/dev/autochg would represent autochanger.

Naming Disk Devices

Device files for disk drives should reside in the I dev directory. Disk drives
require both character and block device files. Disks use the I dev /rdsk
directory for the character device files and the I dev I dsk directory for the block
device files.
An exception exists for magneto-optical devices , which use the /dev/rac and
/dev/ac directories. Description of file-name conventions for magneto-optical
devices follows standard disk devices in this section.
File names for disk devices on Series 300/400 use the following format :

/dev/ [ r ] dsk/c#d# [ l# ] s#
where:
r Indicates a raw (character) interface to the disk.
c# Specifies the controller number. The # should be replaced with
a capitalized hexadecimal representation of the select code.
d# Specifies the bus (target) address, a number representing the
entire device.
[1#] The logical unit number (1) is used to identify specific units in
integrated devices.

System Configuration 1 1 -45

11
s# The s# stands for section number. This is typically zero,
except when using software disk striping (refer to the System
Administration Tasks manual for information) .
File names for disk devices on Series 700 use the following format :

/dev/ [ r ] dsk/c###d# [ l# ] s#
where:
r Indicates a raw (character) interface to the disk.
c### Specifies the controller bus module and function number; the
### can be:
201 = core 1/0 (the default)
4# = EISA bus (4) and slot number (#). If specifying a
device on an EISA bus, only two (not three) digits follow the
c (for example, 41).
d# Specifies the target number, the address set on the device
itself. This number represents the entire device.
[1#] Disk arrays require a logical unit number, which designates
specific device units .
s# The s# stands for section number. This i s typically zero,
except when using software disk striping (refer to the System
Administration Tasks manual for information) .
On Series 800 systems , default disk device files are shipped as follows:

/dev/ [ r ] dsk/c#d#s#
where:
r Raw interface to the disk.
c# Controller, represented by the lu number.
d# Drive number; always 0 .
s# Section number.
The assignment of controller, drive, and section numbers is described in the
System Administration Tasks manual. Additional information about disks can
be found in disk (7) in the HP- UX Reference Manual.

1 1 -46 System Configuration

 11

Naming Magneto-Optical Disk Devices

Magneto-optical disk devices use the /dev/rac and /dev/ac directories for
character and block device files, respectively. Device files for magneto-optical
disk devices on HP 9000 systems use the following format :

/dev/ [ r ] ac/c#d#_pp
where:
r Indicates a raw ( character ) interface to the disk.
c# Specifies the controller bus module and function number; the #
can be:
Capitalized hexadecimal representation of select code ( Series
300/400)
201 = core 1/0 (Series 700 default )
4# = EISA bus (4) and slot number (# ) .
Logical unit number ( Series 800)
d# Specifies the target number, the address set on the device
itself. This number represents the entire device.
Underscore ( always present ) .
pp Disk platter number and side (a or b ) .
For further information, consult the chapter, "Setting u p Devices Using
HP-UX Commands," in the Installing Peripherals for your system. Also see
autochanger(1) in the HP- UX Reference Manual.

Naming Nine-Track Magnetic Tape Drives

The following naming convention is recommended for magnetic tape devices
because it connects most of the mode flags with the device name:

/dev/ [ r ] mt /c#d# [ hml ] [ c ] [ n ]

where:
r indicates a raw ( character ) device.
c# indicates the controller number ( option ally specified by the
system administrator ) .
d# is the device number.
hml indicates the density:

System Configuration 1 1 -4 7
11
• h (high) for 6250 bpi.
• m (medium) for 1600 bpi .
• 1 (low) for 800 bpi .
c indicates data compression.
n indicates no rewind on close.
For example, I dev /rmt/2mn is raw device 2 at 1600 bpi with no rewind and no
compression.
Additional information about 9-track magnetic tape is available in mt(7) of the
HP- UX Reference Manual.

Naming DDS-Format Tape Drives

Device files for DDS-format tape drives are named as follows :

/dev/rmt/c#d#m l c [ n ]
where:
r indicates a raw (character) device.
c# specifies the controller number, in capitalized hexadecimal
notation.
d# is the device number.
mlc indicates density. Use either m for medium density (standard
DDS format) or c for compressed density.
n indicates no rewind on close.
Additional information about DDS cartridge tape can be found in mt(7) of the
HP- UX Reference Manual.

Naming QIC-Format Tape Drives

QIC format is an industry standard for the quarter-inch cartridge (QIC) tape
drives, available as S C SI devices on Series 700/800 systems . QIC-format device
files are named as follows:
/dev/rmt/#m [ n ]
identical to DDS format , or

/dev/rmt/#qic
1 1 -48 System Configuration
 11
for default AT&T-style QIC format, or

/dev/rmt/#qic [ 120 1 1so 1 s2s ] [ n ] [ b ]

r indicates a raw (character) device.
# is the S CSI logical unit number.
1201 1so1s2s indicates specific QIC format , if needed.
n indicates no rewind on close.
b indicates Berkeley style (if b is omitted, indicates AT&T style)
For further information about QIC-format cartridge tape files, see mt(7) or
insf (lM) of the HP- UX Reference Manual. For information on backing up
your system to QIC cartridge tape, consult "Backing Up and Restoring Your
Data" of System Administration Tasks manual for your system.

Naming Cartridge Tape Drives

As shipped, device files for cartridge tape drives are named as follows:

/dev/ [ r ] ct/ [ r ] c# [ d# ] [ s# ]
where:
r indicates a raw interface to the cartridge tape; the second r is
reserved to indicate that this cartridge tape is on a remote
system.
c# indicates the controller number.
d# ?Ptionally indicates the drive.
s# optionally indicates a section number.
The assignment of controller, drive, and section numbers is described in the
System Administration Tasks manual; also see in ct(7) of the HP- UX Reference
Manual.

Naming HP-IB Devices (Series 800 only)

As shipped, HP-IB device files are named as follows:

/dev/hpib/# [ a# ]
where:

System Configuration 1 1 ·49

11
# the first # specifies the bus number ( assigned by the system
administrator )
a# if present , a# specifies the address on the bus. Device files
without an address suffix denote the raw bus. See hpib(7) for
full explanation of auto-addressed files vs. raw bus files .

Naming Modem Device Files

Modem device files are tricky because you must create several of them,
depending on which software release you run. If running the current software
release, you must create two modem device files:
/dev/ttydNN dial-in terminal
/dev/culNN dial-out line
/dev/cuaNN dial-out line
In HoneyDanBer UUCP and cu, the system uses /usr/lib/uucp/Systems to
dial the connection.
If you are using earlier software release or a non-HoneyDanBer uucp , the
system dials its connection using /usr/lib/dial it . c, which requires you to
create a third modem device file:
/dev/cuaNN auto-dial unit
Besides the peculiarities of modem device file names, you might pay special
attention to the modem minor numbers. The last two di gi ts of the minor
number indicate whether a terminal is hard-wired or connected by modem.
Normally, you have a modem on one port and your terminals on different ports
( for example, a modem on port 0 and a dial-in terminal on port 1 ) .
Some devices have very specific minor-number requirements . For example, the
following table shows the minor number for a modem connection at select code
9:

Type Device file name Minor number

dial-in /dev/tty0 1 Ox090000
dial-out /dev/cul01 Ox09000 1
dial-out /dev/cua0 1 Ox09000 1

1 1 -50 System Configuration

 11
Note that /dev/cuaNN is used to establish the connection, whereas
/dev/culNN is used once the connection is established. You can link them
together, as follows:
# ln /dev/cul01 /dev/cua0 1
Direct-connect devices also have special needs. The minor number must end in
the digit 4. If you have a MUX card, the minor number for a direct-connect at
port 1 would be Ox0d0 104:
..

Type Device file name Minor number

direct-connect /dev /ttyd0 1 Ox0d0 104

System Configuration 1 1 -5 1
 12
12
H P - UX Peripherals

Computers are used t o process massive amounts of data, which must b e

stored as large files online. A variety of peripheral devices meet this storage
need. This chapter describes the following technologies that handle and store
magnetic media:
• Nine-track magnetic tape drives .
• DDS-format tape drives.
• Cartridge tape drives .
This chapter also describes :
• The CD-ROM file system.
• Rewritable optical disk drives.
• Terminals and modems.
• HP-IB and S CSI guidelines .

HP-UX Peripherals 1 2- 1
 Magnetic Tape

Note Series 700 systems do not support nine-track magnetic tape

drives ; see discussion of DDS-format tape drives instead.
12

Perhaps the closest thing to an industry standard for mass media, nine-track
( 1/2-inch) magnetic tape ( "magtape" ) serves as a low-cost, high-capacity
medium to store information. Magnetic tape is also the most interchangeable
medium between different hardware and operating systems.

Magnetic Tape Physical Characteristics

Magnetic tape is a medium similar to everyday audio cassette tape, but it
is used to store digital information. All standard magnetic tape is 1/2-inch
wide, and is available in 600, 1200 and 2400 foot reels. (A 2400 foot reel is
approximately 1 foot in diameter.) The size of the reels, hubs, tape width and
other mechanical properties are all specified by ANSI standard.
On the back of the reel is a removable soft plastic write ring. Every magtape
drive has a sensor mechanism to detect the presence of this ring. When a ring
is present the tape can be written to by the host and cannot be written when
absent (it is write protected ) . Normally, once a tape is written, the ring is
removed and left out indefinitely except when being rewritten.

Tape Density
The measure of the amount of information which can be stored in a given area
of tape is known as tape density. Bits per inch (bpi ) , a common measure of
tape density, is the number of bits per track , recorded per inch on the tape.
For nine-track tape, eight data bits and one parity bit are written across the
width of the tape simultaneously. Thus for nine-track tape, bpi is synonymous
with characters per inch (cpi ) . One of these characters is sometimes called a
frame.

1 2-2 HP-UX Peripherals

Logical Organization of Nine-Track Magnetic Tape
The load point , or beginning of tape, is a foil mark placed about 10 feet from
the beginning of a tape. The foil mark is a short piece of silver tape placed on
one edge of the tape on the non-recorded side by the manufacturer. 12
When you load a tape (put the tape i n the drive, and press "load" ) , the drive
searches forward until the lo'ad point is found and placed under the sensor. The
first write is then treated specially: several inches of tape are skipped and then,
when using PE or GCR formats, a special burst of data is written to the tape
( which is invisible to the user) . This is the identify burst. Data is recorded
after the identify burst in the usual way. The first read expects the identify
burst , and quietly skips over it . Some smart drives, such as the HP 7978, can
determine the tape density from the identify burst ( 1600 and up) .

File Marks
A special command to the drive writes a :6.le mark to the tape. A file mark
is a special type of record recognized by the drive and reported as a Boolean
condition during reading. It is not possible to write a file mark as ordinary
data.
Single file marks are used to separate logical files on tape. Two consecutive file
marks are used to signify the logical EOT. Data is undefined past the logical
EOT.

End of Tape Markers

There is both a logical end of tape (EOT) and a physical EOT (see
Figure 12- 1 ) . Logical EOT is two consecutive file marks. Physical EOT is a foil
mark about 25 feet from the end of the reel. Like the load-point foil mark, the
physical EQT foil mark is placed on the tape by the manufacturer.
Note that the distance between the EOT detector and the read/write head
may vary among different model tape drives. So, one drive may return an EOT
indication associated with the lOOOth record on the tape, while another drive
may return an EOT indication with the 999th or the lOOlst record. For small
records this variation may be large; for large records this variation is probably
small.

HP-UX Peripherals 1 2·3

 Iden tify Double File

I
B u rst Doto Mork
A
( �
_l_
12

•
DOD • • •

DD -
Loa d Physical
Point EOT

Figure 1 2 - 1 . Magnetic Tape Format

Magnetic Tape Capabilities

A tape drive can read and write to the media, rewind to the load point ,
forward or back space one record, and forward or back space to the next file
mark. Tape drives also rewind to unload: the tape is rewound and taken off
line. Some tape drives actually rewind the tape out of the threading path;
others simply set an interlock that requires manual intervention to release the
tape.
When digital information is written to a tape, it is written in a series of tracks
(like an eight-track car stereo ) . Most magtape today is written in a nine-track
format . Older systems wrote six tracks plus a parity bit , resulting in 7 tracks.

Records and Record Sizes

A series of frames written to the media is known as a record. The physical
record size is variable. The maximum limits on record size range from 16 KB
to 60 KB , depending upon the tape drive. Beyond these limit , the drive rejects
the request and there are no write/read retries. The maximum record sizes are:

1 2-4 HP-UX Peripherals

HP 7971 1600 bpi-32 KB
HP 7974 1600 bpi-16 KB
800 bpi-16 KB
HP 7978A 1600 bpi-16 KB 12
6250 bpi-64 KB
HP 7978B 1600 bpi-32 KB
6250 bpi-60 KB
HP 7980 1600 bpi-64 KB
6250 bpi-64 KB

Cyclic Redundancy Check

When writing a tape, a number of frames are written by the drive in a single
transaction. This collection of frames is called a record. Part of the record,
but invisible to the user, is a cyclic redundancy check (CRC). The CRC is
recorded as some additional frames on the tape. There is a very short blank
section between the true record and the CRC . Following the CRC is a nominal
1 / 2-inch gap of unrecorded tape, known as the inter-record gap or IRG. The
next record follows the gap. If either the frame parity or the CRC is incorrect
when the tape drive reads the tape, an error is generated by the drive. Newer
formats ( 1600 bpi and above ) generate a preamble and postamble to help
synchronize the read logic.

Coding

Tape is recorded in several ways. Older systems use Non Return to Zero
Immediate (NRZI) coding, and record with a tape density of either 200, 556, or
800 bpi ( bits per inch ) . Newer tapes use Phase Encoding ( PE ) and record at
1600 bpi, or they use Group Coded Recording ( GCR) and record at 6250 bpi.
There may be other forms of coding as well, but these are the most common.
The HP 7971 supports a density of 1600 bpi, the HP 797 4 and HP 7979
support both 1600 bpi and ( optionally ) 800 bpi, and the HP 7978 and HP 7980
magnetic tape drives support a density of 1600 and 6250 bpi.
The higher the density, the more information can be stored on a tape. On a
2400 foot tape, an HP 797 4 at 800 bpi can only store 22 MB of data, at 1600
bpi the HP 7974 can store 43 MB , while an HP 7978 storing at 6250 bpi can
write up to 140 MB of data to a tape at a rate of up to 16 MB per minute.

HP-UX Peripherals 1 2·5

 Write/Read Errors
Tape, in its usage for long-term archive and data interchange, is more
error-prone than disks . When your tape drive is reading from or writing to a
tape, and it detects an error, the normal procedure is to backspace the tape
12 over the record and retry the tape operation. An error message is reported to
the user only after the driver gives up . Many more tape errors are caused by
dirty tape heads than by real recording errors , so you should periodically clean
your tape drive as outlined in its service manual.
Tape drives do a form of reading-while-writing, and if the data is not properly
recorded, an error will be detected. The normal procedure is to backspace and
retry writing the record once, and if that fails , to backspace, write a long gap
and try again on a section of tape farther down. A long gap is several inches of
erased tape. That 's why we said an IRG is "nominally " 1/2 inch long.

Preventive Maintenance
There are several maintenance procedures for tape. A tape can be completely
erased (degaussed) , or the beginning of the tape can be discarded and a new
load point put on (stripped) . There is also a tape cleaning ·and certifying
machine that will knock off any loose oxide and check that the tape will record
properly over its full length (certified) . This always makes any data on the
tape unusable. Commerci al shops certify their tapes fairly often, and discard
them if they get too short or fail to certify. It is also an excellent idea to clean
the tape head and guides of your drive periodically as they tend to accumulate
loose oxide and other contaminants .

Tape Streaming
Tape drives are designed to transfer data by one of two methods - st art/ stop or
streaming .
The HP 7971 is a start/stop tape drive, and is designed to stop and restart
the tape fairly quickly. The HP 7971 transfers data with very little buffering
between the computer and the tape drive's read/write head; the drive must
stop the tape between records, and wait for the next record.
The HP 797 4, HP 7978, and HP 7980 are streaming magnetic tape drives.

1 2-6 HP-UX Peripherals

A streaming magnetic tape drive is designed to move continuously, reading
data from a buffer or writing data to a buffer, not stopping between records, as
start/stop tape drives do. Streaming increases the rate at which a tape drive
can write data onto tape. Before a tape drive can write data onto a tape, the
drive read/write head must be positioned at the proper place on the tape, 12
and the tape must move across the head at the proper speed. After writing a
record on the tape, if a streaming drive has already received the data for the
next record from the computer, it can continue to move the tape across the
head without slowing down to write the next record.
If the drive has not received the data for the next record after writing a record
on the tape, then the drive must reposition the tape. This involves stopping
the forward motion of the tape, backspacing the tape to some point preceding
the beginning of the next record to be written, stopping the tape, and waiting
for your computer to send the data for the next record. The average data
transfer rate is much higher when the drive streams than when it repositions,
especially for the HP 7978 . The HP 797 4 supports both a start-stop and a
streaming mode. The HP 7978 supports only a streaming mode. Both drives
are much faster than the HP 7971 when they stream. When they do not
receive data fast enough to stream, the HP 797 4's performance is similar to the
HP 7971 ; the HP 7978 and HP 7980 are much slower.

Immediate Response Mode

To help your computer send data fast enough to permit the drive to stream,
the HP 7974, HP 7978, and HP 7980 support immediate response mode.
Ordinarily the actions of your computer and the drive are serialized. Your
computer sends data to the drive. Then the drive writes the data to the tape.
After the data is written, the drive returns status information to the host
indicating whether the write succeeded or failed. When immediate response is
enabled the drive returns status before it writes the data to the tape.
This is accomplished by the drive buffering the data it receives from your
computer in high speed memory which is built into the drive. The transfer rate
between the host and this buffer memory is much faster than the transfer rate
would be if the drive transferred the data directly to the tape. Because the
drive returns status to your computer very quickly, the host 's and the drive's
activities overlap , so the average transfer rate to the drive has a much better
chance of being fast enough to permit the drive to stream. Even when the

HP-UX Peripherals 1 2-7

 drive has t o go through a reposition cycle, i t can still b e buffering additional
records from the host .
Even with immediate response enabled, the HP 7974 and HP 7978 tape drives
typically don't stream continuously because the programs running on the Series
12 300 don't collect their data from the disk fast enough to supply it to the tape
drive. However, they still perform faster than the HP 7971 stop / start tape
drive.
An identical concept applied to CS / 80 cartridge tape is referred to as
immediate report .

1 2-8 HP-UX Peripherals

DDS-Format Tape Drive
Digital data storage (DD S ) , based on digital audio-tape technology, provides
a means for storing large amounts of data more compactly than on nine-track
magnetic tape and faster than on cartridge tape. A single DDS tape cassette 12
fo r a HP Series 6400 Model 1300S tape drive can hold u p to 1 .3 gigabytes of
data.

Comparison of DDS-Format and Nine-Track Magnetic Tape

• data coding: DDS-format and nine-track magnetic tapes use different coding
format .
• EOD : When a DDS-format drive writes a sequence of records to tape and
then rewinds, it inserts an EOD area at the end of the recrods. Subsequent
reads ignore any data after the EOD area. Nine-track tape drives do not
perform any comparable activity.
• Write-protecting media: DDS-format cassettes use a small write-protect
tab to protect the media from being overwritten. Nine-track tapes use
write-protect rings.
• Online/Offiine: Online/offl.ine mechanisms are similar for both technologies.
Ifa cassette is loaded in a DDS-format drive and the device is power-cycled
or HP- UX is rebooted, the mechanism will cause the drive to go offiine. To
bring it back online, eject the cassette and reload it.
• Streaming and immediate report : DDS-format drives support both streaming
and immediate report modes .
The DDS-format drive responds to all HP-UX commands appropriate to
magnetic tape devices, including cpio , t ar, and mt .

Logical Organization of DDS-Format Tape

Both DDS-format tapes and nine-track magnetic tapes can be viewed as a
sequence of records and tape marks. Two concluding tape marks represent ·

logical EOD (End of Data) for both formats.

The following example illustrates the sequence of events that occur during a
transaction with the drive:

HP-UX Peripherals 1 2-9

 When a tape is loaded, the DDS-format drive going online and positions the
tape at LBOT (Logical Beginning of Tape) . In this example, two commands
are issued:
tar cf /dev/rmt/Omn dir_A
12 t ar cf /dev/rmt /Om dir_A
The first tar command causes a sequence of 10-KB records to be written.
When the archive is finished, the driver automatically writes two filemarks
(since the last operation is a write) and backspaces one file mark (since
the selected device was Omn , for no rewind on close) . The tape head is now
positioned between the two filemarks when the second tar operation is called.
The second sequence of tape records is written (writing over the second
filemark) . When the second archive is completed, the driver's close routine
writes two filemarks and rewinds (Om) the tape to LBOT. The tape then
logically looks like this:

File 1 File 2 EOD

I
I I �I� I 11 11 111
111
LBOT Records F i lem a r k 2 F i l emarks

Figure 1 2-2. Organization of DDS-Format Tape

Tape archives such as these are sometimes called filesets, since they usually
consist of many files . A tape can consist of one or more files or filesets, with
the files separated by single :filemarks . The last file is followed by a double
filemark, representing EOD . The mt command is used to space between the
files .

1 2- 1 O HP-UX Peripherals
For More Information
For information on DDS-format tapes and their operation, refer to the
following resources:
• User's manual for the HP Series 6400 Model 1300 DDS-Format Drive 12
( available as 1300H for HP-IB and 1300S for S CSI ) .
• Installing Peripherals manual, for information on customizing HP-UX for a
DDS-format drive.
• mt(7) manual page in the HP- UX Reference.

HP·UX Peripherals 1 2- 1 1
 Cartridge Tape Drives
In addition to nine-track magnetic tapes and DDS-format tapes,
Hewlett-Packard manufactures a series of 1 /4-inch data cartridge tapes used
12 for the installation and updates of HP-UX. The cartridge tapes can also be
used for inexpensive backups . These data cartridges, model HP 88140, have
most of the benefits of nine-track magnetic tape but are cheaper and easier to
handle. However, they do not offer the same level of data interchange between
non-HP-UX machines as the nine-track or DDS-format tapes and are much
slower to use.
Like nine-track magnetic tape drives , cartridge tape drives can stream data in
immediate response/report mode. See the discussion of "Immediate Response
Mode" earlier in this chapter.
HP 9 144/45 cartridge tape drives are supported on Series 700 systems . They
require an optional EISA HP-IB card, customization of your default df ile
with the HP-IB device drivers , and appropriate device files.

1 2· 1 2 HP·UX Peripherals
CD-ROM File System {CDFS)
CD-ROM is an acronym for Compact Disc-Read Only Memory, a technology
for mass distribution and easy retrieval of large amounts of information.
Compact Discs ( or CDs ) contain approximately 550 megabytes ( MB ) of data 12
per disk. The information on the CD is virtually permanent ; you can read data
from a CD, but cannot write to it . Data on CDs is prepared and mastered
using a specialized publishing process .
In addition to the information contained in this section, you can learn more
about CD-ROM technology from these HP manuals:
• CD-ROM: The Basics , part number 5954-9709-discusses the introductory
concepts of CD-ROM technology.
• HP Series 6100 Model 600/S CD-ROM Drive User's Guide , part number
A1999-90002-describes the HP Series 6100 Model 600 / S SCSI CD-ROM
Drive.
• HP Series 6100 Model 600/A HP-IB CD-ROM Drive User's Guide, part
number C l 707-90000-describes the HP Series 6100 Model 600 / A HP-IB
CD-ROM Drive. ( Note, the HP-IB CD-ROM drive is not supported on the
Series 700. )

Overview of Implementation
Thjs section discusses the file system format for CD-ROM data, and the
standards which define it.

CD-ROM Standard Documents

Currently, there are two standard documents which define the file system
format for CD-ROM. One was produced by the CD-ROM Ad Hoc Advisory
Committee ( popularly called the High Sierra Group , abbreviated HSG ) . The
standard document produced by this group is called The Working Paper for
Information Processing (Volume and File Structure of Compact Read Only
Optical Discs for Information Interchange). This document is available from
the National Information Standards Organization ( NIS O ) .
The second standard evolved from the HSG standard and was produced by
the International Organization for Standardizatfon ( ISO ) . The name of the

HP-UX Peripherals 1 2- 1 3
 standard document i s Information Processing (Volume and File Structure of
CD-ROM for Information Interchange ( IS O . 9660: 1988 ( E ) ) .
The file system formats described by these two standards differ in some areas ,
but are largely identical. HP-UX supports CD-ROM media that conform to
12 either specification, hiding the differences from you.

System Call Changes

Because the CDFS format is so different from that of HFS , significant changes
had to be made to the kernel in order to support it . These changes were made
in such a way as to hide low-level details from the user. The result is that all
system calls have the same calling syntax and the same behavior as they did
on read-only HFS . All system calls work transparently with CDFS , exactly as
they do with HFS . Those calls which cannot complete successfully on read-only
media, such as write(2), fail gracefully; they return errno=EROFS.
One new system call, fsct1(2), was added to enable users to get CDFS-specific
information from a CDFS volume. Even though fsctl currently supports only
CDFS file systems, its architect will permit its functionality to be expanded to
include HFS ( and any other file systems that may eventually be supported ) .
The calling syntax for fsctl is:

fsctl (fildes, command, outbuf, outlen)

where fildes is an open file descriptor for a file on a CDFS volume, command
is a value specifying the information desired, outbuf is a buffer in which
the information is placed, and outlen is the length of outbuf . The available
commands are:
CDFS_DIR_REC Returns the directory record for the file or directory
indicated by fildes. Maximum size for outbuf is 256 bytes.
CDFS_XAR Returns the extended attribute record, if any, for the file or
directory indicated by fildes . Maximum size for outbuf is
64 KB.
CDFS_AFID Returns the abstract file identifier for the primary volume
whose root directory is specified by fildes, terminated with
a NULL character. Maximum size for outbuf is 38 bytes.

1 2- 1 4 HP-UX Peripherals
CDFS_BFID Returns the bibliographic file identifier for the primary
volume whose root directory is specified by fildes,
terminated with a NULL character. Maximum size for
outbuf : 38 bytes .
CDFS_CFID Returns the copyright file identifier for the primary volume 12
whose root directory is specified by fildes , terminated with
a NULL character. Maximum size for outbuf: 38 bytes.
CDFS_VOL_ID Returns the volume ID for the primary volume specified by
fildes , terminated with a NULL character. Maximum size
for outbuf : 33 bytes .
CDFS_ VOL_SET _ ID Returns the volume set ID for the primary volume specified
by fildes , terminated with a NULL character. Maximum
size for outbuf: 129 bytes.
For example, suppose the volume ID for a CDFS volume is desired. The
following code retrieves this information:

# include < sys/types . h>

# include < sys/vf s . h>
# include <fcntl . h>
# include < sys /cdf s . h>
# include < sys/cdf sdir . h>
extern int errno ;
main ( )
{
int f ildes , ret ;
char outbuf [129] ;

f ildes = open ( 11 /cdrom11 , O_RDONLY) ;

if ( (ret = fsctl (f ildes , CDFS_VOL_ID , outbuf , 129 ) ) < 0) {

printf ( 11 Fsctl failed (Y.d) , errno = Y.d\n11 , ret , errno) ;
exit ( 1 ) ;
}

HP-UX Peripherals 12-15

 Note that for the following commands which retrieve volume-wide information:
CDFS_VOL_ID , CDFS_VOL_SET_ ID , CDFS_BFID , CDFS_CFID , CDFS_AFID
fildes should be an open file descriptor for the root directory of the volume.
12 For the others ( CDFS _DIR_REC , CDFS _XAR ) , fildes should b e an open file
descriptor for the directory or file of interest .

Command Changes
Compared to the kernel effort , only minor code changes were necessary in the
HP-UX commands. This is due to the fact that most of the low-level details
are hidden within the system call and kernel work, thus allowing commands to
function independently of file system type.
Most HP-UX commands work transparently under CDFS . A few commands,
however, are not applicable to CDFS volumes, either because of the read-only
nature of the medium, or because they are inherently HFS-oriented. These
commands are unsupported under CDFS , and include the following:
clri( lM)
convertfs( lM)
diskusg( lM)
fsck( lM)
fsclean( lM)
fsdb(lM)
mediainit( lM)
mkf s ( lM)
ncheck( lM)
newf s ( lM)
tunef s ( lM)
No new commands have been added as a result of CDFS support .

Enabling CDFS
Before being able to mount and use CDFS volumes, the appropriate kernel
code must be configured into the kernel.

1 2· 1 6 HP-UX Peripherals
Series 300/400/700 Configuration for CDFS
To add the CDFS code to the Series 300/400/700 kernel, add the cdfs keyword
into the appropriate df ile. If the S CSI driver is needed for the CD-ROM drive
that 's been chosen, add the scsi keyword as well. Reconfigure the kernel using
conf ig and make , or the sam program. 12

A block special file is needed to mount the CDFS volume. If desired, a

character special file can also be created for raw access to the CD-ROM data.

Series 800 Configuration for CDFS

To configure the Series 800 kernel, edit the appropriate S800 file and add:
include cdfs ;
Also, make sure a location is available for the HP-IB CD-ROM drive in the I/O
configuration portion of the S800 file. Regenerate the kernel using uxgen.
A block special file is needed in order to mount the CDFS volume. If desired, a
character special file can also be created for raw access to the CD-ROM data.

HP-UX Usage
Once the kernel has been reconfigured and the necessary device files created,
the CDFS volume can be mounted. Mounting a CDFS volume is done in
exactly the same way as mounting an HFS volume.
Once mounted, HP-UX commands can be used to list, edit ( without making
changes ) , print , or copy files on the CDFS volume. Any executable files on
the CDFS volume can be directly executed on HP-UX. Other CDFS or HFS
volumes can be mounted on top of the mounted CDFS volume. In short ,
anything that does not require modifying the data on the CDFS volume may
be done.
Many things can also be done programmatically. Files may be opened, read
from, stated, checked for accessibility, etc., via the normal HP-UX system calls
and library routines. Again, only those library routines and system calls which
do not require modifying the CDFS volume data can be used successfully.

HP·UX Peripherals 12-17

 Optical Technology

Rewritable Optical
12
Traditionally, mass storage solutions have fallen into one of two categories­
primary or secondary storage. Primary storage is typically one or more fixed
magnetic hard disks. It is fast , random-access storage with moderately high
capacity and is used as the online system disk.
Primary storage is used to store applications, heavily accessed databases,
or files on which you are currently working and using extensively. It is also
used for applications needing virtual memory, fast data processing, report
generation, and complex calculations.
Secondary storage has consisted of one or more offiine storage devices-usually
a 1/4-inch or 1/2-inch tape drive, or a flexible disk drive on smaller systems.
These devices are used primarily to back up the system disks and are also used
for logging transactions, distributing software, archiving historical data, and
exchanging data between systems.
There's a gap between primary and secondary storage in terms of access time
and cost per megabyte. Average access time for hard disks is measured in tens
of milliseconds . Access time for information on tapes is measured in tens of
seconds for a mounted tape, minutes to hours for a tape in the library. The
cost of magnetic disk storage is a few dollars, while tape storage costs a few
cents per megabyte.
Rewritable optical fills this gap . With an average access time in the . 1- to
10-second range, and a cost of a few cents per megabyte more than tape
storage, optical drives create a new layer in the storage hierarchy called Direct
Access Secondary Storage (DA S S ) . Figure 12-3 illustrates this concept .

1 2- 1 8 HP-UX Peripherals
 Traditional Mass
storage Devices

New Mass Storage

Device: DASS 12

Primary Storage
• High Perfonnance Hard Disks
Performonce: Access in
m i l l iseco n d s

Direct Access Secondary Storage

• Rewritable Optical Disks
Performance: Acc e ss i n
se c onds

Cost: /
Cents M byte

• Applications
/
H istoric a l Archiva l
Stora g e
Unattended Backup/Restore
Docume n t Storage & Retrieva l
• Sequential Tape Drives for
Backup/Interchange

Performance : Acc e s s in
minutes
Cents/Mbyte

Figure 1 2-3. Direct Access Secondary Storage

Why Use Rewritable Optical?

Rewritable optical technology has many strengths. The disks are more
durable than other media, have greater storage capacity, are more reliable, are
removable, and cost far less per megabyte than magnetic disks. Rewritable
optical disks are good when you need direct access to a large amount of
traditionally "offiine" information such as archival or backup files .
The major disadvantage of optical drives is an access time slower than that of
hard disks. Due to the weight of the optical head, with its laser, lenses, and
mirrors, today's access times for optical disks are 2 to 5 times slower than high

HP·UX Peripherals 1 2· 1 9
 performance magnetic hard disks . Therefore, the optical disk library system
would not be good as a hard disk replacement .

Hewlett-Packard's Rewritable Optical Products

12
The Hewlett-Packard rewritable optical products range from a standalone drive
for small systems to an autochanger multi-disk setup for large systems and
networks. The following products are supported on HP-UX 6.5 and subsequent
releases of HP-UX.

HP Series 6300 Model 650/A - Optical Disk Drive

The Model 650/ A is a standalone rewritable optical disk drive. It uses 5 .25
inch ( 130 mm) magneto-optical disks (a type of rewritable optical technology
that complies with the continuous composite (C*C) format ) . One optical disk
holds 650 MB of data (325 MB per side) . However, because there is only one
read/write head assembly in the drive, you must eject the cartridge and :flip it
over to access the second side.
Using a 2400-rpm rotational speed, the Model 650 /A achieves a data transfer
rate of at least 340 KB per second. It connects to the host system with a SCSI
interface and can be accessed by S CSI commands as a conventional magnetic
disk drive.
You can use an optical disk drive like any other disk drive. You can rewrite
data an unlimited number of times and can store the disks for at least 10
years making data storage worry-free. Unlike 1/2-inch tape that you must
re-tension to keep archived files readable, MO disks are maintenance-free and
very durable. In addition, because of their small size MO disks require less
storage space than 1/2-inch tapes.

Optical Disk Drive Guidelines

The Model 650/ A can be used as a temporary or emergency boot device. If
you use the optical disk drive as a boot device, you must insert the disk before
you turn on the system, and you must not remove the disk until after you turn
off the system.
If you use the optical disk drive as a part of your mounted file system, do not
eject the disk until you have unmounted it .

1 2-20 HP-UX Peripherals

HP Series 6300 Model 20GB/A - Optical Disk Library System
The Model 20GB/A Optical Disk Library System ( ODLS ) is a device which
allows automatic, convenient access to vast amounts of information. The disk
library system can have a library of up to 32 magneto-optical disks, offering a 12
total capacity of 20.8 gigabytes.
The disk library system consists of several elements: one or more optical drives,
slots to store the 32 disks, a mechanical picker, and a mail slot which lets you
insert and remove disks .
. Within the disk library system cabinet the mechanical picker, guided by the
host computer, selects , moves, rotates, and inserts disk cartridges into the drive
mechanism ( s ) . Figure 12-4 illustrates the disk library system concept .
Since the disk library system drive ( s ) are the same as i n a Model 650/ A , you
can use the same optical disk in either of the two devices.

HP·UX Peripherals 12-2 1

 �

·
.::
·•
• • '1
.. · ..�. .
.
.

.
.
. ,. .

·. '

.. ,
;
..

:
"
,

\
.

•
Figure 1 2·4 Optical Disk L "brary
• . System concept

1 2-22 HP·UX Peripherals

Optical Disk Library System Guidelines
Each of the 64 surfaces in the disk library system (32 disks, double-sided) is
treated like a hard disk. A surface can hold a file system, can be stored to
and rewritten like a hard disk, has a device file associated with it , and can be 12
accessed with the same HP-UX commands as a magnetic disk.
While each surface on the disk is treated like a magnetic disk, the disk library
system is not a hard disk replacement . With this in mind, follow these
guidelines when using an optical disk library system:
• Boot Device: Do not use any surface in the optical disk library system as the
boot device. It is not supported as a boot device.
• Virtual Memory Swap Space: Do not put swap space on any surface in the
disk library system. Swap space is not supported.
• Directories on the Search Path: Do not put directories that are on optical
disk surfaces in your search path ( defined by the PATH environment
variables) . This could cause performance delays .
• Mounting Surfaces:
o Mount no more than one surface in any branch of the file tree.
Although you can mount surfaces anywhere in the file system, for best
performance no more than one surface should be mounted in any branch of
the file tree. Addition al disk exchanges will result when traversing these
paths.
o Minimize the number of surfaces mounted for writing.
You can have as many read/write surfaces as you want ; however you
should minimize the number of surfaces mounted read/write.
If the system's power fails, all write-mounted surfaces will require a file
system check (fsck). The time required to check the file system if you
have many optical disk surfaces mounted could be excessive.
o Do not put surfaces in /etc/checklist .
This slows down the booting process considerably since each surface must
be exchanged to be mounted and possibly file system checked. A better
strategy is to write a script to mount surfaces after the system is booted.

HP-UX Peripherals 1 2-23

 This way, activity on the rest of the system can proceed while the surfaces
are being mounted.
• Copying Files:
o Don't copy files straight from side A t o side B of the same disk. Doing so
12
requires many exchanges. Rather, copy the file intermediately to the hard
disk and then from the hard disk to the other optical surface.
o If you have a one-drive Disk Library System, don't copy straight from one
optical disk surface to another.

1 2-24 HP-UX Peripherals

Terminals and Modems
The following sections explain some conceptual information on terminals, and
configuring terminals into your system. The number of terminals and ports you
can add depends on the model and configuration of your computer system. 12
Use the procedures i n the System Administration Tasks manual t o set up
terminals.

Terminal Configuration
The Installing Peripherals supplied with your computer discusses the hardware
aspects of connecting a terminal to your system. This section offers the
software configuration information you need.
After connecting the hardware, terminals must be configured so they can
communicate with HP-UX. If a particular configuration option is not available
on your HP terminal, the option is already properly chosen ( as a default value )
by the terminal.
Use the Installing Peripherals to determine the typical terminal and data
communications parameters. The manual supplied with the terminal describes
how to use the function keys to configure the terminal. Generally, you press a
key that chooses the "terminal configuration" option and alter the appropriate
fields by answering prompts from the terminal's configuration program.
Except when using the terminal as a system console ( discussed below ) , you
may use any baud rate that the terminal can handle. The baud rate setting
on the terminal must match the baud rate parameter in the getty command
located in the terminal's entry in the /etc/ inittab file as discussed below.
Otherwise, it must reference a gettydefs entry that cycles the baud rate to
select the correct one.
If you are using the terminal as the system console, set the terminal's baud
rate at 9600. After the system is installed and running, you may change some
of the configuration parameters to suit your own needs. Further information is
provided in the next section ( "Special Considerations for Terminals" ) and in
getty(lM), gettydef ( 4) , and inittab( 4) in the HP- UX Reference.

HP-UX Peripherals 1 2-25

 Special Considerations for Terminals
For terminals, as well as modems, printers, or X.25 packet
assembler/disassembler (PAD ) that connect via RS-232 cable,
12 consider the following:
• Determine if a MUX card needs to be added or moved. On Series 800
systems using CIO /Mid-bus architecture, several MUX cards are already
configured in the CIO slots. To change the configuration, you need to
reconfigure the system using uxgen.
• Select which port to use for the device. On Series 800 systems attach the
device to an available port on the terminal port distributor (TP D , also
known as the junction panel).

/etc/inittab Entries for Terminals

To add a terminal to the system, you must perform the steps described
in System Administration Tasks manual. Part of this procedure involves
adding entries to the /etc/inittab file. This file displays a login prompt on
configured terminals. /etc/inittab is described in Chapter 2, "Constructing
an HP-UX System" and Chapter 4, "Controlling Access to the System" of the
System Administration Tasks manual. This section discusses entries specific to
terminals.
Most terminal entries in /etc/inittab have the form:

id: run-level: action : /et c/getty -txxx specialfilename N # comment field

Typical values for the id, run-level, and action fields are:
• id is a unique string up to 4 characters in length identifying the entry

• run-level is 2 (the default multi-user run-level)

• action is respawn; this indicates that the specified command (getty in this
case) is to be re-invoked once the process terminates (typically, when the user
logs off) .
The process field contains the /etc/getty command for a terminal. It can be
followed by up to three parameters:
-txxx is the optional time-out option for use with modems.

1 2·26 HP·UX Peripherals

specialfilename is the filename of the terminal's or modem's character
special file ( e.g. , tty0p4 ) , but not the complete pathname
( e.g., /dev/tty0p4) . The named file must reside in the
I dev directory.
12
N specifies a speed indicator for getty; a value of 9600 is
common for hardwired (9600 baud terminal ) lines, a value
of 1200 is common for dial-up (300/ 1200 baud modem )
lines. Graphics displays ignore the speed indicator.
The N parameter indexes into the / etc/gettydefs file. It
is possible to set up a terminal to cycle through 22 different
baud rates by formatting the gettydef s file properly. This
file supplies information on the speed and terminal settings
and what the login prompt should be. If the user enters a
"break" on the terminal keyboard, it indicates that the
terminal speed is incorrect , which causes getty to use the
next entry from the gettydefs file.
( For details on getty parameters , see getty(lM) in the HP- UX Reference.)
On a multiuser system, be certain t o set u p / etc/ inittab terminal entries
for each terminal connected to the system. For example, to add a terminal on
/dev/tty0p4 the etc/ inittab could be:
p4 : 2 : respawn : / etc/getty tty0p4 9600 #t erminal at rob ' s desk
An id of p4 was chosen to correspond to the last two digits of the terminal's
special file, although any id up to 4 characters long would have sufficed. In
general, though, try to use ids that enhance the entry's readability. The
run-level is 2 ( multi-user, default run-level ) , and the action field is respawn,
causing getty to restart and the "login:" prompt to be displayed again after a
user logs out .

HP·UX Peripherals 1 2·27

 HP-I B Guidelines
Depending on your computer system, you can connect to an HP-IB bus
through one of several I/O paths, as summarized in the following table:
12
Table 1 2- 1 . HP-IB Connectivity
Series Type 1/0 Path
300 Internal Human Interface Board
External DIO I /II
400 External EISA
700 External EISA
800 External HP-PB
CIO

All external HP-IB connections require that your system have HP-IB I/O
hardware (optional on some systems). In all cases , you will need appropriate
device drivers configured into the kernel and device files created. Each device
attached to the bus must have a unique HP-IB address. See the Installing
Peripherals for further information.
When connecting multiple HP-IB devices, daisy-chain configurations are more
reliable than multiple devices connected at a single point .
Cabling requirements affect HP-IB operating speed. Consult the Installing
Peripherals and documentation shipped with your peripheral for guidelines.
When using HP-IB and other multiple-access external buses, consider what
priority the peripherals should have for accessing system resources. HP-IB
connections provide either standard-speed or high-speed data transmission,
depending on the slowest peripheral connected to the bus. (The exception to
this is the standard-speed internal HP-IB connection on the Series 300, which
transmits data only at standard speed. )
For example, a high-speed HP-IB might be used t o advantage for accessing
the system disk containing root and swap , while standard-speed HP-IB might
be configured with a cartridge-tape drive, which takes longer to respond and
might interfere with higher-speed devices .

1 2-28 HP-UX Peripherals

The system printer can be installed on a standard-speed HP-IB interface, with
bus address of 1. Install the printer separate from the system root device.
( The printer can be on high-speed HP-IB , as long as there are no high-speed
devices on the same card. ) Printers and plotters cannot be on the same HP-IB
interface as disks. 12

On the Series 300 and 400, the system root device ( hard disk ) is typically
located at select code 14, bus address 0 on a HP 98625B or HP 98262A
high-speed disk interface card.
On a Series 300, the internal HP-IB is pre-set at select code 7. You can
change this select code in the Boot ROM configuration mode. The Boot ROM
configuration mode is discussed in the Appendices of the Installing Peripherals
manuals. On the Series 300, 400, and 800, consider the following factors:
• An HP 7974 or HP 7978 nine-track tape drive should be placed on a
high-speed HP-IB interface, if possible. You can also use that same HP-IB
for the root device.
• Plotters and the HP 9 1 1 1 graphics tablet should be placed on separate
standard-speed HP-IB interfaces when possible. Typical HP-IB addresses are
5 and 6 for plotters and graphics tablets, respectively.
• A void putting flexible disk drives and cartridge tape drives on the same
HP-IB interface as the root device, except as noted above.

HP-IB Electrical Guidelines

• Always turn on the HP-IB device before turning on the computer system.
Turn off the computer system before turning off the HP-IB device.
• Never turn on or off an HP-IB device while connected to a computer system
that is turned on. This could corrupt data on the HP-IB bus.
• All devices attached on an active HP-IB subsystem should have their power
switched on. Devices with power switched off load the bus to excess.
• Never connect or disconnect an HP-IB device while the computer system is
1

running or the bus is active.

• Before resetting switches on any HP-IB device, turn off the power to the
device.

HP-UX Peripherals 1 2-29

 SCSI Device Guidelines
A SCSI bus may use either single-ended ( "unbalanced" ) or differential
( "balanced" ) electrical signal lines to transfer data. Do not mix single-ended
12 and differential devices and buses; they are electrically incompatible.
The SCSI interface can support multiple SCSI devices simultaneously. A bus
terminator must be attached to the end of the chain of devices, or the bus will
not operate. The terminator must be compatible with either the single-ended
or differential S CSI bus .
See the system configuration guides, Installing Peripherals, and the
documentation shipped with your peripheral device for specific information
about which SCSI devices are supported by your system, whether your
device is single-ended or differential, cabling requirements and terminating
considerations.
Consider also the following list of SCSI guidelines:
• Use of third party peripherals is at user's risk and is not supported by
Hewlett-Packard.
• Never attach a single-ended S CSI device to a differential SCSI bus, or vice
versa.
• Each end of a SCSI bus must be terminated. Terminator resistors are
typcially installed in the host adapter; thus, the host adapter must be on one
end of the SCSI bus. This terminator provides matching impedance on the
bus circuits. Without the terminator the bus will not work.
• The last SCSI device in the chain, even if it is the only one, must have
a terminator installed to its second connector. This terminator provides
matching impedance on the bus circuits. Without the terminator the bus will
not work.
• Make sure there are no unterminated cables ( that is, all cables are attached
to a device at both ends ) .
• All devices must be connected to a common ( single point ) system reference
ground. The system ground must be isolated from other electrical devices
such as copying machines , arc welders and air conditioners . HP supplied
cables supply correct grounding.

1 2-30 HP·UX Peripherals

• Keep all devices powered on during and after system boot-up.
• Do not connect or disconnect any SCSI devices while the system is powered
on.
• Do not turn power on or off to any S CSI device while connected to a 12
powered-up system. This could result in data corruption or a system panic,
which in turn might lead to file system corruption.
• All devices must have an unique bus address between 0 and 6. The host
adapter's SCSI bus address is typically 7.
• S CSI bus address 7 is reserved for the SCSI interface. Do not use it for SCSI
device bus addresses.
• The BootROM searches from bus address 6 to 0. Your root disk should have
a higher bus address that any other device on the bus. It is recommended
that the root disk be set to SCSI address 6 .
• If you need t o change the bus I D on a SCSI device, perform the task i n the
following sequence:
1. Shut down your system and turn power off.
2. Turn off the device.
3. Change the bus ID on the device.
4. Turn on the device. Power on all SCSI peripherals and make sure they
have time to complete their selftest before powering on the SPU ( System
Processor Unit ) .
5. Reconnect the power cord to your system.
6 . Turn on your system.
• Because SCSI cable impedance and construction can have a significant effect
on signal quality, only HP cables are recommended.
• Ensure that the total cable length ( including external and internal cables )
does not exceed SCSI bus limits. For single-ended SCSI, the limit is 6
meters. For differential S CSI, the limit is 25 meters.
The length of the SCSI bus should be kept as short as possible. However, do
not use cables less than 0.5m in length. Refer to your system configuration
guides for internal cable length of supported peripherals.

HP-UX Peripherals 1 2-3 1

 13
Networking

Networking enables you t o transfer data among different systems. Networks

can be simple (for example, linking just a few machines of the same type) or 13
powerful (having machines from different vendors with different operating
systems on each).
Most networking capabilities require that you have optional networking services
installe d. What services you have depend on what options you used when
ordering HP-UX.
This chapter summarizes briefly:
• Network architecture and the O SI model
• HP-UX networking software
• HP-UX networking documentation

Networking 1 3· 1
 Network Architecture and the OSI Model
A network architecture is a structured, modular design for networks. HP 's
network architecture conforms to the reference model of Open Systems
Interconnection (OSI), a network architecture mode developed by the
International Standards Organization (IS O ) . NS, ARPA, and NFS Services are
implemented based on the OSI model.
In the OSI model, connectivity is divided among seven logically distinct
modules called layers , each of which performs a specific data communication
13 function. Interfaces between layers allow communication with adjacent layers
or with a peer layer on a remote computer.
Table 13- 1 shows each layer and its function.

Table 1 3 - 1 . OSI Model and Layers

Layer and Name Layer Function

7-Application Provides user services, such as file transfer and terminal access
to the network .
6-Presentation Manipulates user data such as in data compression.
5-Session User 's interface into the network. Establishes and manages
connections between users.
4-Transport Responsible for data integrity, ensures that data arrives at a
remotely networked computer without errors, controls flow,
retransmits corrupt data, suppresses duplicated data.
3-Network Routes messages from one node to another.
2-Data Link Packs raw bits for transmission , detects transmission errors,
controls access to the media.
1-Physical Defines the electronic characteristics of the transmission media.

1 3·2 Networking
HP-UX Networki ng
Networking software corresponds to specific layers of the OSI model and can be
thought of in three broader categories:
• links, corresponding to layers 1 and 2 .
• transports, corresponding t o layers 3 through 6.
• services, corresponding to layer 7.
Table 13-2 shows the relationships of links and transports to HP-supported
networking services . For example, the LAN link enables access to Network 13
Services (NS) through the TCP /IP transport layer or to FTAM through the
OTS layer.

Table 1 3-2. HP 9000 Networking Products

Links Transports Services

LAN TCP/IP NS, ARPA, NFS
OTS X .400 , FTAM
X .25 TCP/IP NS, ARPA
OTS X .400 , FTAM
802.4 HP OSI Express FTAM, MMS

Some of these components are bundled into integrated products. For example,
the TCP /IP transport is bundled with LAN and X.25 products; the 802.4 link
is included in the HP OSI Express MAP 3.0 product .

Networking 1 3-3
 Networking Documentation
For a good overview of all networking services, refer to the manual Networking
Overview . Also, networking reference pages are located in the HP- UX
Reference . The designation (3N) denotes networking library.
As a system administrator, you are most likely to deal with networking
issues from the perspective of services. Table 13-3 summarizes the services
implemented in HP-UX and manuals available for each element in the
preceding table. (Manuals can be ordered through your local HP sales
13 representative.)

1 3·4 Networking
 Table 13·3. Available Networking Services

Service Capabilities Related Documentation

ARPA/ Connects HP 9000 to other HP 9000 • Using ARPA Services
Berkeley systems or multivendor systems running • Installing and
Services ARPA/Berkeley Services from HP or A dministering ARPA
non-HP software supplier. Requires Services
optional ARPA Services software. • Berkeley !PC
Programmer's Guide

NFS/9000 Allows users on a networked computer to • Using NFS Services 13

access files from another HP-UX o r UNIX • Installing and
system over the network as though those A dministering NFS
files resided on the client 's file system. Services
Requires optional NFS Services software. • Programming and
Protocols for NFS Services

UUCP Enables users to transfer files among UUCP volume of HP-UX

HP-UX and other UNIX systems, and to Con cepts and Tutorials
execute commands on remote systems.
Computers must have UUCP installed
and be connected via modem lines.
UUCP utilities reconcile differences in file
format between systems. UUCP is
provided as part of standard HP-UX.
NS/9000 Provides network file-transfer capability • Using Network Services
between HP 9000, HP 3000, HP 1000 and • Installing and
multivendor systems running NS, HP's A dministering Network
proprietary network implementation. Services
Used to access applications and • LLA Programmer's Guide
databases. Requires optional Network • NetIPC Programmer's
Services software. Guide
• NS Cross-System NFT
Reference Manual

Networking 13·5
 Table 1 3-3. Available Networking Services (continued)

Service Capabilities Related Documentation

OTS/9000 Provides networked communication using • Installing and
OSI Transport Services protocols between Administering HP OSI
multivendor systems. HP-UX OSI Transport Services
supports FTAM (File Transfer, Access, • HP OSI Express 802.4
and Management service) , X .400 (a Hardware Installation and
standard electronic mail service) , MMS Configuration Guide
(Manufacturing Message Specification • Installing and
13
used for multivendor manufacturing A dministering HP OSI
automation applications) . Includes Express MAP 3. 0 Link
programmatic access to the transport • Troubleshooting HP OSI
layer via the X/Open Transport Interface Express MAP 3. 0 Link
(XTI) . Requires optional OSI Transport • A dministrator's Reference
Services software. Manual for HP OSI
Express
• XTI Programmer's Guide
• Session Level
Programmer 's Guide
X.400/9000 Enables HP 9000 system users to sending Installing and Administering
and receiving electronic mail messages X.4 00/9000
throughout a multivendor network.
Requires optional X.400 software.
X .25 Connects HP 9000 systems to public and • X. 25 Programmer's Guide
private X .25 wide-area networks (WANs) . • X. 25: The PSN
Connection
LAN/9000 Provides network connection from HP • Installing and
9000 systems to Ethernet and local area Administering LAN/9000
networks. Enables information exchange Series 300
through NetIPC and Berkeley Sockets, • Installing and
two interprocess communication facilities. Administering LAN/9000
Allows implementation of link-level access Series 800
(L L A) for special-purpose network • LLA Programmer's Guide
protocols to access network interface • Berkeley IPC
drivers directly. Requires optional Programmer 's Guide
LAN /9000 software . • Net/PC Programmer's
Guide

1 3-6 Networking
 Table 1 3-3. Available Networking Services (continued)

Service Capabilities Related Documentation

FTAM/900( Provides multivendor file system • HP FTAM/9000
interoperability, permitting text and • HP FTAM/9000
binary file transfer , access to remote files, Programming Guide
or management of remote files of any • HP FTAM/9000 User's
defined data type between HP9000 and Guide
diverse OSI end systems , regardless of
hardware or operating system. Requires
13
option al FTAM/9000 software.
MMS/9000 Enables MAP-connected cell controllers • MMS/9000 Reference
to direct control of MAP 3.0 conformant Manual
factory-floor devices via • MMS/9000 Programmer's
program-to-program messaging and Guide
binary file transfer services.

Networking 1 3-7
 14
System Accounting

HP-UX allows users to share computer resources such as disk space,

memory, and the CPU. Some computing environments, such as universities
and government contractors, need to keep and report detailed accounting
information on the use of these resources. Time-share environments also
require detailed records of resource use.
HP-UX is equipped with programs and scripts that tabulate resource usage
14
and generate reports. In addition, you can develop custom scripts that use the
available HP- UX accounting tools.
This chapter introduces the principles and tools of system accounting, as
indicated in the next section. For accounting procedures, refer to the System
Administration Tasks manual for your system.

System Accounting 1 4· 1
 What I s in This Chapter?
HP-UX system accounting allows you to accomplish accounting tasks using
many versatile commands. This chapter describes concepts underlying use of
these accounting commands in the following sections:
• "Accounting System Planning and Billing" suggests how to approach setting
up and using an accounting system.
• "Overview of System Accounting" gives background information useful for
using system accounting.
• "System Accounting Set-up Guidelines" summarizes what happens when you
set up , enable and disable system accounting.
• "System Accounting Files" shows the directory structure and contains brief
definitions of system accounting files .
14
• "Disk Space Usage Accounting" illustrates the use of the accounting
commands that monitor disk space utilization on a per-user basis.
• "Connect Session Accounting" describes the commands that record and
report connect session accounting information.
• "Process Accounting" discusses methods for generating per-process
accounting data and reports.
• "User Fees" introduces an HP-UX tool for charging fees to users.
• "Accounting Summaries and Reports" surveys the daily and monthly
accounting reports that can be generated to monitor system performance and
bill users.
• "Disk Quotas" describes a system of tools that enable you to limit disk usage
by user or file system.

1 4·2 System Accounting

Accounting System Planning and Billing
As a system administrator, you might be asked to provide data that can be
used to bill individuals or groups for their use of computer resources .
To develop an equitable way of sharing costs, you'll need to gather some
information:
• how to group people for billing
• what criteria to apply for billing computer resources
Then, with the accounting software described in this chapter, collect
appropriate data and apportion costs by use. In this way, heavy and light users
are fairly charged.

Grouping Users for Billing

14
You can group individuals logically for data collection purposes, so that
resource costs can be billed to these groups in proportion to the amount of
resources they consume. You might do this by using a small database with
login name or user ID number associated with the entire group or functional
area.
Another possibility is to use the GCOS ( personal information ) field of
I etc/passwd, so that the information can be extracted using grep or awk.
Note that the chfn(l) command allows users to change their entries in
I etc/passwd, so with this approach, you might want to restrict access to chfn.
Data collected with HP-UX accounting tools and a grouping technique can be
used as input to a more global spreadsheet or charging scheme.

Criteria for Billing Computer Resources

Find out what criteria or costs are key to account for on a given machine.
For example, system-administration time for tasks such as backups might be
critical in a heavily file-system based environment . In an NFS cluster or file
server, you might care more about CPU cycles. Figure out the most critical
elements , use the appropriate accountancy tools to derive your data, and
apportion charges based on your findings . Thus, you might decide to apportion
charges as follows, based on what you perceive as key costs:

System Accounting 1 4-3

 100% charges = 30% CPU + 703 disk space usage

Time Considerations
When collecting data by connect-session and process accounting, you can bill
usage of computer resources differently during busy or less busy periods. When
reporting computer time usage, system accounting distinguishes between prime
and non-prime time, which is defined in the /usr/ l ib /acct/hol idays file.
Prime vs Non-prime Connect Time. Prime time is the time during the day when
the computer system is most heavily used-typically from 9 :00am to 5:00pm.
Non-prime time is the remaining time during the day when the system is used
less-typically from 5:00pm to 9:00am.
Prime time is in effect only on weekdays (Monday through Friday). Non-prime
time is in effect during the weekends (Saturdays and Sundays) and on any
14 holidays specified in the hol idays file.
You can specify prime and non-prime time on your system by editing the file
/usr/lib/acct/hol idays .
Updating the Holidays File. The file /usr/lib/acct/hol idays contains the
information that System Accounting needs to distinguish between prime and
non-prime time. Here is a typical hol idays file:

1 4-4 System Accounting

 * Prime/Nonprime Table for HP-UX Accounting System
*
* Curr Prime Non-Prime
* Year Start Start
*
1990 0800 1700
*
* Day of Calendar Company
* Year Date Holiday
*
1 Jan 1 New Year ' s Day
50 Feb 19 Washington ' s B irthday
103 Apr 13 Spring Holiday
148 May 28 Memorial Day
185 Jul 4 Independence Day 14
246 Sep 3 Labor Day
326 Nov 22 Thanksgiving Day
327 Nov 23 day after Thanksgiving
358 Dec 24 Christmas Eve
359 Dec 25 Christmas
Note the following characteristics of the holidays file:
• Comment Lines-Comment lines begin with an asterisk ( * ) and can appear
anywhere in the file.
• Year Designation Line-This line should be the first non-comment line in the
file and must appear only once. The line consists of three four-digit numbers
representing year, time (in 24-hour format) that prime time starts , and time
that non-prime time starts.
The time field entry 2400 is identical to 0000 .
• Company Holidays S ince few people use the computer during holidays,
-

non-prime time applies for the 24 hours of those specified days.

Note As delivered, the holidays file contains valid entries for typical
prime/non-prime time, and holidays. Edit this file to reflect
your organization's requirements.

System Accounting 1 4·5

 Collecting Data Strategically
Having determined your billing plan, you can use accounting tools to determine
usage, collect statistics, and input data. This information can then be used as
input to either the chargefee billing command or a more global billing scheme,
which might also include variables such as occupancy and depreciation.

14

1 4-6 System Accounting

Overview of System Accounting
HP-UX commands let you:
• Monitor disk space usage of individual users.
• Record connect-session data ( logins /logouts ) .
• Collect resource-utilization data ( such as memory usage, execution times for
individual processes.
• Charge fees to specific users and groups.
• Generate summary files and reports used to analyze system performance and
bill users for resource consumption.
As described in "Accounting System Planning and Billing," system accounting
can be used effectively to quantify and bill for computer system usage.
Regardless of complexity, determining those statistics involves two aspects:
• Gathering and displaying data 14
• Generating reports
Figure 14- 1 shows the elements that contribute to this process in relationship
to one another. In a general sense, the "environment snapshot" refers to values
generated by the various commands at the time they are invoked. The content
of various files also can be merged for intermediate summaries , and the output
of one command is often used as input to other commands.

system reporting
accounting t---i� & summarizing reportS
commands commands

LG200 1 72_001

Figure 1 4- 1 . System Accounting Overview

System Accounting 1 4-7

 Gathering Data
HP-UX System Accounting provides numerous tools for gathering data,
which is then used to generate reports and possibly charge fees . Most system
accounting focuses on:
• Disk space usage
• Session accounting
• Process accounting
We discuss these categories one by one in subsequent sections of this chapter.

Automating Data Collection

Gathering data can be automated.
You can use the cron command to automate routine accounting tasks, by
14
setting cron to execute various accounting commands at regularly scheduled
intervals. For example,
• Run the ckpacct command every hour to ensure that the process accounting
file pacct does not become too large. If pacct grows past a set maximum
number of blocks, turnacct switch is invoked, which creates a new pacct
file. ( Other conditions may also limit the size of the process accounting file
or turn process accounting off; for more details , refer to the discussion of
ckpacct in the "Process Accounting" section of this chapter. ) The advantage
of having several smaller pacct files is that runacct can be restarted faster if
a failure occurs while processing these records.
• Execute runacct each night , to process active fee files, as well as
process-accounting, connect-session, and disk-usage accounting files. The
runacct command produces command and resource-usage summaries by
login name.
A report of the previous day's accounting information can be created by
running prdaily , once runacct has been executed.
• The chargefee program can be used to charge fees to users. It adds records
to the file fee. These records are processed during the next execution of
runacct and merged in with total accounting records .
These suggestions will make more sense once you have read through the
chapter and decided upon your overall approach.

1 4·8 System Accounting

Generating Reports
Raw data either represents a snapshot of the state of the computing system
or is generated by system-accounting tools and collected in files from which
reports can be generated. Flow charts in subsequent sections show how data is
synthesized into assessible form, called total accounting records.

Total Accounting Records

These records , created by various accounting commands, contain summary
accounting information for individual users. The format of total accounting
records conforms to a structure t ac ct . h, explained in acct( 4) of the HP- UX
Reference.
Here are some examples of information contained in these records:
• ID and user name of the target user 14
• Total number of processes a user has begun during a given accounting period
• Fees for special services rendered to the user
These records provide the basic information for many reports generated by
system accounting. Commands covered in later sections of this chapter show
how these records are created and used by system accounting.

Summary Reports
The data tabulated in total accounting records can provide insight on specific
areas. They can also be combined to produce comprehensive summaries
of computer resource usage. Several tools for summarizing and reporting
accounting information are discussed later in this chapter.

System Accounting 1 4-9

 System Accounting Set-up Guidelines
Since not all computing environments require accounting capabilities, and
accounting does impact performance, HP-UX System Accounting is provided as
a selectable fileset. Before you can generate accounting data and reports , you
must enable the accounting software.
To set up system accounting, you perform three steps:
1 . Update / etc/re to automatically start System Accounting when the system
switches to multi-user mode.
2. Create crontab entries to automate creation of accounting data.
3. Set PATH to include the directories where accounting commands reside.
These steps are discussed in detail in the System Administration Tasks manual
for your system.
14
Administering System Accounting in a Cluster (Series 300/400/700
only)
System accounting in a clustered environment is straightforward, as long as you
realize where data is stored and processes execute. Some issues need special
consideration; for example, on an HP-UX cluster:
• You must enable or disable accounting on each cluster client .
• Some tasks , such as disk-space accounting, provide useful data only when
administered from the root server.

Note Throughout this chapter, we note cluster-specific details. For

further information on clusters, see Managing Clusters of HP
9000 Computers: Sharing the HP- UX File System.

1 4· 1 0 System Accounting
How System Accounting is Turned On
When HP-UX is switched into multi-user mode, the system initialization shell
script re executes the command st artup , which starts system accounting by
performing the following functions:
• Calls acctwtmp to add an initial record to / etc/wtmp . This record is marked
by storing "acctg on" in the device name field of the I etc/wtmp record.
• Turns on process accounting via turnacct on, which executes accton with
the filename argument /usr/adm/pacct . This directs the kernel to collect
process-accounting data in the file /usr/adm/pacct . ( Additional information
on turning on process accounting is found later in this chapter. )
• Removes work files left in the /usr/adm/acct/sum directory by runacct .

14

.._
.. .. /ate/wt. m p
....,___,,

/ate/re

remove work
files from
/usr/adm/acc1/su
directory

LG200 1 72_1l02

Figure 1 4-2. How startup Enables System Accounting

System Accounting 14- 1 1

 How System Accounting is Turned Off
When a system that runs accounting is turned off using shutdown, the
shutacct command is executed. The purpose of shutacct is to stop System
Accounting by performing the following functions:
• Writes a termination record to / etc/wtmp by calling the command acctwtmp ,
which writes "acctg oft" in the device name field.
• Turns process accounting off by calling turnacct off . ( Additional
information on turning off process accounting is found later in this chapter. )

Logging I n for System Accounting

To maintain the integrity of accounting data files, we recommend that you
establish a separate accounting user account , with password, for use when
14 performing all accounting functions.
The login name for system accounting is adm; the user ID for adm is 4. The
adm login is a member of the group adm, whose group ID is also 4. The home
directory for the adm login is /usr I adm.

1 4- 1 2 System Accounting
System Accounting Files
System Accounting uses a multi-level directory structure to organize its many
accounting files, some of which are permanent and others temporary. Each
directory in this structure stores related groups of files, commands, or other
directories.

I
/ ( root )

usr

/ ""' etc
ho
odrn I
__
Active data ­
/ wtrnp
re
shutdown 14
col lection files

pocct

===---- 1 ------
acct

I I I I
nite sum fiscal acct

d a ta fi les daily s u m m a ry m o n thly s u m m a ry holidays file

p rocessed files crea ted files c reated and
daily by by runocct by rnonacct accounting
runocct commo n d s

Figure 1 4·3. System Accounting Directory Structure

System Accounting 1 4- 1 3
 • /usr/adm-contains all active data-collection files, such as pacct and fee.
• /usr/adm/acct-contains the nite, sum, and f i scal directories described
below.
• /usr/adm/acct/nite-stores data files that are processed daily by runacct .
• /usr/adm/acct/sum-cumulative summary files updated by runacct are
kept here.
• /usr/adm/acct/fiscal-periodic ( monthly ) summary files created by
monacct are stored here.
• /usr/lib/acct-System Accounting commands reside here.
• I etc-contains wtmp , and shell scripts re and shutdown.

Files in the /usr /adm Directory

Note that in an HP-UX cluster ( Series 300/400/700 only ) , the /usr/adm
14 directory is a context-dependent file (CDF).

Filename Contents
dtmp Output from the acctdusg or diskusg programs .
fee Output from the chargefee command (ASCII total accounting records ) .
pacct The current active process accounting file.
pacct? Prior process-accounting files switched via turnacct swit ch.

1 4-14 System Accounting

Files in the /usr/adm/acct/nite Directory

Filename Contents
act ive Used by runacct to record progress. It contains warning and error
messages. act ive mmdd is the same as act ive after runacct
detects an error.
ctacct . mmdd Total accounting records created from connect session accounting
where mmdd is the month and day the file was created.
ctmp Output of acctcon1-connect session records.
daycms ASCII daily command summary used by prdaily .
daytacct Total accounting records for current day.
disktacct Total accounting records created by the dodisk command.
fd2log Diagnostic output from the execution of runacct ( refer to crontab 14
entry ) .
lastdat e The last day runacct was executed, in dat e l # + %m %d format .
( Refer to d at e ( l ) in the HP- UX Reference for a description of
+%m%d date format . )
lock and lock1 Used to control serial use of runacct .
lineuse Terminal ( tty ) line usage report used by prdaily .
log Diagnostics output from acctcon1 .
logmmdd Same as log after runacct detects an error.
reboots Contains beginning and ending dates from wtmp, and a listing of
reboots.
stat ef ile Used to record the current state being executed by runacct .
tmpwtmp wtmp file , corrected by wtmpf ix.
wtmperror Error messages, if any, from wtmpf ix .
wtmperror mmdd Same as wtmperror after runacct detects an error .
wtmp . mmdd The previous day's wtmp file.

System Accounting 14· 1 5

 Files in the /usr/adm/acct/sum Directory

Filename Contents
ems Total command summary file for current month in intern al
summary format .
cmsprev Command summary file without latest update.
day ems Command summary file for previous day in internal summary
format .
loginlog Shows the last login date for each user.
rpt mmdd Daily accounting report for date mmdd .
taeet Cumulative total accounting file for current month.
taeetprev Same as taeet without latest update.
14 taeet mmdd Total accounting file for date mmdd .
wtmp . mmdd Saved copy of wtmp file for date mmdd . Removed after reboot .

Files in the /usr/adm/acct/fiscal Directory

Filename Contents
ems ? Total command summary for month ? in internal summary format .
fiserpt ? Report similar to prdaily for the month ? .
taeet ? Total accounting file for the month ? .

1 4- 1 6 System Accounting
Disk-space Usage Accounting
System accounting provides the means to monitor disk space used by individual
users. Before reading this discussion, you might want to review the "HFS File
System" chapter of this manual. Also, if you are interested in setting limits on
disk-space usage, refer to the section at the end of this chapter, entitled "Disk
Quotas."
Disk-usage commands perform three basic functions, which are illustrated in
the figure below:
• tabulate disk usage by file system
• report disk usage (in blocks) for individual users
• summarize disk usage in a form that can be merged with other accounting
records, to produce reports by commands such as prtacct or runacct .
14

file
find name
list
disk
clodlsk . Total
usage acctdlsk
Accou nti
report Records
file-system
I nodes

LG200172_003

Figure 1 4-4. Disk-space Usage Accounting

Reporting Disk Space Usage

Two commands-acctdusg and diskusg-report disk usage by showing the
number of disk blocks individual users are consuming. Each command has
slightly different options and differs in how it produces accounting information.
Note that neither acctdusg nor diskusg produce any output on a cluster
client that does not have a file-system disk attached to it . For information on
working in a clustered environment (supported on Series 300/400/700 only), see
Managing Clusters of HP 9000 Computers: Sharing the HP- UX File System.

System Accounting 14- 1 7

 Computing Disk Resource Consumption by Login-acctdusg
Acctdusg is a script that takes its input from a list of path names created by
the f ind command. For each file in the list , acctdusg identifies the owner
of the file, computes the number of blocks allocated to the file, and adds
this amount to a running total for the file's owner. acctdusg displays the
information accumulated for each user: user ID, login name of the user, and
number of blocks used.
The syntax for acctdusg is:

acctdusg [ -u no_owners ] [ -p p_file ]

Acctdusg has two useful options:
-u no_owners If-u is given, path names of any files without apparent
owners are written into the file no_owners . This option could
14
potentially find users who are trying to avoid disk charges.
A shell script called grpdusg, described in System
Administration Tasks, displays disk accounting information
for users in a given group and uses the -u option.
-p p_file The password file I etc/pas swd is the default file used by
acctdusg to determine ownership of files. If the -p option is
used, then acctdusg will use p_file instead. This option is
not needed if your password file is / et c/passwd.
For more information about acctdusg, refer to acct(lM) in the HP- UX
Reference
Generating Disk Accounting Data by User ID-diskusg
D iskusg generates disk usage information by examining the inodes of the file
systems associated with a disk, to create its output . ( Refer to inode( 4) in the
HP- UX Reference and the "File Systems" chapter of this manual for more
information on inodes. )
This command reports disk usage information in the same format as
acctdusg-user ID , login name of the user, and total disk blocks used, but
because it reads the code directly, diskusg is faster and more accurate than
acctdusg.

1 4- 1 8 System Accounting
The output of diskusg is normally used by acctdisk to create disk total
accounting records. In addition, diskusg is normally called by dodisk.
The syntax of the diskusg command is:

diskusg [ options) [ device special files]

The following options can be used with diskusg:
-s Enables diskusg to generate reports from existing diskusg output
and streamline the output by combining information about a single
user found in multiple reports. This option is used to merge data
from separate files , each containing the output from using diskusg
on different devices.
-v Useful for finding users who are trying to avoid disk space
accounting charges. When this option is specified, diskusg writes
14
records to stderr ( standard error output ) showing the device file
name, inode number, and user ID of files that have no apparent
owner.
- i fnmlist Causes diskusg to ignore the data on those file systems whose file
system name is in fnmlist . fnmlist is a list of file systems separated
by commas or enclosed within quotes .
-p p_file This is the same as the -p option of acctdusg .
- u u_file This option produces identical output as the -v option, but -v
writes its output to stderr, whereas this option writes its output
to the file u_file .

Comparing acctdusg and diskusg

Differences between diskusg and acctdusg are best illustrated by comparing
their results. In the two examples that follow, disk usage information was
generated for all users whose files reside on the disk whose device file is
/dev/dsk/OsO .
In the example using acctdusg, the command was executed from the
mount-point of the file system.

System Accounting 1 4- 1 9
 $ f ind I -hidden -print I acctdusg
00350 fred 11
0035 1 bill 30
00352 mike 17
00353 Sarah 13
00354 molly 18
00000 root 3
00004 adm 36
0000 1 bin 2434
When using diskusg, the device file of the disk is specified.
$ diskusg /dev/dsk/OsO
0 root 106 16
1 bin 778
14 4 adm 96
350 fred 14
35 1 bill 32
352 mike 20
353 S arah 16
354 molly 22
355 j ulie 2
50 1 guest 2

From these examples, you can see:

• Acctdusg places leading zeros on user IDs; diskusg does not .
• Acctdusg counts files only under each user's $HOME directory. Files owned
by users outside their home directory hierarchy ( for example, files in the
/tmp directory ) are counted as files without owner.
• Two extra users-j ul ie and guest-show up in the diskusg output. This
occurs because the directories of these two users are empty; therefore, no
disk usage totals are generated by acctdusg. However, diskusg looks at
inodes and sees that j ulie and guest are actually using two blocks for the
directories themselves.
• If two or more users have links to a particular file, acctdusg divides disk
space usage for the file between each user. For example, if three users have

14-20 System Accounting

 links to a 300-block file called report . dat , each user is charged for 100
blocks of this file.

Creating Total Accounting Records

Both acctdusg and diskusg produce intermediate information that can be
redirected into commands that produce total accounting records.
Two commands are used to create total accounting records - acctdisk , and
dodisk.
acct disk Acctdisk uses standard input from acctdusg and diskusg and
produces disk total accounting records that can then be input
to prtacct or runacct to produce summary reports. Acctdisk
has no options and is typically invoked by dodisk.
dodisk Dodisk produces total accounting records by using the 14
diskusg, acctdusg, and acctdisk commands. Dodisk is
normally invoked by cron to create disk total accounting
records for daily usage by System Accounting. The syntax for
dodisk is:
dodisk [ -o ] [files . . . ]
In the default case, dodisk creates disk total accounting
records on the special files whose names are stored in
/ etc/checkl ist ; the special file names are supplied as input
to diskusg, which pipes its output to acctdisk, which in turn
creates total accounting records.
If the -o option is used, dodisk creates total accounting
records more slowly by using acctdusg instead of diskusg.
If files are used, disk account'ng is limited to these file systems
only. When the -o option is sed, files should be mount points
of mounted file systems; if o 'tted, files should be the device
file names of the disk.
Dodisk is one of several shell scripts described on acctsh(lM)
in the HP- UX Reference .

System Accounting 14-2 1

 Other Disk-related Commands
Besides the accounting commands described in this section, you can get
valuable information using two other HP-UX commands:
emph l du l ( l ) Summarizes disk usage by giving the number of 5 12-byte
blocks contained in files and directories.
emph l bdfl ( l ) Reports the number of free disk blocks on a specified file
system.

A Security Note
It is possible for malicious users to defeat disk-space accounting by assigning
their files to other users with chown(2) or chown(l), which by default , all
users can execute. If necessary, restrict access to these commands for some or
14 all users with the s etpri vgrp ( 1M) command. To restrict chown use to the
superuser, add the following line to / et c/re :
setprivgrp -n CHOWN
To permit one or more groups of users to use chown, add a line for each group
to I etc/re, similar to the following:

setprivgrp group_name CHOWN

Bear in mind, however, that restricting access to chown might cause problems
when running various commands or applications .

Accounting for Disk-space Usage in a Cluster (Series 300/400/700 only)

All cluster nodes in an HP-UX cluster typically share the section of /etc/re
that sets privileged groups . This arrangement enables all nodes to behave
identically toward privileged groups. However, if you wish, you can m ake
I etc/re a context-dependent file, with different contents for each cluster node.
See Managing Clusters of HP 9000 Computers: Sharing the HP- UX File System
for more information.

1 4-22 System Accounting

Connect-session Accounting I
Connect-session accounting is used to record � ata about when a user logs
t
into a system. This is the form of accounting u sed by time-share systems and
electronic bulletin boards to charge subscribe s only for the time spent using
system resources, that is, the user's "connect ime."
In HP-UX, the programs login and init record connect sessions by writing
records into I etc/wtmp. The I etc/wtmp file is key to connect-session
accounting, as you can see in Figure 14-5. System accounting commands can
display or fix this file, and can produce total accounting records for this file.

14

login acctcon2

I nit Total
prctmp
Accounting
Records
Report

l.G2Q0172_004

Figure 1 4-5. Connect-session Accounting

The Key Connect-session Accounting File-/etc/wtmp

The / etc/wtmp file is one of three-/etc/wtmp , / etc/utmp , and / et c/btmp­
that hold data for such commands as last ( l ) , who( l), write(l), and login( l ) ;
they are defined by the utmp . h header file. Refer to utmp(4) in the HP- UX
Reference for detailed specifications.

System Accounting 1 4·23

 Connect-session accounting commands allo w you to write records to I etc/wtmp
and to display and manipulate I etc/wtmp data, which cannot be read directly.
You can also create total accounting records from / etc/wtmp .
Records in I etc/wtmp contain the following information:
• the terminal name on which the connect session occurred
• the login name of the user
• the current time/date at login or logout
• other status information

Connect-session Accounting Commands

Six commands are used to account for connect sessions:
• acctwtmp writes records to I etc/wtmp .
-

14 • fwtmp displays the information contained in I etc/wtmp .

-

• wtmpf ix fixes connect-session records that span date changes (refer to

-

date ( l ) in HP- UX Reference. ) . Also validates login names in connect

session records.
• acctconl summari z es / etc/wtmp in ASCII readable format , producing one
-

line per connect session.

• acct con2 takes input of the format produced by acctconl and produces
-

total accounting records as output .

• prctmp displays the session record file, called by default ,
-

/usr/adm/ acct/nite/ ctmp .

Writing Records to wtmp - acctwtmp

The command acctwtmp allows you to write records to I etc/wtmp for any
reason. Acctwtmp is typically invoked by startup and shutacct to record
when system accounting is turned on and off, respectively.
The syntax of the command is:

acctwtmp " reason"

where reason is a very brief comment string in which to describe why you are
writing the record to / etc/wtmp . Note that acctwtmp does not directly write
records to I etc/wtmp : it writes a record containing the terminal name, current
time, and reason string to standard output . To actually write the record to

1 4-24 System Accounting

/ etc/wtmp , you must append the output from acctwtmp to the / et c/wtmp file
as follows:
acctwtmp 11 reason11 > > / etc/wtmp
The reason string may be any combination of letters, numbers, spaces, and the
dollar sign ( $ ) , but may not exceed 1 1 characters in length. ( reason must be
enclosed in double quotes as shown. )
For example, you might want to add a terse remark about why you needed to
reboot, such as "out of swap" or "reconfigure" .
Refer to acct( lM) in the HP- UX Reference for specifications on acctwtmp .

Displaying Connect Session Records - fwtmp

To display the contents of /etc/wtmp , you can use the command fwtmp. When
14
no options are used, fwtmp uses standard input records of the format contained
in I etc/wtmp ; it writes to standard output the ASCII readable equivalent of
the input records.
The output of fwtmp can be edited, using an HP-UX editor such as vi, and
then rewritten to / etc/wtmp using special fwtmp options described below.
The syntax of fwtmp is:

fwtmp [ - i c ]
Ifno option is specified for the fwtmp command, input is in binary format and
can be converted to ASCII readable format . Output results differ, depending
on whether you use - i , - c , or - ic :

Option Description
- ic Input is in ASCII readable form and is to be converted to binary form. This
is essentially the opposite of using fwtmp without any options .
-i Both input and output are in ASCII readable format . This is the same as

performing an ASCII to ASCII copy.

- c Both input and output are in binary format-a binary to binary copy.

System Accounting 1 4-25

 The following example shows the output produced by fwtmp :
$ £vtmp < /etc/vtmp
syst em boot 0 2 0000 0000 479472540 Mar 12 03 : 49 : 00 1985
root co console 0 7 0000 0000 479475173 Mar 12 04 : 32 : 53 1985
acctg on 0 9 0000 0000 479493135 Mar 12 09 : 32 : 15 1985
mike a1 ttya1 352 7 0000 0000 479493590 Mar 1 2 09 : 40 : 00 1985
mike a1 ttya1 352 8 0011 0000 479496000 Mar 12 10 : 20 : 00 1985
sarah 07 tty07 353 7 0000 0000 479518335 Mar 12 16 : 32 : 1 5 1985
bill 10 tty10 351 7 0000 0000 479521475 Mar 12 17 : 24 : 35 1985
Sarah 07 tty07 353 8 0011 0000 479522478 Mar 12 17 : 41 : 18 1985
bill 10 tty10 351 8 0011 0000 479526487 Mar 1 2 18 : 48 : 07 1985
co console 0 8 0011 0000 479526488 Mar 12 18 : 48 : 08 1985
acctg 0££ 0 9 0000 0000 479526493 Mar 12 18 : 48 : 13 1985
system boot 0 2 0000 0000 479389800 Mar 12 05 : 00 : 00 1985

The output contains the following information:

• The login name of the user who logged in or out .
14
• /et c / initt ab ID (this is usually the number of the line on which the
connect session took place ) .
• Name of the device on which the connect session occurred .
• Process ID of the user who logged in or out .
• Entry type. This field contains information on the type of record-for
example, it shows whether the record is a login record (entry type=7 ) , logout
record (entry type=8), or if the record was written by acctwtmp (entry
type=9 ) . Refer to utmp( 4) in HP- UX Reference for more details on this field.
• Exit status for connect session. Refer to login(l) and utmp(4) in the HP- UX
Reference for details .
• Number of seconds elapsed from January 1 , 1970 to time of entry.
• Date and time of entry.

Fixing wtmp Errors - wtmpfix

Every entry in /etc/wtmp is categorized with a one-digit code in the entry
type.
For example, when a user logs into HP-UX, the login program stores the
value seven (7) in the entry type field of the connect session record. When the
same user logs out , an entry type of eight (8) is recorded. You can see this by

1 4-26 System Accounting

examining the sample output created by fwtmp in the previous section. Note
that in the example, login records precede their corresponding logout records in
chronological order.
If the system date is changed using date(l) while users are logged in,
time-stamped ordering can become inconsistent . An inconsistent state can
result in logout records to precede login records and can confuse the commands
that create connect-session total accounting records .
The command wtmpf ix fixes corrupted / etc/wtmp files by taking /etc/wtmp
binary records as input , modifying the time/date stamps , and writing the
corrected output to a temporary file in binary wtmp format . The syntax for
wtmpfix is:

wtmpf ix [ files]
See the System Administration Tasks manual and fwtmp(lM) in the HP- UX 14
Reference for information on using wtmpfix.
Recording Connect-session Statistics
Total accounting records for connect-session accounting are created by the
acctcon2 command.
Prior to generating the total accounting records , however, you use acct con1 to
produce columnar output from I etc/wtmp and prctmp to provide headings for
your columnar output . Both acct con1 and acctcon2 are found in acctcon(lM)
of the HP- UX Reference . Prctmp is a shell script found in acctsh(lM).

System Accounting 1 4-27

 Tabulating Output for Connect-session Accounting -acctcon 1

Acctcon1 converts the sequence of login and logout records read from
I etc/wtmp into output in columnar AS CII format . This output can be
supplied as input to prctmp or acctcon2 , two commands that generate
connect-session reports.
The columnar output of acctcon1 resembles that of fwtmp ( shown previously ) ,
but sorts by different categories, including device address and duration of
connect time.
$ acctcon1 < /etc/vtmp
20095488 353 sarah 1665 2478 479518335 Tue Kar 12 16 : 32 : 1 5 1985
521012224 352 mike 479493590 Tue Kar 12 09 : 40 : 00 1985
520095488 351 bill 0 5012 479521475 Tue Kar 12 17 : 24 : 35 1985
52101 1712 0 root 41047 6488 479475173 Tue Kar 12 04 : 32 : 53 1985

The output contains the following information:

14
• Device address ( in decimal equivalent of major/minor device address ) at
which the connect session occurred.
• User ID for the connect session record.
• User login name.
• Number of prime connect-time seconds used during the connect session.
• Number of non-prime connect seconds .
• Number of seconds elapsed from January 1 , 1970 to connect-session starting
time.
• Date and starting time time of connect session.

1 4-28 System Accounting

In addition to its default usage, acctcon1 has four options:

Option Description
-p This option tells acct con1 not to produce one record per connect session.
Instead, acctcon1 simply echoes its input-one line per wtmp record­
showing line name, login name, and time (in both seconds and day /time
format). Using this option is similar to using fwtmp, except that this
option doesn't show status information, whereas twtmp does.
-t acctcon1 maintains a list of lines on which users are logged in. When it
reaches the end of its input , it emits a session record for each line that still
appears to be active. It normally assumes that its input is a current file, so
that it uses the current time as the ending time for each session in progress.
The -t flag causes it to use, instead, the last time found in its input , thus
assuring reasonable and repeatable numbers for non-current files .
-1 file This option causes a line usage summary report to be placed in file . This 14
report shows each line's name, number of minutes used, percentage of total
elapsed time used, number of sessions charged, number of logins, and
number of logins and logoffs. This report can be used to keep track of line
usage , identify bad lines, and find software/hardware oddities.
Note that hang-up , termination of login, and termination of the login shell
each generate logoff records; therefore , the number of logoffs is often three
to four times the number of connect sessions .
-o file Using the -o option (for example, acct con1 -o £_overall ) creates a
concise file containing a record of the accounting period, giving starting
time, ending time, number of reboots, and number of date changes.

Acct con1 options can be combined, as shown in the following example, which
uses both -1 and -t options. Also, the standard output of acct con1 is
redirected to the file ctmp :

System Accounting 1 4-29

 $ acctcon1 -t -1 line_us e < /etc/wtmp > ctmp
$ cat l ine_use
TOTAL DURATION IS 899 MINUTES
LINE MINUTES PERCENT # SESS # ON # OFF
console 856 95 1 1 1
tty07 69 8 1 1 1
ttya1 40 4 1 1 1
tty10 84 9 1 1 1
TOTALS 1049 4 4 4

Heading Columns of acctcon 1 Output-prctmp

The prctmp command simply puts headings above the output created by
acctcon1 , thus making the report more readable.

14 Since prctmp takes its input from standard input , pipe the output from
acctcon1 into prctmp as follows:
$ acctcon1 < / etc/wtmp I prctmp

Creating Connect-session Total Accounting Records -acctcon2

Connect-session total accounting records are created by the acctcon2

command, found in acctcon( lM) of the HP- UX Reference.
Acct con2 takes standard input created by acctcon1 and converts it into the
structure of tacct . h, explained in acct( lM) of the HP- UX Reference.
The total accounting records created by acct con2 are sent to standard output ,
so to store these records , you must redirect standard output to a file.
In the following command, the total accounting records generated by
/etc/wtmp is redirected into the file ctacct :
$ acctcon1 < / et c/wtmp I acct con2 > ct acct

1 4-30 System Accounting

Process Accounting
Process accounting quantifies resource use by process ( that is, by programs
executed ) . Statistics include memory usage, CPU time, number of
input / output transfers for individual processes. This information is
accumulated in a process-accounting file, called by default /usr/adm/pacct .
Note that several accounting commands do not work properly unless this name
is used.
Information contained in /usr/adm/pacct includes:
• the user ID of the process 's owner
• the name of the command that invoked the process
• the amount of time it took the process to execute
Once system accounting is installed and turned on, the termination of every
process is recorded . The kernel writes a record of the terminating process into 14
the current process accounting file, /usr/adm/pacct .
For greater detail on the contents and format of process accounting records,
refer to acct(4) in the HP- UX Reference.
Sever al accounting commands exist to display, report , and summarize process
accounting information. In addition, certain commands turn process accounting
on or off and ensure that /usr I adm/pacct does not grow out of bounds.
The following figure shows the flow of information and commands involved in
process accounting:

System Accounting 1 4-3 1

 /etc/wt mp

Total
acctprc2 Aooountlng
Records

Display

14 Report

LG200 1 72_005

Figure 1 4-6. Process Accounting

Turning On Process Accounting

Before System Accounting can generate process accounting data, process
accounting must be turned on. This is done with the startup command,
run from the system initialization shell script / etc/re . If you have updated
I etc/re for System Accounting ( as described earlier in "System Accounting
Set-up Guidelines" ) , process accounting is activated automatically and you
seldom need to use the commands described here. These commands are
described in case you ever need to manually turn on or off process accounting.
Process accounting can be turned on and off using either of two commands:
turnacct and accton. Turnacct is a multi-faceted shell script found in
acctsh(lM) and accton is found in acct(lM), both located in the HP- UX
Reference. ( Both commands are used to turn accounting off also, but we will
discuss that later. )
turnacct on Turnacct on is the command used most often to activate
accounting. Only the superuser and the adm login can execute
this command.
Turnacct on works as follows :

1 4-32 System Accounting

 1 . Checks to see if the process accounting file /usr I adm/pacct
exists, and if not , turnacct creates the file.
2. Turns on process accounting by invoking accton with the
default filename argument /usr/adm/pacct . This causes all
process accounting data to be stored in the pacct file.
acct on Accton, specified with a filename, turns on process accounting,
calls acct(2), and has the kernel write process-accounting
records to the file. Note that the specified file must exist before
executing accton; otherwise, accton fails . Only the superuser
can execute accton.
If no file name is given, process accounting is turned off; the
kernel stops writing process accounting records to the process
accounting file.
14

Turning Off Process Accounting

If you updated the /etc/shutdown shell script to include shutacct ( as
described earlier in "When System Accounting is Turned Off'' ) you seldom
use these commands, because the shutacct command automatically turns off
process accounting. The following commands are explained, should you ever
need to deactivate process accounting manually.
Process accounting can be turned off by using either of two commands:
turnacct off and accton ( with no filename argument ) . Both commands tell
the kernel to stop writing records to the current process accounting file.
turnacct off Turnacct off turns off process accounting by invoking the
accton command without the optional filename argument .
Turnacct off is executed only by the superuser and the adm
login.
accton When accton is invoked without the optional filename
argument , process accounting is turned off.
As shown in the System Data Flow Diagram, accton tells the
kernel to stop writing process accounting records by using the
system call acct .

System Accounting 1 4-33

 Managing the Size of the Process Accounting File pacct
On a multi-user system, many processes execute during a single hour, causing­
process accounting files to become quite large. System Accounting has
built-in mechanisms that ensure that the default process accounting file
/usr/adm/pacct does not grow out of bounds.
The two commands used for this purpose are: turnacct switch and ckpacct .
Note that these commands work only on the default process accounting file,
/usr/adm/pacct .

Checking the Size of pacct-ckpacct

The command ckpacct is typically invoked by cron hourly to ensure that the
current process accounting file /usr/adm/pacct does not become too large.
If disk space becomes critically short , process accounting is turned off until
14
sufficient space is available.
The syntax of ckpacct is:

ckpacct [ blocks]
If the size of pacct exceeds the blocks argument , 1 000 by default if blocks is
not specified, then turnacct switch is executed. turnacct switch renames
the current pacct file and creates a new pacct file.
If the amount of free space falls below a threshold of 500 blocks, ckpacct
automatically turns off process accounting via turnacct off . When free space
exceeds the specified threshold, process accounting is reactivated.
The kernel can also enforce a size limit on the size of pacct . This takes
precedence over the limit set by ckpacct . Refer to acctsh ( 1M) and acct(2) in
the HP- UX Reference for more details.
If /usr I adm/pacct becomes too large, a new /usr I adm/pacct file is created
via turnacct switch.
The turnacct swit ch renames the current /usr/ adm/pacct file so that it no
longer receives data, and creates a new, empty /usr I adm/pacct file.

1 4-34 System Accounting

Changing Process Accounting Files -turnacct switch

turnacct switch is used to create a new /usr I adm/pacct file when the
current /usr/adm/pacct file is too large. The action of turnacct switch can
be summarized as follows:
1 . Process accounting is temporarily turned off.
2. The current pacct file is renamed to pacct incr , where incr is a number
starting at 1 and incrementing by one for each additional pacct file that is
created via turnacct swit ch.
3. After the old pacct file is renamed to pacct incr , a new, current pacct file
is created.
4. Process accounting is restarted; the kernel starts writing records to the
newly created pacct file.
14

Displaying Process Accounting Records - acctcom

The acctcom( lM) command allows you to display records from any file
containing process accounting records . Acct com generates a columnar report
with descriptive headings for each column. Each line of the report represents
the execution statistics accumulated by a particular process during its lifetime.
Typically you use this command to display records from the /usr I adm/pacct
files (pacct , pacct 1 , pacct2 . . . ) .
Acct com is a very versatile command with many options. Its syntax follows:

acctcom [ [ options ] [file ] ] . . .

If no file is specified, acctcom uses the current /usr/adm/pacct file as input .
Input can also be taken from standard input . Some acctcom options allow you
to select only the records that you want to see; other options control the format
of the report .

System Accounting 1 4-35

 Sample Report using acctcom

The following sample report illustrates the default use of acct com.
$ acct com
ACCOUBTllG RECORDS FROM : Thu Kar 21 12 : 52 : 26 1985
COKKAl'ID START EID REAL CPU KEAi
HAKE USER TTYIAKE TIME TIME (SECS ) (SECS ) SIZE (K)
#acct on root console 1 2 : 52 : 26 12 : 52 : 26 0 . 12 0 . 10 19 . 00
ls sarah tty07 14 : 04 : 08 14 : 04 : 08 0 . 28 0 . 23 1 6 . 50
ckpacct adm '!' 14 : 30 : 00 14 : 30 : 05 5 . 13 1 . 45 24 . 00
pvd bill tty10 15 : 09 : 07 15 : 09 : 07 0 . 48 0 . 22 22 . 50
:find sarah tty07 18 : 51 : 37 18 : 51 : 39 2 . 73 0 . 15 26 . 50
t abs root console 19 : 10 : 18 19 : 10 : 18 0 . 92 0 . 13 23 . 50
stty root console 19 : 10 : 19 19 : 10 : 19 0 . 88 0 . 08 26 . 00
mail bill t ty10 19 : 10 : 21 19 : 10 : 22 1 . 78 0 . 23 28 . 50
nevs root console 19 : 10 : 23 19 : 10 : 23 0 . 73 0 . 12 23 . 00
acct com adm ttyaO 19 : 53 : 16 19 : 53 : 38 22 . 58 2 . 55 28 . 50
#dat e root console 19 : 58 : 26 12 : 58 : 26 0 . 12 0 . 10 19 . 00
14
Definitions o f Information Produced by acctcom

Default output of acctcom ' includes the following:

1 4-36 System Accounting

 Column Header Definition
COMMAND NAME The name of the command or program that spawned the
process .
Commands requiring superuser privileges are listed with a
preceding # (see #dat e in sample output above) . This feature
ensures that actions that can only be accomplished as superuser
are visible via system accounting records .
USER The login name of the user who created the process .
TTYNAME The terminal from which the process was executed. If the
process was not executed from a known terminal (for example,
if it was executed via cron ) , a question mark( ?) appears .
START TIME When the process began executing, specified in hh : mm : ss
format .
EID TIME When the process finished executing. 14
REAL ( SECS ) How many seconds elapsed from START TIME to END TIME.
, CPU ( SECS) How much CPU time a process used while executing.
MEAN SIZE (K) How much memory (estimated in kilobytes) a process used
while executing.

The MEAN SIZE estimate is determined from the current process's memory
usage at each system clock interrupt , and is therefore subject to statistical
sampling errors. Only the memory resident pages of a process are counted; no
pages in the swap space are counted. Shared code and data are divided among
the processes using them. MEAN SIZE is divided by the number of processes
sharing the code or data.

Additional Capabilities of acctcom

Listed below are columns not displayed in a standard acct com report , but can
be displayed by using acct com options. The acctcom options used to generate
these columns are described in the next section.

System Accounting 1 4-37

 Column Header Definition
F For a process created by fork that does not call exec, this
column takes the value 1 ; otherwise , this column shows a 0 .
STAT This column displays the system exit status if the process
terminated by calling exit . When a process terminates
normally, this field shows a 0 . If a command terminates
abnormally, the value returned is the signal number that caused
the process to terminate. If a core file image is produced by the
signal, the value is a signal number plus 0200 . For specific
details of signal values, see signal(5) in the HP- UX Reference .
HOG FACTOR The hog factor is computed as the CPU time divided b y REAL
time; it provides a relative measure of the available CPU time
used by the process during its execution. For example, a hog
factor of less than 0.50 indicates that the process spent less
than half of its time using the CPU. A hog factor of 0. 75
14
indicates that a process spent 75% of its time using the CPU.
KCORE MIN This calculation provides a combined measurement of the
amount of memory used (in kilobytes) and the length of time it
was used (in minutes) . It is computed as follows:
KCORE MIN = CPU ( SECS) * MEAN SIZE (K) / 60
CPU SYS This is the portion of total CPU time that was spent executing
operating system code, such as system calls (for example,
writing to disk) .
USER ( SECS ) This is the remaining portion of CPU time. User CPU time is
the amount of time actually spent executing a process's code
(rather than system code) .

1 4-38 System Accounting

 Colunm Header Definition
CPU FACTOR Whenever you execute a command, the CPU spends part of its
time actually executing the command's code (user CPU time)
and spends the rest of its time performing system functions,
such as writing to the disk or terminal (system CPU time) .
That is, total CPU time is comprised of both CPU SYS and USER
CPU time:
CPU ( SECS ) = CPU SYS + USER ( SECS)
The CPU factor shows the ratio of user CPU time to total CPU
time :

CPU FACTOR = USER ( SECS ) / ( CPU SYS + USER ( SECS ) )

For example , if a command has a CPU factor of 0.35, that
means that the CPU spent 35% of its time executing user code
14
and 65% performing system functions.
CHARS TRNSFD The number of characters (bytes) read and/or written by the
command is displayed in this column.
BLOCKS R/W This column shows the number of file system blocks read and/ or
written as a result of executing this command. This number is
not directly related to CHARS TRNSFD and may vary each time
the command is executed, because BLOCKS R/W is affected by
directory searches made before opening files, other processes
accessing the same files, and general file system activity.

Formating Options of acctcom

The options to acctcom( lM) enable additional statistics to be reported, some of

which are discussed in the previous section:

System Accounting 1 4-39

 Option Description
-a Display at the end of the report statistical averages. Total number of
commands processed shown as ( cmds=xxx)
• average real time per process (Real=x . xx)
• average CPU time per process ( CPU=x . xx)
• average USER CPU time per process {USER=x . xx)
• average SYS CPU time per process {SYS=x . xx)
• average characters transferred ( CHARS=x . xx)
• average blocks transferred {BLK=x . xx)
• average CPU factor {USR/TOT=x . xx)
• average HOG factor (HOG=x . xx)
-b Display the process records in reverse order: most recently executed
commands shown first .
-f Print the fork/exec flag (F column ) and process exit status (STAT column ) .
14
-h Display the optional HOG FACTOR column, instead of the standard mean
memory size MEAN SIZE (K) .
-i Replace the standard MEAi SIZE (K) column with the optional 1/0
counts-CHARS TRNSFD and BLOCKS R/W .
-k Replace the standard MEAN SIZE (K) column with KCORE MIN.
-m Show the default column MEAi SIZE (K) when MEAi SIZE (K) is superceded
by another option, such as KCORE MIN .
-r Include the optional CPU FACTOR column in the report .
-t Show separate CPU times for system {CPU SYS) and user (USER (SECS ) ) .
-v Suppress printing column headings .
-q This option produces only the averages as displayed by the -a option,
without the individual process accounting records.
-o ofile Copy the input process accounting records to ofile .

1 4·40 System Accounting

Record Selection Options of acctcom

The options described here allow you to select the records that are included
in the report produced by acctcom. See your System Administration Tasks
manual for usage examples.

Option Description
-1 lin e Display only the processes executed from the user terminal /dev/ /ine .
- u user Show only the processes belonging to user.
-g group Show only processes belonging to group , which may be specified by
either group name or group ID .
-s time Select processes existing at or after tim e . Time is specified in 24-hour
format- hr[: min [:sec]] .
-e time Select processes existing at or before tim e . Time is specified in 24-hour
format hr[: min [:sec]] . 14

-S time Select processes starting at or after time where time is in 24-hour format.
-E time Display only processes terminated at or before tim e , where time is in
24-hour format hr: [min [: sec]] . Note both the -S and -E options with the
same time argument cause acct com to display only processes existing at
the specified time .
-n pattern Show only the commands matching pattern . pattern can be a regular
expression as described in ed( l ) , except that + means one or more
occurrences.
-H factor Display only processes whose hog factor exceeds factor.
-0 time Show only processes whose system CPU time exceeds time , specified in
seconds.
-c sec Show only processes whose total CPU time ( SYS + USER ) exceeds sec
seconds.
-I chars Display only processes transferring more characters than the limit given
by chars .

System Accounting 1 4·4 1

 Process-accounting Report of Commands - acctcms
The acctcms( lM) command takes /usr/ adm/pacct as input , and produces
summary accounting information by command instead of process . The action
of acctcms( lM) can be summarized as follows:
1 . acctcms reads the data in process-accounting records to accumulate
execution statistics for each unique command name. This information is
stored in internal summary format-one record per command name.
2. Command summary records created in step 1 are sorted according to the
acct cms options specified.
3 . Command summary records are written to standard output in the internal
summary format mentioned in step 1 . This format is not readable.
The syntax of the acctcms command is:
14

acctcms [ options ] f iles

where files is a list of the input process accounting files for which the command
summary report is to be generated. The options are discussed in subsequent
sections.

Producing a Readable Report - the -a option

By default , the output of acct ems is in internal summary record format ; if you
display it to your terminal, all you see is gibberish. To get a AS CII, readable
report , use the -a option.
The -a option causes acctcms to produce a report with descriptive column
headings . Total and average ( mean ) execution statistics for each command are
displayed-one line per command-along with total and average statistics over
all commands in the report .

Sample Report by acctcms

The AS CII reports produced by acctcms contain more than 80 characters per
line. When displayed at an 80-column terminal, the lines wrap around on the
screen , and if printed on an 80-column printer, some of the rightmost columns
of the report are lost . Therefore, compensate as follows:

1 4-42 System Accounting

• Use a printer with compressed print capabilities, so that the entire report fits
on standard computer paper; or
• Use a wide-capability printer to display all the information, such as one that
prints 132 columns.
The following example shows a command summary report for the current
process accounting file ( if no file is specified, the /usr/adm/pacct file is used ) .
Note also that total execution statistics for all commands are grouped under
the command name TOTALS.
$ acct cms -a proc . figs
TOTAL COllllAID SUMKAB.Y
COllllAID IUMBER TOTAL TOTAL TOTAL llEAI llEAI HOG CHARS BLOCKS
IA!IE CllDS KCOREllUI CPU-1111 REAL-llII SIZE-K CPU-1111 FACTOR TRISFD READ

TOTALS 61 17 . 63 0 . 38 164 . 49 46 . 25 0 . 01 0 . 00 104553 1027

14
acct ems 17 12 . 13 0 . 16 0 . 35 76 . 72 0 . 01 0 . 45 49192 306
sh 8 2 . 43 0 . 09 152 . 86 26 . 79 0 . 01 0 . 00 9043 163
more 3 0 . 73 0 . 02 10 . 50 31 . 00 0 . 01 0 . 00 21618 83
11 6 0 . 61 0 . 04 0 . 11 16 . 50 0 . 01 0 . 33 571 5 95
acct com 4 0 . 58 0 . 02 0 . 07 28 . 50 0 . 01 0 . 30 15319 42
cat 4 0 . 19 0 . 01 0 . 35 22 . 97 0 . 00 0 . 02 31 1 2 52
rm 2 0 . 11 0 . 00 0 . 02 22 . 22 0 . 00 0 . 29 0 29
chmod 2 0 . 10 0 . 00 0 . 01 22 . 00 0 . 00 0 . 35 0 15
acct on 2 0 . 08 0 . 00 0 . 02 19 . 00 0 . 00 0 . 29 0 22
sed 2 0 . 08 0 . 01 0 . 04 14 . 50 0 . 00 0 . 13 73 38
echo 2 0 . 05 0 . 00 0 . 02 20 . 00 0 . 00 0 . 16 22 21

Header Descriptions

Descriptions of the columnar data produced by acct cms are shown in the
following table.

System Accounting 1 4-43

 Column Header Description
COMMAND NAME The name of the command for which execution statistics are
summarized. Unfortunately, all shell procedures are lumped
together under the name sh, because only obj ect modules are
reported by the process accounting system.
NUMBER CMOS The total number of times that the command was invoked.
TOTAL KCOREMIN The total amount of kcore minutes accumulated for the
command. (Refer to the section "Additional acctcom
Capabilities" earlier in this chapter for more information .)
TOTAL CPU-MIN The total CPU time allocated to the named command.
TOTAL REAL-MIN Total allocated real-time minutes are displayed in this column.
MEAN SIZE-K The average amount of memory (in kilobytes) consumed by the
command.
14 MEAN CPU-MIN The average CPU time allocated per command invocation is
shown here; the following equation shows how it is computed:
MEAN CPU-MIN = TOTAL CPU-MIN / NUMBER CMOS
HOG FACTOR The average hog factor over all invocations of the command. H
is computed as :
HOG FACTOR = TOTAL CPU-MIN / TOTAL REAL-MIN
CHARS TRNSFD The total number of characters transferred by the command.
Note that this number may sometimes be negative .
BLOCKS READ A total count of the physical blocks read and written by the
given command. (Refer to the section "Displaying Process
Accounting Records-acctcom" in\ this chapter for details on
the significance of this total.)

Note When only the - a option is specified, the report is sorted in

descending order on the TOTAL KCOREMIN column: commands
using more TOTAL KCOREMIN are shown before those using fewer
TOTAL KCOREMIN. This gives a relative measure of memory used
over time by the various commands: commands listed early in
the report use more memory resources than commands listed
later.

1 4·44 System Accounting

Additional acctcms Options

Several options besides -a can be used to control the format of reports

generated by acctcms . Some options specify on which field to sort the report ;
other options control the printing of prime / non-prime time usage.

Option Description
-c Sort the commands in descending order on TOTAL CPU-MIN , rather than
by default TOTAL KCOREMIN. This report can be used to determine which
commands are using most of the computer's CPU time.
-n Cause the report to be sorted in descending order on the column named
NUMBER CMDS . Commands toward the start of this report are the ones
used most frequently; commands toward the end are used least often.
-j Combine all commands invoked only once on one line of the report . This
line is denoted by having "***other" in the COMMAND NAME column. This
14
option is useful for shortening a report that has many one-invocation
commands.
-o Used only with the -a option , -o causes the report to be generated only
for commands that were executed during non-prime time ( as specified in
the holidays file ) . You can use this option to get a non-prime time
command summary report .
-p Also used only with the -a option , -p elicits a report generated only for
commands that were executed during prime time ( as specified in
holidays ) . This option is used to get a prime time command summary
report .

System Accounting 1 4-45

 Option Description
-apo When the options -o and -p are used together with -a, a combination
prime and non-prime time report is produced. The output of this report
is identical to that produced by -a alone, except that the NUMBER CMDS,
TOTAL CPU-MIN , and TOTAL REAL-MIN columns are divided into two
columns-one for prime time totals , the other for non-prime time.
( Prime time columns have a (P) header, while non-prime time columns
are headed by (NP ) .)
-s [files] Specifies that any named input files following the -s on the command
line are already in internal summary format . This option is useful for
merging previous acct cms reports with current reports . The following
example uses -s to create a command summary report from previous
process accounting files (pacct? ) and the current process accounting file
(pacct ) . The final ASCII report is stored in the file as cii_summary.
14
$ acctcms pacct? > old_summary
$ acct cms pacct > new_summary
$ acctcms -as old_summary new_summary > as cii_summary

Creating Total Process-accounting Records

Two commands- acctprc1 and acctprc2-are used to create total accounting
records from the process accounting files . The output from acctprc 1 is
supplied as input to acctprc2 , which produces the total accounting records .
These commands are invoked by runacct to produce daily accounting
information.
acctprc 1 Acctprc1 produces readable process-accounting information,
mainly for input into acctprc2 . Acctprc 1 reads process
accounting records from standard input , adds login names
corresponding to the user ID of each record, and for each
process writes an AS CII line showing:
• the user ID of the creator of the process
• the user's login name
• prime CPU time in ticks ( a "tick" is one fiftieth of a second )
• non-prime CPU time, also in ticks
• mean memory size ( in pages-typically 4 Kbytes )

1 4-46 System Accounting

 The format of acctprc1 is:

acctprc 1 [ ctmp ]
where ctmp contains a list of login sessions of the form created
by acctcon1 , sorted by user ID and login name.

Note The number of sessions should be 1000 or less. Running more

than 1000 sessions causes the accounting system to "hang"
( that is, suspend indefinitely ) . If that happens, sessions must
be killed using kill command and restarted.

To use acctprc 1 , input must be redirected from a process

accounting file, such as /usr/adm./pacct . The following
example creates a file, as cii_ptacct , containing ASCII
14
process accounting information useful for creating process total
accounting records.
$ acctprc 1 <pacct > ascii_ptacct
Normally, acctprc 1 gets login names from the password file
/ etc/passwd, which is sufficient on systems where each user
has a unique user ID . However, on systems where different
users share the same user ID , the ctmp file should be specified;
it helps acctprc1 distinguish different login names that share
the same user ID .
acctprc2 Acctprc2 takes input produced by acctprc 1 , summarizes the
records by user ID and name, and writes the sorted summaries
to standard output as total accounting records . The following
example creates total accounting records for all processes in
the current process accounting file pacct ; the total accounting
records are stored in the file ptacct .
$ acctprc1 <pacct 1 acctprc2 >ptacct

System Accounting 1 4-47

 User Fees - chargefee
Although you can use system accounting reports as input to other fiscal
accounting packages, HP-UX system accounting also provides some capability
for charging fees . Chargefee, a shell script on acctsh ( lM ) of the HP- UX
Reference, allows you to charge fees to specific users by allowing you to tally
generic units per specific login n:ame.
For example, you might apportion fees among several departments for use of
computer resources or charge fees to users for fixing damaged files.
The syntax of this command is:

chargefee login_name number

where login_name represents the user to whom number units are being charged
14 and number represents how many units to charge to the user.

Note number can be any whole number from -32K to 32K; when
charging fees, remember that a user's total fees must be within
this range.

( login )--•I /etc/wt.mp ---<chargefe),___

e �_... /usr/ad rn/fea

l..G200172Jl08

Figure 1 4-7. Charging Fees

Chargefee accumulates fee charge records in the file /usr/adm./fee. These

records are then merged with other accounting records via runacct .
See the System Administration Tasks manual for detailed implementation.

1 4-48 System Accounting

Accounting Summaries and Reports
Five commands and shell scripts summarize and report data for a variety of
accounting purposes:
prtacct Display total accounting records.
acctmerg Combine total accounting files.
runacct Generate daily summary files and reports.
prdaily Create and display daily summary files and reports generated
from runacct output .
monacct Create monthly summary files and reports generated from
runacct output .
In terms of importance, you will probably find that runacct is your main daily
accounting tool. For explanatory purposes, however, we first discuss prtacct
and acctmerg. 14

Displaying Total Accounting Records - prtacct

The prtacct command allows you to display the contents of a process
accounting file by taking as input total accounting records and displaying them
in AS CII-readable format . Its syntax is

prtacct file 11 heading11

where file is the name of the total accounting file being displayed and
"heading" is a descriptive header produced by prtacct . Since "heading" can
be a phrase, enclose it in double quotes.

l..G200 1 72Jl07

Figure 1 4·8. Displaying Total Accounting Records

System Accounting 1 4·49

 Sample Output by prtacct

Although prtacct is capable of providing merged summary reports, this first

example displays disk total accounting records only. First, the total accounting
records are created by executing disk space accounting commands; then, they
are displayed using prtacct . When studying this report , note the following:
• There are many similarities between this and the sample report produced
by diskusg ( refer to the section "Disk Space Usage Accounting" in this
c,hapter ) .
• Since the total accounting records were created only from disk space usage
accounting commands, only columns relating to disk space usage have
non-zero values .
$ £or £ile_system in ' cat /etc/checklist '
> do
14 > diskusg $£ile_system >dtmp . ' basename $£ile_syst em '
> done
$ diskusg -s dtmp . • l sort +On +1 l acctdisk >disktacct
$ prt acct diskt acct "DISK TOTAL ACCOUITIIG RECORDS"

Kar 26 17 : 01 1986 DISK TOTAL ACCOUITIIG RECORDS Page 1

LOGII CPU (KIIS ) KCORE-KIIS COllECT (KIIS) DISK I OF I OF I DISK FEE

UID I.A.KE PRIME IPRIKE PRIME IPRIME PRIME IPRIKE BLOCKS PROCS SESS SAKPLES
0 TOTAL 0 0 0 0 0 0 1 1698 0 0 10 0
0 root 0 0 0 0 0 0 10616 0 0 1 0
1 bin 0 0 0 0 0 0 778 0 0 1 0
4 adm 0 0 0 0 0 0 96 0 0 1 0
360 £red 0 0 0 0 0 0 14 0 0 1 0
361 bill 0 0 0 0 0 0 32 0 0 1 0
362 mike 0 0 0 0 0 0 20 0 0 1 0
363 sarah 0 0 0 0 0 0 16 0 0 1 0
364 molly 0 0 0 0 0 0 22 0 0 1 0
366 julie 0 0 0 0 0 0 2 0 0 1 0
601 guest 0 0 0 0 0 0 2 0 0 1 0

Report Format of prtacct

Prtacct produces a columnar report with one line per total accounting record,
with the following column headings :

1 4-50 System Accounting

 Column Header Description
UID User ID of the user for whom the total accounting record was
created.
LOGIN NAME The user 's login name .
CPU (MINS ) Total amount of CPU time ( in minutes ) consumed by the user,
during prime and non-prime time.
KCORE-MINS This represents a cumulative measure of memory and CPU time
consumed by a user during prime and non-prime time. ( See also
"Additional acctcom Capabilities" earlier in this chapter for a more
information ) .
CONNECT The amount of time a user spent logged into the system, cited by
(MINS) PRIME and NPRIME time.
DISK BLOCKS The total number of disk blocks allocated to the user .
# OF PROCS The total number of processes spawned by the user. 14

# OF SESS How many times the user logged in.

# DISK How many times disk accounting was run to obtain the average
SAMPLES number of disk blocks listed in the DISK BLOCKS column.
FEE The number of fee units charged via chargef e e .

Origin o f Columnar Data by prtacct

Process-accounting commands are used to tabulate data in the CPU (MINS) ,

KCORE-MINS , # OF PROCS columns.
Connect-session accounting commands are used to tabulate data in the
CONNECT (MINS) and # OF SESS columns.
Disk-space accounting commands are used to tabulate data in the DISK BLOCKS
and # DISK SAMPLES columns.

Merging Total Accounting Files - acctmerg

Acctmerg combines the contents of separate disk, process , and connect session
total accounting files into a single comprehensive total accounting file. Data is
consolidated by user name and ID .
Acctmerg is typically executed by runacct .

System Accounting 1 4-5 1

 Acctmerg reads standard input and up to nine additional files, all in total
accounting record format . Thus, acctmerg accepts input from:
• acctdisk for disk-usage accounting
• acct con2 for connect-session accounting
• acctprc2 for process accounting.
The syntax for acctmerg is:

acctmerg [ options] [file ]

where options control the report format and how records are merged and file
can be one to nine files ( plus standard input ) being merged into a single total
accounting file, written to standard output .

Command Options tor acctmerg

14
The following options can be used to control the report format and how the
total accounting records are combined. ( For more details, see the acctmerg(lM)
manual page in HP- UX Reference.)
-a Produce output in ASCII, much like that of prtacct
reportage, but without report headings.
-i Accept input to acctmerg in AS CII format.
-p Echo input records without merging or processing, and display
output as with the -a option.
-t Produce a single total accounting record of all input . For the
AS CII version of this record, use the both -t and -a options,
each specified with its own hyphen.
$ acctmerg -t -a <tot _ acct_recs
- u Merge records by user ID only, rather than user ID and name.
-v Produce verbose ASCII output, with more precise notation
( showing full logarithmic and exponential factors ) for floating
point numbers .

1 4-52 System Accounting

Creating Daily Accounting Information - runacct
Runacct is a comprehensive daily accounting procedure, a shell script found on
the acctsh(lM) manual page of the HP- UX Reference. Typically, runacct is
invoked daily by cron during non-prime hours when users are logged off. This
ensures for maximum accuracy.
As shown in the following diagram, runacct processes and merges disk,
connect session, process , and fee accounting data, and produces summary files
and reports. It prepares cumulative summary files for use by prdaily and for
billing purposes.

14

cron

Daily
Interim
prdaily Summary
Reports ReporlS

Up to Nlne Monthly
Total Accounting monacct Summary
Records Reports

l.G200172_008

Figure 1 4-9. How runacct Generates Summary Reports

This section discusses the following aspects of runacct :

• files processed by runacct
• the states through which runacct progresses while executing
• recovery from runacct failure
• restarting runacct
• reports produced by runacct

System Accounting 1 4-53

 Files Processed b y runacct

The following files, processed by runacct , are of particular interest to the

reader. ( File names are given relative to the directory /usr/ adm/acct . )
ni te/lineuse Usage statistics for each terminal line on the system. This
report is useful for detecting bad lines. If the ratio of logoffs
to logins on a particular line exceeds 3 to 1 , that line is likely
failing.
ni te/ daytacct Total accounting records from the previous day.
sum/tacct Accumulated total accounting records for each day's total
accounting records (nite/daytacct ) . sum/tacct can be used
for billing purposes. It is restarted each month or fiscal period
by the monacct shell script.
14 sum/dayems Produced by acctems , sum/dayems contains the daily
process-accounting command summary. The ASCII version of
this file is in ni te/ dayems .
sum/ems Accumulates each day's command summaries ( sum/ dayems ) .
A new sum/ ems file is created each month by monacct . The
ASCII version of this file is in ni te/ ems .
sum/loginlog Maintains a record of the last time each user logged in.
sum/rprt mm dd Main daily accounting report created by runacct . The name
for this report is created automatically by the system with mm
being the month and dd the day of the report . This report can
be printed via prdaily.

1 4-54 System Accounting

Safeguards of runacct

Runacct does not damage files. A series of protective mechanisms recognize

errors, provide intelligent diagnostics, and terminate processing so that
runacct can be restarted with minimal intervention. To ensure these goals, the
following actions are performed by runacct :
• Runacct 's progress is recorded by writing descriptive messages to the
nite/act ive file.
• All diagnostic output generated while runacct executes is redirected to the
file nite/fd2log.
• If the files lock and lock 1 exist when runacct is invoked, an error message
is displayed and execution terminates .
• The lastdat e file contains the month and day when runacct was last run
and is used to prevent runacct from being executed more than once per day. 14

• If runacct detects an error, a message is written to /dev/console, mail is

sent to root and adm, locks are removed, diagnostics files are saved, and
execution is terminated.

States of runacct

Runacct accomplishes its tasks by proceeding through various states, each

associated with specific tasks. These states enable runacct to be restarted
easily.
For example, in one state, total accounting records for connect sessions are
created; in another, disk, connect session, process, and fee total accounting
records are merged to create one total accounting file.
As runacct executes, it records its progress by writing the name of the most
recently completed state into the /usr/adm/stat ef ile file. After processing
for each state ends, runacct examines statefile to determine which state to
enter next . When runacct reaches the final state (CLEANUP) , the lock and
lock1 files are removed, and execution terminates .

System Accounting 1 4·55

 State Action
SETUP Execute the command turnacct switch. The process accounting files,
pacct ? , are moved to Spacct? . mmdd . The /etc/wtmp file is moved to
nit e/wtmp . mmdd with the current time added on the end.
WTMPFIX Check nit e/wtmp . mmdd for correctness by wtmpf ix . Since some some
date changes cause acct con1 to fail, wtmpf ix adj usts time stamps in
nit e/wtmp . mmdd when it encounters a date-change record.
CONNECT! Write connect-session records to ctmp. The lineus e file is created, and
the reboot s file , showing all boot records found in nit e/wtmp . mmdd, is
created.
CONNECT2 Convert ctmp to connect-session total accounting records in the file
ctacct . mmdd.
PROCESS Execute acctprc1 and acctprc2 programs to convert the
process-accounting files Spacct ? . mmdd to total accounting records in
14 ptacct? . mmdd. The Spacct and ptacct files are correlated by number
so that if runacct fails, Spacct files will not have to be reprocessed.
One precaution should be noted: when restarting runacct in this state,
remove the last ptacct file. If you don't , runacct will not terminate .
MERGE Merge process and connect-session total accounting records to form
nit e/daytacct .
FEES Merge any ASCII tacct records from the file fee into nite/daytacct .
DISK On the day after the dodisk shell script runs, merge nit e/disktacct
with nit e/daytacct .
MERGETACCT Merge nit e/daytacct with sum/tacct , the cumulative total accounting
file. Each day, nit e/daytacct is saved in sum/tacct mmdd , so that
sum/tacct can be recreated if it becomes corrupted or lost .
CMS Merge today's command summary into the cumulative summary file
sum/ ems . Produce ASCII and internal-format command summary files .

1 4-56 System Accounting

 State Action
USEREXIT Execute any installation-dependent ( local) accounting programs. For
example, you might want to execute commands that generate daily
billing data for individual users , such as the shell script acct_bill
described in System Administration Tasks section "Sample System
Accounting Shell Scripts" .
CLEANUP Clean up the temporary files, run prdaily and save its output in the file
swn/rprt mmdd , remove the locks , then exit .

Recovering from runacct Failure

Sometimes runacct fails and terminates abnormally. The primary reasons for
runacct failure are:
• a system crash
• not enough disk space remaining in /usr 14

• a corrupted wtmp file

For information on how to recover from runacct failure, consult the System
Administration Tasks manual.

Daily and Monthly Reports of runacct

Each time runacct is executed, it generates five reports:
• Daily Line Usage Report
• Daily Resource Usage Report
• Daily Command Summary
• Monthly Total Command Summary
• Last Login

Daily Line Usage Report of runacct

The Daily Line Usage Report summarizes connect-session accounting since the
last invocation of runacct . It logs system reboots, power failure recoveries ,
and any other records dumped into / etc/wtmp via acctwtmp, and provides an
analysis of line utilization.
The FROM/TO banner in the first part of the report identifies the period covered.
The times are the date-time the last report was generated b y runacct , and
the date-time the current report is generated. It is followed by a log of system

System Accounting 1 4-57

 reboots, shutdowns, power failure recoveries, and any other records dumped
into wtmp by the acctwtmp command.
The second part of the Daily Line Usage Report is an analysis of line
utilization. The TOTAL DURATI ON shows how long the system was in a
multi-user state. The columns of the report are defined as follows :

Column Description
LINE The terminal line or access port being reported on.
MINUTES The total number of minutes that the line was in use during the
accounting period.
PERCENT The percentage of TOTAL DURATION that the line was in use :
PERCENT = (MINUTES / TOTAL DURATION) * 100
# SESS The number of times that this port was accessed for a login session.
14
# ON Historically, this column displayed the number of times a port was used
to log a user on. However, since login is no longer be executed to log in
a new user, this column should be identical to # SESS.
# OFF This column reflects the number of times a user logs off and any
interrupts that occur on the line . Interrupts occur on a port when getty
is first invoked, which happens when the system is brought to run-level 2 .
If # OFF exceeds # O N b y a large factor, the multiplexer, modem, or
cable is probably deteriorating, or a bad connection exists somewhere.
The most common cause is an unconnected cable dangling from the
multiplexer.

The I etc/wtmp should be monitored regularly, as this file from the source of
connect-session accounting data. If I etc/wtmp grows rapidly, execute acctcon1
to determine the noisiest line. If interruptions occur at a high rate, general
system performance is affected .

Daily Resource Usage Report

The Daily Resource Usage Report summarizes resource usage, by merging all
total accounting records and displaying the information by individual user.
This report gives a by-user breakdown of system-resource usage. The format
of this report is identical to that produced by the prt acct command. ( For

1 4-58 System Accounting

definitions of the data and format of this report , refer to the discussion of
prtacct in the "Displaying Total Account Records - prtacct" section of this
chapter. )

Daily and Monthly Command Summary

The Daily Command Summary shows resource-usage data about individual

commands since the last invocation of runacct . It is useful in identifying
the most widely used commands, and applying that knowledge of command
characteristics when tuning your system.
This report is sorted by TOTAL KCOREMIN, an arbitrary but useful yardstick for
judging CPU usage.
The Monthly Total Command Summary is identical to the Daily Command
Summary, except for the duration of time reported. The Monthly Total
Command Summary summarizes commands from the start of the fiscal period 14
to the current date, and reflects the data accumulated since the last invocation
of monacct .
The output of these reports is identical to that produced by acctcms . For
definitions of the data found in this report , refer to the discussion of acct cms
in the "Process Accounting" section of this chapter.

Last Login

The Last Login simply shows the last date and time each user logged into the
system. It is useful for finding likely archive material or for removing unused
user directories.

System Accounting 1 4-59

 Creating Daily Summary Reports - prdaily
Prdaily is invoked by runacct as runacct finishes executing. Prdaily
formats the previous day's accounting data, and places the report in the file
/user/ adm/acct/ sum/rpt mmdd where mmdd is the month and day of the
report .
Prdaily can also be used to display any daily report file created by runacct .
The syntax of prdaily is:

prdaily [ -1 ] [ -c ] [ mmdd ]
where:
• The -1 option prints a report of exceptional usage by login name for the
specified date. This option might reveal which users are consuming excessive
14
amounts of system resources. The limits for exceptional usage are kept
in the file /usr/lib/acct/ptelus . awk and can be edited to reflect your
installation's requirements.
• The - c option is valid only for the current day's accounting, and is used to
identify exceptional resource usage by command. The limits for exceptional
usage are maintai�ed in the file /usr/lib/acct/ptecms . awk and can be
edited to reflect your system's needs.
• mmdd is an optional report date. If no date is specified, prdaily produces a
report of the current day's accounting information.

14-60 System Accounting

Creating Monthly Accounting Reports - monacct
The monacct shell script is invoked once a month ( or accounting period )
to summarize daily accounting files and produce a summary files for
the accounting period. The resulting output is stored in the directory
/usr/adm/acct/f iscal . After creating its monthly reports, monacct removes
the old daily accounting files from the directory /usr/adm/acct/sum and
replaces them with new summary accounting files .
Monacct is found on the acctsh(lM) manual page of HP- UX Reference.
The syntax of monacct is

monacct number
where number specifies month or period ( Ol=January, 12=December ) . If
number is not specified, monacct assumes it is being invoked for the current 14
month. This default is useful if monacct is executed via cron on the first day
of each month.
Descriptions of the files created in the acct/fis cal directory follow:
• ems ? contains the total command summary file of the accounting period
denoted by ? . The file is stored in internal summary format . To display this
file, use the acctcms command. The following example shows how to display
this file for June:
$ acctcms - a -s /usr/adm/acct/nit e/f iscal/cms06
• f i scrpt ? contains a report similar to that produced by prdaily. The
report shows line and resource usage for the month represented by ? . To
display the fiscal accounting file for November, enter:
$ cat /usr/adm/acct/nite/f iscal/f iscrpt 1 1
• t acct ? i s the total accounting file for the month represented by ? . To
display this file, use the prtacct command. The following example shows
how to display the total accounting summary file for January:

$ prtacct /usr/adm/acct/f iscal/tacct 0 1 " JANUARY TOTAL ACCOUNTING"

System Accounting 1 4-6 1

 Disk Quotas
Disk quotas are used to limit the number of files and file blocks a user can own
per file system. Using disk quotas , the system administrator can assign file and
block quotas to users on any writable file system.
HP-UX disk quotas work on both standard and NFS-mounted file systems and
in a clustered environment ( Series 300/400/700 only ) .
Disk quotas consist of tools to
• manage disk resources equitably by limiting disk usage
• help prevent users from accidentally filling a file system

Planning to Set up Disk Quotas

14 Disk quotas are already configured into your operating system. Using disk
quotas involves two basic steps:
• determining which file systems warrant disk quotas
• establishing soft limits ( quotas ) and hard limits ( limits ) on users' disk
consumption
Use the quotas option to / etc/checklist entries and the quotacheck( lM)
command to I etc/bcheckrc to have the system invoke disk quotas
automatically at boot time.

Deciding Which File Systems Require Disk Quotas

You need to decide which file systems warrant disk-quota controls . Typically,
only file systems containing users ' home directories and possibly the /usr file
system require disk quotas . Do not assign quotas to the /tmp directory, which
is historic ally a less restricted environment that is periodically purged of excess
files by other means .
You can set disk quotas on all users on a file system, or only specific users,
by using the edquota command, discussed later in this chapter. You can also
establish separate limits for the number of files ( designated as inodes ) and
number of blocks ( given in units of one kilobyte ) per user.
When setting up disk quotas, be sure these file systems have adequate available
space for data collection. Note that the quot as file grows based on the numeric

1 4-62 System Accounting

value of the user ID . Therefore, if you are setting up disk quotas, you might
want to keep the numeric value of your user IDs to low numbers.

Setting Quotas and Limits

Imposing disk quotas involves setting soft and hard limits on the amount of file
space a user can consume. The soft limit , or quota, is the amount of file space
allocated for a given user, and can be exceeded temporarily. The hard limit , or
limit , is the maximum amount of file space allowed by a given user.
The grace period for exceeding the quota is set by the system administrator, or
left at the default of one week. Once the grace period expires , a quota becomes
a hard limit .
If a user exceeds the hard limit , a disk quota mechanism in the operating
system kernel prevents files from being written to disk, thus compelling the
user to remove files . 14

Any file open in violation of the disk quota cannot be written and its original
contents are destroyed. Operations that truncate or remove then rewrite files
might fail, because the system refuses to allocate new resources to the offending
user. Such operations might result in loss of data. For this reason, users
should be warned to heed system warnings when they exceed quotas and take
corrrective action as soon as possible, when the system warns them they have
exceeded quota.

How Disk Quotas Work

The following diagram shows the model by which disk quotas are maintained:

System Accounting 1 4-63

 Data
Structures
disk-quotas
commands quotactl (2)
quotas
File

L.G2001 72Jl09

Figure 1 4- 1 0. Disk Quotas Flow Chart

All disk-quota commands work via the system call quotactl(2), which
14
communicates with the kernel via two disk-quota data structures:
• struct dqblk data structure, defined in quota . h
• struct mount mount tables defined in mount . h . The mount table has a flag
indicating whether quotas are enabled.
The quotactl(2) system call commands are used to update both the data
structures and the quotas file with mappings of each file system. The quotas
file retains information on disk-quota settings and utilization statistics for the
operating system.
When file systems are mounted, the kernel sets up internal tables in memory
containing statistics on usage and limits. The kernel uses these internal tables
to maintain information about disk resources allocated or released for users
with quotas. The information is :flushed to the quotas files on each file system
whenever the file system is unmounted ( such as during a shutdown ) , to prevent
losing the information.

Disk-Quotas Data File-quotas

A quot as file, created at the root of each affe cted file system, stores limits
and utilization statistics for each user for the entire file system. To enable
disk quotas, you must create a quotas file. It is not an ASCII file, but can be
accessed in ASCII form by using the edquota command.

1 4-64 System Accounting

Each user for whom quotas are being established in a file system must be
specified as an entry in the quotas file. The edquot a command allows you
to specify individual disk-quota values for users. It also lets you establish
template values for a prototypical user, and copying those values to selected
users on the system.
The quotas file can be written to under the following circumstances:
• when edited with edquota( lM)
• when a file system is unmounted
• when quotacheck writes to disk
Information is stored in quotas according to the elements of the dqblk data
structure, defined in quot as . h. These include:
• block limit (limit ) -the maximum allowable number of disk blocks.
• block quota ( quota)-the preferred limit of disk blocks .
14
• block usage-current block count being consumed.
• file limit-maximum number of allocated files.
• file quota-preferred file limit .
• file usage-current number of allocated files.
• time limit for excessive block use.
• time limit for excessive files.
The quotas file is read by repquota( lM) for current quota statistics. It is
written ( updated ) by quotacheck and whenever the file system is unmounted.

How Disk Quotas Appear to Users

As long as they do not exceed allocated disk quotas, users will not notice the
presence or enforcement of disk quotas. However, upon logging in, login alerts
any user who exceeds the allocated disk quotas of this fact .
Users can execute the quota(l) command to display their own disk usage and
limits. If they exceed disk allocation, they need to remove files to conform to
disk-quota requirements.
Superusers can use quota(!) to view the limits of any user or repquota( lM) to
report on entire file systems .

System Accounting 1 4-65

 Disk Quota Commands
A variety of commands are available to the system administrator for managing
disk quotas . These commands fall into several categories:
• enabling and suspending disk quotas
• maintaining file-system consistency while implementing disk quotas
• setting and editing user quotas
• displaying and reporting on disk quotas

Enabling and Suspending Disk Quotas- quotaon and quotaoff

quotaon(lM) activates disk quotas on one or more file systems. Its syntax is:
/etc/quotaon [ -v ] -a
14
/etc/quotaon [ -v ]file_sys . . .

quotaoff( lM) suspends enforcement of disk quotas on one or more file systems.
Its syntax is:

/etc/quotaoff [ -v ] -a
/etc/quotaoff [ -v ]file_sys . . .

The following options and arguments apply to both quotaon and quotaoff :
-v Produce verbose output.
-a Enable or suspend disk quotas on all mounted file systems
cited with the quota option in I etc/ checklist .
file_sys Enforce or suspend disk quotas on specified mounted file
systems. The file_sys argument is the name of either the
directory on which the file system is mounted or the block
special file of the file system.
Either -a or file_sys must be specified.
Both quotaon and quotaoff work by calling quotactl(2). quotaon and
quotaoff update the status field of devices located in /etc/mnttab to indicate
whether quotas are being enforced.

1 4-66 System Accounting

Maintaining File-system Consistency-quotacheck

quotacheck( lM) is used to maintain usage statistics on a file system with disk
quotas . Each file system being checked must have a quotas file at its root .
Quotacheck examines each specified file system and builds a disk-usage table
of user IDs and blocks used, which it compares to a table stored in that file
system's quotas file. If inconsistencies are detected, both the quotas file and
the kernel quotas table are updated. ( Inconsistencies typically develop if a
system mounted with disk quotas has been used for a period of time with
quotas enforcement suspended. )
Quotacheck is executed at various times:
• During a reboot , quotacheck(lM) can be run from /etc/bcheckrc, before
activating disk quotas with quotaon.
• When interactively mounting a file system assigned quotas , quotacheck
14
should be run after the mount .
/etc/quotacheck [ -v ] -a
/etc/quotacheck [ -v ]file_sys . . .
Like quotaon and quotaoff , quotacheck uses a -v ( verbose ) option, -a
option instructing the system to check all mounted file systems listed in
/etc/checklist , and file_sys argument , for checking specific mounted file
systems.
Since quotacheck accesses the raw device in calculating the actual disk usage
for each user, the file systems checked should be inactive when quotacheck is
executed. This is most efficiently done in single-user mode.

Setting and E diting User Quotas-edquota

edquota(lM) is the command used to edit the quotas file, something you must
do to set disk quotas . Its syntax for editing user quotas is:

/etc/edquota [ -p proto-user ] user . . .

I etc/ edquota -t
The first syntax line shows two ways of specifying quotas for a user:

System Accounting 1 4-67

 • Use the user argument t o specify disk quotas for one or more users.
• Use the -p proto-user option with the user argument to copy disk quotas
from a prototypical user to one or more users.
When invoked, edquota creates a temporary AS CII file with a one-line
representation for each user specified on the command line, showing the current
disk quotas set for that user. You can then modify or add quotas to the file
using an editor ( environ variable EDITOR, or vi by default ) .
A temporary file created by edquota for editing a new user's disk quotas might
look like this:
fs /mnt blocks (soft=100 , hard=120) inodes (soft=O , hard=O)
If the quota command is invoked for that user, the following is displayed:
$ quota -v userl
14
Disk quot as for userl (uid 101 ) :
Filesystem usage quot a l imit t imeleft files quot a l imit t imeleft
/users 0 100 120 0 0 0

Editing Disk-Quota Time Limits. Time limits are used to enforce disk quotas.
Enforcement changes from warning message to restriction after a specified time
limit . The syntax for specifying the time limit for a given file system is:

/etc/edquota -t
When invoked, edquota -t creates an AS CII representation of the quotas file
much like this:

fs I blocks time limit =O (default ) , files time limit =O (default )

The default time limit is defined in quot a . h as one week. You can specify the
time limit in terms of seconds, hours, days, weeks, and months.
If a user reaches the quota ( soft limit ) , the system generates a warning, begins
to keep track of how much time the user exceeds quota, but allows the user to
continue working.
The quota output for the user might then look like this :
11

1 4-68 System Accounting

 $ quota -v user1

Disk quotas £or user1 (uid 101) :

Filesystem usage quot a limit t imele£t £iles quot a limit t imele£t
/users 114 100 120 4 days 40 0 0

However, if the user exceeds the time limit by ignoring warnings to remove
files or reaches the limit by creating more files or blocks than alloted, the
system prevents the user from writing files. In fact , any file the user attempts
to open and write might be destroyed and the following message (in this case,
by friendly vi(l)) is displayed:
DISK LIMIT REACHED - WRITE FAILED (/users )
Your edited changes will be lost if you do not complete
a successful write command before exit ing . If you were
writing out to your original file , the previous contents 14
of that f ile have been destroyed . Contact your System
Administrator BEFORE exiting if you need ass istance .
[Hit return to cont inue . ]
The user is thus compelled to remove files to below the assigned disk quota.
One thing to bear in mind (and inform users on your system): do not panic
when you get a "WRITE FAILED" message. If you work in a windows
environment, open up another window. If you are not running X-Windows or
some other window manager, suspend the job. From the shell, remove some
unnecessary files, then return to the process. The process is still in memory,
and once you return to it , if you have brought the number of files or blocks
below quota, you will then be able to save the file.

Displaying Users' D isk Quotas-quota

The single most-commonly used disk-quota command is quota(l), because it

can be used by all users to stay abreast of their own disk-usage status.

System Accounting 1 4-69

 Quota displays a user's disk quota and usage. Used without options, the
command displays warnings about mounted file systems where usage
exceeds quota. Remotely mounted file systems ( NFS ) for which disk quotas
are enforced are reported also. ( See section on rquotad below for more
information. ) The -v option shows all current statistics for the user. For a new
user with disk quotas , the output for the quota command might look like this:
$ quot a -v

Disk quot as for user1 (uid 101 ) :

Fil.esystem usage quot a limit t imeleft fil.es quota limit t imeleft
/users 0 100 120 0 0 0

As the user creates files, the number of blocks and files the user owns is shown
in the usage and files columns.

Enforcing D isk Quotas Remotely-rquotad

14
If the user has quotas on remotely mounted ( NFS ) file systems, quota
can report them using the remote quota server daemon, rquotad.
/usr/etc/rpc . rquotad is invoked remotely via inetd(lM).
The syntax of rquotad is:

/usr/etc/rpc . rquotad
Manipulating Disk Quotas Programmatically-quotactl

quotactl(2) is a system c all that manipulates disk quotas. Using quotactl(2),

you can :
• Turn quotas o n for a file system.
• Turn quotas off for a file system.
• Retrieve quota usage information.
• Change quota usage information.
• Set file and block quotas and limits.
• Flush quota cache information to disk (quotas)

Summarizing Disk Usage and Quotas-repquota

The repquota(lM) command is used to read the kernel's internal disk-quota

tables and print a summary of disk usage and quotas for specified file

1 4-70 System Accounting

systems . The summary shows the current number of files, amount of space (in
kilobytes ) , and assigned quotas for each user.

repquota quotas
File
quotactl(2)
Summary of Disk Data
Usage and Quotas Structures

l.G2001 72_010

Figure 14-1 1 . H ow repquota Generates Disk-quota Reports 14

In the following sample summary, repquota is invoked, using the v verbose -

option, to report on two specific file systems (I and /users). The -a option
can be used instead to read all mounted file systems in /etc/checklist :
$ repquota -v I /'users
/dev/root (/) :
Block limit s File limit s
User used soft hard t imeleft used soft hard t imeleft
julie 20 100 2000 4 50 400
fred 400 100 2000 50 min 20 50 400

/dev/dsk/1s0 (/users ) :
Block limits File limit s
User used soft hard t imeleft used soft hard t imeleft
bill 593 1000 1 200 192 500 700
julie 947 1000 1 200 700 500 700 IDT STARTED
fred 4150 3000 6200 4 days 800 1000 1500

The output shows three users, bill, j ulie, and fred. The system is set up
so that they can create a small number of files in the root directory, but are
allotted more space in the /users file system. j ulie and fred have files in the
root directory as well as in /users .

System Accounting 1 4· 7 1
 User fred has exceeded his quota (soft limit) of lK blocks allocated on both I
and /users . login warns him of the condition whenever he logs in, but having
neglected to heed warnings and decrease block numbers below acceptable levels,
fred now has only 50 minutes before the soft limit is enforced as a hard limit
in the I directory. He faces the same restriction in /users if he fails to cut his
block usage below the 3000K-block quota.
Once time runs out , the t imeleft field will read EXPIRED . The kernel will
allow fred to log in, but the login program will report to him that he has
exceeded his quota. Any login process that creates space on the disk will fail.
Any command that he runs that creates new blocks will fail. All he will be able
to do is delete files or the contents of files, until he has brought his account
below quota.
User j ulie faces a similar problem. Although she has not exceeded her block
14
allocation, she has not only exceeded her quota for number of files, but has
reached the hard limit of number of permissible files. Once she exceeded quota,
she too received login warnings . Having reached the hard limit , she is unable
to create files on /users .
The t imeleft field shows j ulie's transgression somewhat cryptically with the
notation "NOT STARTED" . When a user reaches allotted hard limits, whether
block or file, time is irrelevant . The system no longer permits the user to do
anything other than comply with disk quotas.

1 4-72 System Accounting

Glossary

address
In the context of peripheral devices, a set of values that specify the location
of an 1/0 device to the computer. On the Series 300/400, the address is
composed of up to four elements: select code, bus address, unit number
and volume number. On the Series 800, the address is composed of up to
five elements: bus converter, module number, slot address, device switch
setting, and logical unit number. Addresses on the Series 700 are composed
of device, module number, and switch setting.
Addresses are covered in more detail in Chapter 1 1 , "System Configuration"
and the manual Installing Peripherals.
advisory lock Glossary
A file lock placed on a file to inform (warn) other processes wanting to
access the file that it is already being accessed and potentially being
modified. Advisory locks are useful only among cooperating processes that
use file locking. (See Chapter 8 , "HFS File System" .)

access
An interaction between a subject and object resulting in information
:fl.owing from one to the other.
access control list (ACL)
A set of entries associated with a file that specify permissions for all
possible user-ID/group-ID combinations. (See acl(5) in the HP- UX
Reference and HP- UX System Security.)
access control mechanism (ACM)
An algorithm and data structure that supports access control decisions.
It mediates decisions about whether specific subjects can access
specific system objects in specific ways. An ACM might be implicit

Glossary- 1
 (contextual) ,e.g. "only user ID 0 can access files in /etc" . It might be
explicit , as in a list of objects and their access rights. An explicit ACM
might be distributed (data stored with protected objects), as in the UNIX
file permissions scheme, or it might be centralized, as in a file describing
access rights for all protected system objects.
allocation policy
Allocation policy governs how disk space is distributed. When using the
Logical Volume Manager (LVM) , the allocation policy governs how disk
space is distributed to logical volumes and how extents are laid out on an
LVM disk.
(See Chapter 9, "Logical Volume Manager" .)
archived library
Object files that are linked to a program when called, regardless of whether
another copy of that library already exists in memory. Compare with
shared library.

ARPA hostname
Needed for ARPA Services. A name consisting of one field containing any
Glossary printable character except spaces, newlines, or the pound ( See NS_ARPA

Services documentation.) sign ( # ) . Assigned in /etc/hosts .

attended mode
Occurs when the user selects an operating system during system startup ,
rather than letting the boot ROM automatically find an operating system.
(See Chapter 2 , "System Startup" . )

attribute
In an HP-UX cluster, the attributes are characteristics of the cluster
nodes , such as the cluster node name or processor type. The attributes
are collectively referred to as the cluster node's context . (See Managing
Clusters of HP 9000 Systems.)
audit trail
A set of records that collectively provide documentary evidence of
processing, used to aid in tracing from original transactions forward to
related records and reports, and/or backwards from records and reports to
their component source transactions.

G lossary-2
audit ID
An ID associated with each user when trusted systems are used. The audit
ID does not change, even when executing programs with a different user ID .
( See Chapter 5, "Process Management" and HP- UX System Security.)
auto creation
If you create a CDF, but specify only the path name of the CDF and not
a subfile, the system automatically creates a CDF subfile named after the
cluster node name attribute. This is known as autocreation. ( See Managing
Clusters of HP 9000 Systems .)
available memory
Memory is used by the system for executing user processes. It is the
amount of physical memory ( main memory ) on your system not reserved
by the HP- UX kernel for device drivers, system data structures, and other
kernel needs. ( See Chapter 7, "Memory Management" . )
background process group
A process group that is currently not running as the foreground process
group. Typically, only one process group at a time is the foreground process
group in a session, while all other process groups are background process G lossary
groups.
bad block relocation
A feature of LVM for recovering corrupted blocks of data from HP-FL and
SCSI disks, by redirecting the data to another block of media on the disk.
( See Chapter 9, "Logical Volume Manager" . )
binding
The process by which the operating system's self-configurating capabilities
find hardware on the bus architecture and associate it with relevant
software ( device files and device drivers ) . ( See Chapter 1 1 , "System
Configuration" . )
bit
The smallest unit of information in a digital computer. Its value can be
either 0 or 1. Bits are normally grouped together in bytes and words .

G lossary-3
 bits per inch
A common measure of "tape density." Indicates how many bits can be
stored per inch of magnetic tape. ( See Chapter 12, "HP-UX Peripherals" .)
block
The fundamental unit of information HP-UX uses for access and storage
allocation on a mass storage medium. Block size on an HFS file system
can be either 4 KB or 8 KB, and is set at file system creation. The default
block size is 8 KB. However, to present a more uniform interface to the
user, most system calls and utilities use "block" to mean 512 bytes,
independent of the actual block size of the medium. (See Chapter 1 1 ,
"System Configuration" . )
block device
A hardware device that transmits and receives data in blocks . (See Chapter
1 1 , "System Configuration" . )
block mode
Also termed buffered I/O . Data is held in the buffer cache, then transferred
one block at a time. Block size for buffered I/O is not the same as
Glossary block size on the file system. Block size for buffered I/ 0 is defined as
BLKDEV _IO SIZE in /usr/include/sys/param . h. Compare with raw
mode. (See Chapter 1 1 , "System Configuration" .)

boot or boot-up
The process of loading, initializing and running an operating system. (See
Chapter 2, "System Startup" . )
boot area
Eight kilobytes of the disk that are reserved for system use. This area
contains the LIF volume header, the directory that defines the contents of
the volume, and the bootstrapping program. On the Series 300 and 400, the
8 KB are at the beginning of the disk; on the Series 700, they are at the
end of the disk; and on a standard implementation of Series 800, they are
in their own section (typically section 6) or logical volume. On a Series 800
implementing either LVM for the root file system or the SwitchOver/UX
product , the LABEL file is stored in the boot area. (See Chapter 8, "HFS
File System" .)

Glossary-4
boot process
A process that runs during the HP-UX phase of system startup. For
example, I etc/re and I etc/brc are boot processes . ( See Chapter 2,
"System Startup" . )

boot ROM
The boot ROM is small, machine language program that resides in your
computer's read-only memory. When you boot or re-boot the system, the
computer starts the boot ROM program which takes control of the system.
The boot ROM tests computer hardware, finds some devices accessible
through the computer, and loads an operating system. ( See Chapter
2, "System Startup" in this manual or Chapter 5, "System Boot-Up
Problems" in Solving HP- UX Problems .)

bpi
See bits per inch.
buffer cache
A buffer which remains in memory and which is used by the file system for
buffering writes to the file system. ( See Chapter 8, "HFS File System" . )
G lossary
bus
Hardware that carries data and signals between hardware modules within
the system. Typically, a bus is implemented as soldered parallel wires either
on a frame to which boards are plugged, on a board housing multiple chips,
or internal to a processor for transmission to and from memory, other
processors, or other buses.
bus address
Part of an address used for devices ( such as HP-IB , SCSI, EISA, or
HP-FL ) ; a number determined by the switch setting on a peripheral that
allows the computer to distinguish between two devices connected to the
same interface. A bus address is sometimes called a device address; no two
devices on the same HP-IB can have the same bus address. ( See Chapter
1 1 , "System Configuration" . )
bus converter
A board that serves as an internal interface between higher- and
lower-speed bus, typically the SMB and Mid-Bus. The bus converter
handles timing and bus protocol management . Because more than one can

G lossary-5
 be connected to the SMB , bus converters can also increase the number of
modules in a system, thus enabling more devices to be connected.
byte
A basic unit of digital computer information, consisting of 8 consecutive
bits.
byte offset
On the Series 800, a 32-bit quantity that points to the specific location
of the virtual address space being accessed. (See Chapter 7, "Memory
Management" .)
bytes per inode
This specifies the number of data bytes (amount of user file space) per
inode slot . The number of inodes is calculated as a function of the
file system size. The default value is 2048 . (See Chapter 8, "HFS File
System" .)
cache
A small high-speed memory between RAM (main memory) and the
processor that may hold data and text for active processes. Cache designed
Glossary
into computer-system architecture accelerates effective memory transfer
rates and processor speed. If the data is not in cache, the processor consults
the translation tables in RAM that hold the mapping between virtual and
physical addresses of the data. ( See Chapter 7, "Memory Management" .)

central processing unit (CPU)

The instruction-processing module inside the computer, typically including
the processor, :floating-point co-processor, and special registers . On
multi-processing computers (computers with more than one processor), the
term system processing unit (SPU) is used.
certification
The technical evaluation of a system's security features, made as part of
and in support of the approval/accreditation process , that establishes the
extent to which a particular computer system's design and implementation
meet a set of specified security requirements.

Glossary-&
CDF
Context-dependent file. The mechanism used by HP-UX clusters to share a
path name between different cluster nodes in a cluster. A CDF consists of a
hidden directory and one or. more subfiles. ( See Managing Clusters of HP
9000 Systems .)
channel adapter manager ( CAM)
On the Series 800, the type of device driver ( cio_caO ) for the CIO bus that
provides the software interface to device adapters ( such as hpibO, muxO,
and scsi2 ) . The channel adapter manager manipulates data between the
memory mapped I/O and CIO bus .
channel I/ 0 adapter
On the Series 800, a chip on a board or one or several Mid-Bus cards that
serve as an internal interface between the Mid-Bus and the lower-speed CIO
bus. The channel adapter synchronizes the different speeds and bandwidths
of the Mid-Bus and CIO bus. The channel adapter also manages direct
transfer of data to and from main memory with minimal CPU intervention,
and by extension, to the rest of the I/O system. More than one channel
adapter can be installed on a system, thus expanding the number of
possible modules. Glossary

character device
Typically, a device which is not a block device and which transfers data a
character a time. Printers, plotters, terminals, magnetic tape drives, and
pseudo devices are all examples of character devices. (See Chapter 1 1 ,
"System Configuration" . )
clean byte
A flag in the primary superblock which the fsck command uses to
determine whether the file system state is consistent . (See Chapter 2,
"System Startup" and Chapter 8, "HFS File System" .)
cluster
An HP-UX cluster is one or more workstations linked together with a local
area network (LAN ) , but consisting of only one file system, HP-UX clusters
are supported on Series 300/400/700 only. (See Managing Clusters of HP
9000 Computers.)

Glossary-7
 cluster client
A cluster node that does not have the root file system for the cluster on its
local disks. Its file system resides on the cluster server. However, cluster
clients can have locally mounted disks for local data storage. ( See Managing
Clusters of HP 9000 Computers.)
cluster node
A computer in a cluster. ( See Managing Clusters of HP 9000 Computers.)
cluster server
The cluster node that acts as a file-system server for all the clients in an
HP-UX cluster. ( See Managing Clusters of HP 9000 Computers.)
cluster server process (CSP)
A special kernel process that runs on a machine in a cluster to satisfy
requests from other nodes in the cluster. There are two kinds of CSP, the
limited CSP ( LCSP ) and the general CSP ( GCSP ) . ( See Managing Clusters
of HP 9000 Computers.)
code segment
The memory in segment into which a process 's executable code is loaded by
GIossary
exec. S ame as the text segment . ( See Chapter 7, "Memory Management" . )
comment lines
In the context of system accounting, lines in the holidays file that
begin with an asterisk ( * ) in the first column. ( See Chapter 14, "System
Accounting" . )
company holiday lines
Used in the system accounting file, /usr/lib/acct/holidays , to denote
holidays. ( See Chapter 14, "System Accounting" . )
configuring
Telling the operating system what physical hardware is present , by
associating the hardware with necessary software-device drivers and device
files. ( See Chapter 1 1 , "System Configuration" . )
connect session
This denotes the period of time in which a user is connected to the system.

Glossary-8
 It starts when the user logs in and finishes when the user logs out. ( See
Chapter 14, "System Accounting" . )
context
The mode ( either system or user ) a process is in when executing. ( See
Chapter 5 , "Process Management" . )
In HP-UX clusters, context is an AS CII string comprised of a number
of attributes that determine which subfile ( if any ) is accessed from a
CDF. Each workstation on a cluster has an associated context set at boot
time. The context attributes include cluster-node name; processor type;
floating point hardware; file system location; and the string "default" . ( See
Managing Clusters of HP 9000 Computers.)
context-dependent file
A file in a cluster having different contents, depending on which cluster
node uses it . A context-dependent file consists of a hidden file and one
or more subfiles. Also called a CDF. ( See Managing Clusters of HP 9000
Computers.)
controlling process
The session leader that connected to the controlling terminal. ( See Chapte:P l ossary
5 , "Process Management" . )
controlling terminal
Every session has exactly one controlling terminal, which all processes
in that session, by default , use for standard input, standard output , and
standard error. Every controlling terminal is associated with, at most , one
session. ( See Chapter 5, "Process Management" . )
copy-on-write
On Series 300/400, when a process forks, the parent process plans a
duplicate image of itself at child virtual addresses. But instead of actually
copying the image, the parent pages are marked "copy-on-write." The pages
continue to reside in the parent address space, and are not copied until the
process actually writes on the page. On Series 700/800, HP-UX implements
a variation of copy-on-write, called "copy-on-access." ( See Chapter 7,
"Memory Management" . )

G l ossary-9
 cron
This process wakes up every minute to execute commands at specified dates
and times, according to instructions in files contained in the directory
/usr/spool/cron/crontabs . See the cron(lM ) and crontab(l ) entries in
the HP- UX Reference for more details.
CS/80
A family of mass storage devices that communicate with a computer via the
CS /80 (Command Set '80) or SS/80 ( Sub Set '80) command set. In HP-UX
your file system can be on either a SCSI drive or a CS /80 drive.
cylinder
One or more vertical collections of tracks in a disk pack. Disk accesses
within a cylinder do not need a seek. (See Chapter 8, "HFS File System" .)
cylinder group
One or more consecutive cylinders. Each cylinder group contains a
superblock, inodes, cylinder group information, and data blocks. (See
Chapter 8, "HFS File System" .)

G Iossa rVlinder group information

A data structure located in the cylinder group that contains information
about the cylinder group such as numbers of available inodes, data blocks,
and fragments, and bitmaps to free space in the cylinder group. (See
Chapter 8, "HFS File System" .)
daemon
Background process that performs systemwide functions. For example,
pagedaemon, vhand, and swapper are daemons that are part of the HP-UX
memory-management subsystem.
daily command summary
A report produced by system accounting, which summarizes command
usage for the day. (See Chapter 14, "System Accounting" .)
daily line usage report
A report produced by system accounting, which summarizes tty line usage
for the day. (See Chapter 14, "System Accounting" . )

Glossary- 1 o
daily resource usage report
A report produced by system accounting, which summarizes resource
(e.g. , disk, memory, cpu) usage for the day. ( See Chapter 14, "System
Accounting" .)
DASS
Direct Access Secondary Storage. A form of mass storage with
characteristics of both online magnetic disk primary storage and offiine
magnetic tape secondary storage. ( See Chapter 12, "HP-UX Peripherals" in
this manual and autochanger ( 7 ) in the HP- UX Reference Manual. )
data segment
A memory segment into which a process 's static and dynamic data is
stored. ( See Chapter 5, "Process Management" .)
demand paging
A form of virtual memory management in which pages are brought into
memory only as they are accessed. At any one time, only a subset of the
process need be loaded into main memory. (See Chapter 7, "Memory
Management" . )
Glossary
demand-loadable
Even though HP-UX is a demand-paged virtual memory system, it still
attempts (by default) to load an entire process into main memory. If,
instead, code is marked as demand-loadable, then the process's code will be
brought into physical memory only as pages are accessed.
demand-paged virtual memory
This is the type of memory management implemented on Series 300 and 800
computers. See also demand paging.
destination device
The mass storage device on which HP-UX is to be installed or updated.
The destination device must be a hard disk drive.
device adapter
Interface cards (such as HP-IB , HP-FL , S CSI, LAN , and MUX cards)
installe d in the I/O slots of the CIO , HP- PB , EISA , or other bus. Device
adapters make the crucial link that enables the system to communicate
with external peripherals (such as terminals, printers, disk drives) . Device

Glossary- 1 1
 adapters are frequently called 1/0 cards. CIO device adapters are also
called CIO cards; HP-PB device adapters are called HP-PB cards. ( See
Chapter 1 1 , "System Configuration" and Installing Peripherals.)
device adapter manager (DAM)
On the Series 800, the type of device driver that deals with the
device-adapter protocol. For example, hpibO handles HP-IB data protocol
transfer for the CIO bus; hpib 1 handles HP-IB data protocol transfer for
the HP-PB bus. (See Chapter 1 1 , "System Configuration" .)
device manager (DM)
On the Series 800, the type of device driver (such as mux ) that translates
requests from the CIO bus to the attached interface card. (See Chapter 1 1 ,
"System Configuration" .)
device driver
On the Series 300/400/700, the term used for the driver that controls kernel
data structures from the interface card to the external device itself. On the
Series 800, device driver is also used more generally to mean the software
that enables the operating system to transmit data between the computer
Glossary and the external device. Series 800 device drivers include channel 1/0
adapter manager (CAM), device manager (DM), device adapter manager
(DAM ) , and logical device manager (LDM ) . (See Chapter 1 1 , "System
Configuration" .)
device :file
See special :file.
device swap space
Swap space that occupies a disk, section, or logical volume of a disk
reserved exclusively for swapping. It is fast . Because at least one swap
device must be present on a system, device swap space is also referred to as
primary swap space. (See Chapter 7, "Memory Management" .)

Discretionary Access Control (DAC)

A means of restricting access to objects based on the identity of subjects
and/or groups to which they belong. The controls are discretionary in
the sense that a subject with a certain access permission is capable of
passing that permission (perhaps indirectly) on to any other subject (unless
restrained by mandatory access control).

Glossary- 1 2
disk quotas
A subsystem used to limit the number of files and file blocks a user can own
per file system. (See Chapter 14, "System Accounting" . )

disk sectioning
On Series 800 systems, disks can be divided into sections, each one
appearing to the operating system as if it were a separate disk. This
is known as disk sectioning or partitioning. A more flexible Series 800
alternative to disk sectioning is provided by the Logical Volume Manager
(LVM) . ( See Chapter 8, "HFS File System" and Chapter 9, "Logical
Volume Manager" .)
diskless workstation
Same as a cluster client .
driver
Compiled code (supplied with HP-UX) , which resides in the kernel and
handles input and output with peripheral devices. Also called device driver.
( See Chapter 1 1 , "System Configuration" . )
driver number
Glossary
A number identifying a device driver. Same as major number. ( See
Chapter 1 1 , "System Configuration" . )
disk
Collection of recording platters contained in a single disk drive or disk-drive
library.
domain
The set of objects that a subject has the ability to access. A set of
(object , rights) pairs. Each pair specifies an object and some subset of
the operations that can be performed on it . A right in this context means
permission to perform one of the operations.
dynamic swap space
Swap space allocated as needed while the system is running. Compare this
with primary swap space. Both device and file-system swap can be added
dynamically. ( See Chapter 7, "Memory Management" .)

Glossary- 1 3
 effective user ID
A form of user ID that allows users access to files they do not own. (See
Chapter 5, "Process Management" .)
effective group ID
A form of group ID that allows users access to files not in their group. (See
Chapter 5, "Process Management" .)
enforcement mode
A form of file locking in which a user is not allowed to read or write to a
file while another user is accessing the file. In such a case, the requesting
process sleeps until the process that has control of the file releases it . (See
Chapter 8, "HFS File System" .)
exchange
An exchange occurs when an optical autochanger replaces one disk surface
in the optical drive with another. (See Chapter 12, "HP-UX Peripherals" .)
extent
Fixed-size addressable areas of space on an LVM disk or in memory.
On disk, these areas are called physical extents, and are by default 4
Glossary
MB . Physical extents map to areas in memory, called logical extents. If
disk mirroring (HP MirrorDisk/UX product) is used, each logical extent
corresponds (maps) to more than one physical extent . If disk mirroring
(HP MirrorDisk/UX product) is not used , there is a one-to-one relationship
between the two types of extents. ( See Chapter 9, "Logical Volume
Manager" .)
file
A discrete collection of information described by an inode and residing on a
mass storage medium.
file locking
Measures which help to ensure that only one user at a time can access a
particular file or files . See enforcement mode and advisory lock. (Also, see
Chapter 8, "HFS File System" .)
file mark
A special type of record written to tape and recognized as a boolean
condition during reading. Single file marks separate records on tape; two

Glossary- 1 4
 consecutive file marks indicate the end of tape. (See Chapter 12, "HP-UX
Peripherals" .)
file-system swap space
A secondary form of swap space, enabled by sam and set in
I etc/ checklist , that allows a process to use an existing file system if
device swap space is insufficient to meet demand-paging needs.
File-system swap is slower than device swap but varies in size with the
system's swapping activity. (See Chapter 7, "Memory Management" )
file types
The file type is established at the time of the file's creation. The types are:
• Regular files - Contains a stream of bytes. Characters can be either
ASCII or non-AS CII. This is generally the type of file a user considers to
be a file: object code, text files, nroff files, etc.
• Directory - HP-UX treats directories like regular files, with the exception
that writing directly to directories is not allowed. Directories contain
information about other files.
• Block special files - Device files that buffer the I/ O . Reads and writes to
block devices are done in block mode. Glossary
• Character special files - Device files that do not buffer the I/ 0 . Reads and
writes to character devices are in raw mode.
• Network special files - contain the address of another system.
• Pipes - A temporary file used with command pipelines. When you use a
pipeline, HP-UX creates a temporary buffer to store information between
the two commands. This buffer is a file, and is called a pipe.
• FIFO - A named pipe. A FIFO (First In/First Out) has a directory entry
and allows processes to send data back and forth.
• symbolic link - A type of file that indirectly refers a path name.
file system
The organization of files and directories on a hard disk. The HFS file
system is an implementation of the HP-UX directory structure. NFS
provides access to a file system over a network. (See Chapter 8 , "HFS File
System" . )

Glossary- 1 5
 foreground process group
The process group in a session which currently is executing in the
foreground and has control of the controlling terminal. (See Chapter 5,
"Process Management" .)
fragment
A piece of a block. This is the smallest unit of information HFS will
read or write. The lower limit of a fragment is DEV _BSIZE (defined in
/usr/include/sys/param . h) . Fragment size is set at file system creation.
(See Chapter 8, "HFS File System" .)
frame
A single, nine-bit character on a magnetic tape. ( See Chapter 12, "HP-UX
Peripherals" .)
free space threshold
Specifies minimum percentage of free disk space allowed. Once the file
system capacity reaches this threshold, only the system administrator is
allowed to allocate disk blocks. The default is 10% ; if it is less, file system
performance degrades. The free space threshold is set when you create a
Glossary new file system. Same as minfree. (See Chapter 8 , "HFS File System" .)

fs_bsize
The size of disk blocks in the file system. (See Chapter 8, "HFS File
System" . )
generic device file
A device file whose cluster node ID is O , thus allowing it access by all
cluster nodes . (See Managing Clusters of HP 9000 Systems ; (Series
300/400/700 only.)
gateway page
A Series 800 address range used for switching between user and kernel
privileges when making system calls.
group access list
A list of groups whose users are allowed access to a file. (See Chapter 5,
"Process Management" and Chapter 8 , "HFS File System" .)

Glossary- 1 6
halting
Bringing the system to a complete stop, a state in which no processes are
running. The only way to get HP- UX running again is to cycle power or do
a hardware reset. (See Chapter 2 , "System Startup" .)
hard limit
In disk quotas, a hard limit (also termed limit) is the maximum amount of
file space allowed by a given user. (See Chapter 14, "System Accounting" .)
hardware path
The sequence of hardware addresses from bus converter (if applicable) , to
module, to device adapter, to device.
HFS :file system
High performance File System, the file system implemented on HP-UX 9000
systems. (See Chapter 8, "HFS File System" .)
hidden directory
A directory used to implement a CDF. It is called hidden because it is
normally treated and seen as a file. It can be accessed as a directory
only by appending the special character "+" to its name. (See Managing
Glossary
Clusters of. HP 9000 Systems .)
home directory
The directory into which a user is placed after login. The user's home
directory is defined in the I etc/passwd file. ( See Chapter 4, "Login" . )
homogeneous cluster
An HP-UX cluster containing only systems of the same architecture (for
example, all Series 300 and 400 or all Series 700). (See Managing Clusters
of HP 9000 Systems .)
HP-P B
HP Precision Bus. Refers to the hardware I/O architecture of HP
9000 models such as Models 8x2 and 8x7. (See Chapter 10, "System
Architectures" . )
HP-P B adapter
See device adapter.

Glossary- 1 7
 HP-UX directory structure
The hierarchical grouping of directories and files on HP-UX. ( See Chapter
8, "HFS File System" and A Beginner's Guide to HP- UX.)

HP-UX system hostname

Name you assign to your system (from a line on I etc/re). It is used for
UUCP, mail, and other programs.
immediate response mode
A mode of data transfer in which data is read or written immediately to
mass storage, without buffering up the data for more efficient 1/0 . (See
Chapter 12, "HP- UX Peripherals" .)
initdefault
The default run-level for init to use during system startup. (See Chapter
2, "System Startup" .)
inode
A data structure containing information about a file such as file type,
pointers to data, owner, group , and protection information. (See Chapter 8,
"HFS File System" .)
Glossary
interface driver
On the Series 300/400/700, the type of device driver that controls the
kernel data structures concerning memory-mapped and processor 1/0 .
Internet address
An address used by NS_ARPA Services. It consists of two parts: a network
number and a host number. Nodes on the same LAN will have the same
network number and distinct host numbers. (See NS_ARPA Services
documentation.)
I/O channel separation
A capability, when using LVM , for segregating highly 1/0-intensive areas.
For example, you might have a database on one channel and file systems on
another. (See Chapter 9 , "Logical Volume Manager" .)
1/0 adapter
See device adapter.

Glossary- 1 8
1/0 mappings
The range of addresses the operating system uses to deal with I/O devices.
ITE
The Internal Terminal Emulator program which allows a bit-mapped
display to function as a standard computer terminal.
job
An individual instance of a command. Can also be two or more commands
piped together. (See Chapter 5, "Process Management" .)

job control
HP-UX commands and key sequences that allow users to manage jobs more
effectively. (See Chapter 5 , "Process Management" , A Beginner's Guide to
Using Shells , and Using HP- UX .)
kernel
The core of the HP-UX operating system. The kernel is the compiled code
responsible for managing the computer's resources , such as memory, file
system and input/output (1/0). The kernel resides in RAM (Random
Access Memory) whenever HP-UX is running. (See Chapter 1 1 , "System
Glossary
Configuration" .)
kernel stack segment
Virtual address space that contains the run-time stack when a process is
running in kernel mode. (See Chapter 7, "Memory Management" .)
LIF
Logical Interchange Format. LIF is Hewlett-Packard's standard file format ,
used for transferring files between Hewlett-Packard systems. Since LIF is a
standard, files with LIF format can easily be transferred between different
Hewlett-Packard computers (See Chapter 8, "HFS File System" and lif( 4)
in the HP- UX Reference.
logical volume
When using the Logical Volume Manager (LVM) , a logical volume is a
storage device, much like a disk section but of flexible size, that can hold
a file system (including root ) , raw data, application program, or swap .
Logical volumes can be mirrored using an optional product , HP LVM
Mirroring.

Glossary- 1 9
 Because its data is distributed logically (rather than physically ) , a single
logical volume can be mapped to one LVM disk or span multiple disks, but
the logical volume itself is used as a single virtual disk. ( See Chapter 9,
"Logical Volume Manager" . )
Logical Volume Manager (LVM}
An operating system software module that implements virtual (logical)
disks to extend, mirror, and improve the performance of physical disk
access. ( See Chapter 9, "Logical Volume Manager" . )
LVM disk
A disk that has been initialized for LVM; also termed a physical volume.
(See Chapter 9, "Logical Volume Manager" .)
limit
In disk quotas, a limit (or hard limit ) is the maximum amount of file space
allowed by a given user. ( See Chapter 14, "System Accounting" . )
link level address
A unique 12-digit hexadecimal number which is part of every LAN card .
This number appears on the LAN card hardware, on the boot ROM screen,
GI ossary
and can be obtained using the landiag program.
local login script
A . login script , contained in a user's home directory, which provides local
control over the user's working environment . (See Chapter 4, "Login" .)
local swapping
In an HP-UX cluster, cluster clients can have locally mounted disks to
which they can do swapping. This is known as local swapping. (See
Chapter 7, "Memory Management" and Managing Clusters of HP 9000
Systems .)
logical device manager (LDM}
On the Series 800 , the kind of device driver that translates requests between
the interface card and the external peripheral. For example, lprO.
login
The process by which user gains access to HP-UX. This process consists of

Glossary-20
 entering a valid user name and its associated password (if one exists). (See
Chapter 4, "Login" . )
lockable memory
Memory that can be locked into memory by processes via the plock and
shmctl system calls (described in section 2 of the HP- UX Reference). (See
Chapter 7, "Memory Management" .)
logical unit (lu)
The logical unit (lu) specification is created by insf ( lm) to allow one
device driver to handle many devices. The lu is mapped into a table that
the kernel then consults to determine which specific hardware address
the driver wants to access . The lu allows the system administrator
to keep track of all configured devices belonging to the same class of
device. ( See insf(lm) in the HP- UX Reference and Chapter 1 1 , "System
Configuration" . )
magneto-optical (MO)
Magneto-optical is a form of rewritable optical technology. (See Chapter 12,
"HP-UX Peripherals" .)
Glossary
major number
An index into a device driver table in the kernel. It is needed to
communicate with peripheral devices. (See Chapter 1 1 , "System
Configuration" .)
manufacturing automation protocol (MAP)
A communications standard enabling HP-UX to control factory-floor
equipment , including robotics. ( See Chapter 1 1 , "System Configuration" . )
mechanical changer
The part of the optical autochanger that moves optical disks from slot to
drive and vise versa. ( See Chapter 12, "HP-UX Peripherals" .)
mechanical picker
See mechanical changer.
media
In terms of optical products, this is the optical disk that holds the data.

Glossary-2 1
 The term includes the plastic cartridge that houses the optical disk. (See
Chapter 12, "HP-UX Peripherals" .)

memory management unit (MMU)

A hardware component that maps logical address to physical addresses
in the virtual memory system. Also protects against illegal accesses by
processes into each others' address space. (See Chapter 7, "Memory
Management" .)
minfree
See free space threshold.
minor number
Part of a device file; a hexadecimal number made up of driver-specific
information. The minor number often contains device addressing
information such as SCSI address, multiplexer port number, or select code
specific to the peripheral device you are setting up . (See Chapter 1 1 ,
"System Configuration" .)
mirroring
Simultaneous replication of data using an optional product , HP
Glossary
MirrorDisk/UX. This capability ensures a greater degree of data
availability. Mirroring maps logical volumes to multiple LVM disks , thus
providing the means to recover easily from the loss of one copy (or two
copies in the case of three-way mirroring) of data. Mirroring can provide
faster access to data for database applications using more data reads than
writes . (See Chapter 9 , "Logical Volume Manager" .)
monthly total command summary
A report produced by system accounting summarizing command activity for
the month. (See Chapter 14, "System Accounting" .)
multiprocessing (MP )
Computer processing using two, three, or four system processing units
(SPUs), with one source of memory shared among the processors and
peripherals. Multiprocessing is a limited form of parallel processing. ( See
Chapter 5, "Process Management" .)

Glossary-22
multi-user run-level
A run-level of HP-UX when terminals in addition to the system console
allow communication between the system and its users . The multi-user
run-level is run-level 2 as shipped. ( See Chapter 6, "Run-Levels" . )
MUX
MUX is an abbreviation for Asynchronous Multiplexer. Each channel is an
RS-232C port which is normally associated with a /dev/ttyXX file.
named object
Objects which have names , are visible at the TCB interface, and are shared
among users.
NCSC
National Computer Security Center, the government agency that wrote the
guidelines for trusted systems .
node
A computer on a network.
NS _ARPA Services
A combined networking product providing both NS and ARPA services . Glossary
The NS_ARPA Services networking product enables your Series 300 or 800
to transfer files to and from remote hosts, log into remote hosts, execute
commands on remote hosts, and send mail to and receive mail from remote
hosts on the network. ( See NS _ARPA Services documentation. )
NS nodename
Needed for Network Services and the rlb daemon. A name consisting of
three fields separated by periods, i.e. node.domain.organization. Each field
can contain up to 16 alphanumeric, case insensitive characters.
objects
Passive entities that contain or receive information. Access to an object
potentially implies access to the information it contains. Examples of
objects are: records, blocks, pages, segments, files, directories, directory
trees, and programs, as well as bits, bytes , words, fields, processors , video
displays, keyboards, clocks, printers, networks nodes, etc.

Glossary-23
 optical autochanger
A rewritable optical mass storage peripheral which includes the mechanics
to move optical disks in and out of drive(s), the drive(s) , media and
controller electronics. (See Chapter 12, "HP-UX Peripherals" .)
orphaned process group
A process group in which the parent process of every member is either itself
a member of the group or is not a member of the group 's session. (See
Chapter 5, "Process Management" .)
page fault
When the process encounters an address for which no virtual-to-physical
translation currently exists on the system, the system issues a page
fault . The kernel then switches execution to kernel mode to locate the
sought-after virtual address. ( See Chapter 7, "Memory Management" .)
page
The smallest contiguous block of physical memory that can be allocated for
storing data and processes. On all HP-UX systems, a page is 4 KB in size.
(See Chapter 7, "Memory Management" .)
Glossary .
pagmg
The means by which the memory-management subsystem moves pages of
data between RAM and mass storage for execution as a process demands
them. (See Chapter 7, "Memory Management" .)
parent process
The parent of a process-that is, the process that spawned a process . ( See
Chapter 5, "Process Management" .)
parent process ID
The process ID of a process's parent . (See Chapter 5, "Process
Management" .)
password
A private character string that is used to authenticate a username to
HP- UX. (See Chapter 4, "Login" .)

Glossary-24
path name
A series of directory names separated by / characters, and ending in a
directory name or a file name. (See A Beginner 's Guide to HP- UX.)
peripheral device
Any input/output device that can be connected to your system. Disk
drives, tape drives, printers, plotters, and terminals are all examples of
peripheral devices . ( See Chapter 1 1 , "System Configuration" , Chapter 12,
"HP-UX Peripherals" and Installing Peripherals .)
peripheral location
A peripheral's hardware address (minor number) . (See Chapter 1 1 , "System
Configuration" .)
per-process region (pregion)
A logical segment that points to specific segments of a process, including
text (process instructions) , data, u_area and kernel stack, user stack, one or
more shared-memory segments, and shared-library text and data segments.
Pregions hold page protections and the number of pages mapped to each
segment. (See Chapter 7, "Memory Management" .)
Glossary
phantom record
A record that crosses the EOT mark on a magnetic tape. (See Chapter 12,
"HP-UX Peripherals" .)
physical memory
The actual hardware memory components contained within your computer
system. ( See Chapter 7, "Memory Management" .)
physical volume
A disk that has been initialized by LVM for use in a volume group ; an LVM
disk. (See Chapter 9, "Logical Volume Manager" .)
physical volume group
When using LVM , physical volumes (LVM disks) can be further organized
within a volume group into a physical volume group , on the basis of the 1/0
channel or interface adapter to which they are connected, to achieve higher
availability of mirrored data. (See Chapter 9, "Logical Volume Manager" .)

Glossary-25
 PID
Same as process ID .
port
The place of access through which 1/0 comes into contact with the CPU.
PPID
S ame as parent process ID .
primary storage
Typically consists of fixed hard disk( s ) , used for fast , random-access
applications. The primary storage devices are used as online system disks.
(See Chapter 8, "HFS File System" .)
primary swap space
A contiguous area of a hard disk reserved for use by the demand-paged
virtual memory system for swapping. The size and location of primary
swap remain fixed in size and location, but can be changed either by
reconfiguring the kernel or dynamically, while the system is running. ( See
device swap space and Chapter 7, "Memory Management" .)
GlossarJriority
Ranking of processes by relative importance; used by the scheduler in
determining which process the CPU executes next . (See Chapter 5,
"Process Management" .)
process
A process is the environment in which a program (or command) executes .
It includes the program's code, data, status of open files , value of variables,
and more. For example, whenever you execute an HP- UX command, you
are creating a process; whenever you log in, you create a process. ( See
Chapter 5, "Process Management" .)
process group
When a user spawns a job , HP-UX assigns all processes in the job to a
process group . Signals can propagate through all processes in the group .
Each process group has a process group ID and a process group leader.
(See Chapter 5 , "Process Management" .)

Glossary-26
process group ID
An integer number which identifies a process group . It is the same as
the process ID of the process group leader. ( See Chapter 5, "Process
Management" . )
process group leader
Assigned by HP-UX when a process group is created. The PID of the
process group leader is the same as the process group ID of the process
group . ( See Chapter 5, "Process Management" . )
process group lifetime
A period of time beginning when a process group is created and ending
when the last remaining process in a group leaves, caused by either:
• The proces lifetime ends , or
• The process calling the sets id or setpgid system calls ( defined in
section 2 of the HP- UX Reference.
( See Chapter 5 , "Process Management" . )
process ID
An integer, assigned to a process at creation, which uniquely identifies the Glossary
process to HP-UX. That is, no two existing processes can have the same
process ID . ( See Chapter 5, "Process Management" . )
process lifetime
After a process is created with a fork ( ) function, it is considered active.
Its thread of control and address space exists until it terminates. It then
enters an inactive state where certain resources may be returned to the
system, although some resources , such as the process ID , are still in use.
When another process executes a wait ( ) or waitpid ( ) function for an
inactive process, the remaining resources are returned to the system. The
last resource to be returned to the system is the process ID . At this time,
the lifetime of the process ends.
pseudo-swap reservation
On Series 800 , a variation on traditional swapping that allows users to
execute processes in memory without allocating physical swap . ( See
Chapter 7, "Memory Management" . )

Glossary-27
 quorum
When using LYM, quorum is the requirement that a volume group have
at least half the configured LVM disks present to change or activate that
volume group . If there is no quorum, LYM prevents the change. Quorum
is also required to maintain Mirror Write Cache. (See Chapter 9, "Logical
Volume Manager" .)
quota
In disk quotas, a quota (also termed soft limit) is the amount of file space
allocated for a given user, and can be exceeded temporarily. (See Chapter
14, "System Accounting" .)
random access memory (RAM)
Memory cards that plug into the computer's backplane. For the CPU to
execute a process, the entire process must exist in RAM. ( See Chapter 7,
"Memory Management" and Chapter 11, "System Configuration" .)
raw mode
Unbuffered 1/0 . Data is transferred directly between the device and the
user program requesting the I / 0, rather than going through the file system
Glossary buffer cache. Also known as character mode. Compare with block mode.
(See Chapter 1 1 , "System Configuration" .)
re command
This is the system initialization shell script /etc/re. The actions that
it performs depend on the state in which it is invoked. ( See Chapter 2,
"System Startup" .)
real user ID
An integer, assigned to a username at login from the /etc/passwd file,
which uniquely identifies the username to HP-UX.
real group ID .
An integer, assigned to a username at login from the I etc/passwd file,
which identifies what group the user is from. Group IDs are defined in the
file /etc/group.
reboot
Taking HP-UX from a running state, down to a stopped state, and back to
a running state. (See Chapter 3, "System Shutdown" .)

Glossary-28
record
A group of related data items, usually stored contiguously on a mass
storage medium. ( See Chapter 12, "HP-UX Peripherals" .)
region
Data structures unique to a file or chunk of memory being accessed.
Regions inform the process about where the data exists in physical memory.
( See Chapter 7, "Memory Management" .)
remote swapping
The default mode of swapping in an HP-UX cluster. That is, the swapping
is performed for a cluster node on the cluster server. (See Chapter 7,
"Memory Management" and Managing Clusters of HP 9000 Systems .)
rewritable optical
An optical disk technology which can be repeatedly written. ( See Chapter
12, "HP-UX Peripherals" .)
resource
Anything used or consumed while performing a function. The categories
of resources are: time, information, objects (information containers) , or
GI ossary
processors (the ability to use information) . Specific examples are CPU time;
terminal connect time; amount of directly-addressable memory; disk space;
number of 1/0 requests per minute.
read-only memory (ROM)
Memory that can be read from (for example, the boot ROM resides in such
memory) , but cannot be written to. It retains its state (memory) even after
power is shut off.
root
Root refers to the highest level directory (root directory or /).
root file system
The file system containing the HP-UX kernel. The kernel is loaded from the
root file system at system startup . ( See Chapter 2, "System Startup" .)
root logical volume
When using LVM, the root logical volume is the logical volume containing
the HP-UX kernel. (see Chapter 9, "Logical Volume Manager" . )

Glossary-29
 run-level
A system state in which a specific set of processes are run. Run-levels are
defined in /etc/inittab.
secondary loader
Code located in the Boot ROM which executes during system boot-up and
loads the HP-UX system kernel. The Boot ROM loads and passes control
to the secondary loader, which, in turn, loads and passes control to the file
/hp-ux (or backup kernel, if /SYSBCKUP is specified) .
secondary memory
The location (typically disks) where data is stored when not being used.
By comparison, main memory (RAM) stores computer data required for
program execution. Secondary memory (also called secondary data storage)
is commonly termed swap. In the course of executing a program, data and
instructions are swapped (that is, copied) to and from secondary memory.
secondary storage
The storage device(s) , typically tape drives, used to back up and archive
data stored on the syst em disks (primary storage) . Secondary storage is
Glossary also used to log transaction, interchange data and distribute software.
Secondary storage devices always use removable media.
section
On Series 800 only, HP-UX allows you to divide a disk into separate pieces
called sections. To the operating system, these pieces look like separate disk
drives , each of which can have its own file system, swap area, or boot area.
( See Chapter 8 , "HFS File System" .)
security policy
The set of laws , rules, and practices that regulate how an organization
manages, protects, and distributes sensitive information.
select code
On a Series 300 / 400, the part of an address used for devices; a number
determined by switch settings on the interface card. The select code
determines the interface card's location in the processor address space.
Each interface card is in turn connected to a peripheral. Multiple
peripherals connected to the same interface card share the same select code.
( See Chapter 1 1 , "System Configuration" .)

Glossary-30
semaphore
A locking mechanism for ensuring that only one processor accesses critical
kernel resources at a time on a multiprocessor ( MP ) system. ( See Chapter
5, "Process Management" . )
sensitive information
Information that, as determined by a competent authority, must be
protected because its unauthorized disclosure, alteration, loss, or
destruction will at least cause perceivable damage to someone or something.
session
Typically, a session is the same as a login shell and all the jobs spawned
from the command line of that login shell. A session consists of one or more
process groups . ( See Chapter 5 , "Process Management" . )
session leader
Normally, the process that created the session ( i.e. , the login shell ) . ( See
Chapter 5, "Process Management" . )
session lifetime
The period between when a session is created and the end of the process
Glossary
group lifetime of all its process groups . ( See Chapter 5, "Process
Management" . )
set group ID bit
The middle bit of the most-significant octal digit in a file's protection mask.'
When set on a file that is an executable file, this bit causes the effective
group ID of the user invoking the command to be set equal to the group ID
of the file's group. ( See Chapter 8, "HFS File System" . )
set user ID bit
The most-significant bit of the most-significant octal digit in a file's
protection mask. When set on a file that is an executable file, this bit
causes the effective user ID of the user invoking the command to be
set equal to the user ID of the file's owner. ( See Chapter 8, "HFS File
System" . )
shared code
When an executable file is marked as having shared code, this means that
the file's code can be shared among the processes that run it. That is, each

Glossary-3 1
 process need not have its own copy of the code in its code segment ; instead,
there is one copy of the code, which all processes that run the program
share. ( See Chapter 7, "Memory Management" . )
shared library
Object files linked to a program the first time the library is c alled.
Subsequent calls to that library are mapped to the copy of the library
already existing in memory. ( See Chapter 7, "Memory Management" . )
shared library segment
Each shared library segment typically has three pr e gions text , initialized
-

data, and bss . The shared library text can expand like other text segments,
while the initialized data and bss cannot . ( See Chapter 7, "Memory
Management" . )
shared memory segment
Shared memory segments are data segments that can be shared among
processes . They are typically used when multiple processes must share data
( for example, in a windowing system, all window processes must be able to
update common data structures ) . Multiple shared memory segments can be
Glossary attached by many processes using the shmat(2) system call. ( see shmat(2)
in the HP- UX Reference and Chapter 7, "Memory Management" . )
shell
A program that interfaces between the user and the operating system.
HP-supported shells are:
• /bin/sh
• /bin/ csh
• /b in/ksh
• /b in/rsh
• /b in/rksh
• /b in/pam
shutdown
The process of taking the system from multi-user state to a single-user
state, or taking the system from a running state to a state in which no
processes are running.

Glossary-32
shutdown command
A shell script that has the primary function of terminating all currently
running processes in an orderly and cautious manner. ( See Chapter 3,
"System Shutdown" and shutdown(lM) in the HP- UX Reference.)
slot
The physical place in the card-cage of a computer where a card plugs in.
Each slot has a number, which if the slot is used for modules, is multiplied
by four to derive the module number (used in hardware addressing) .
( See Chapter 10, "System Architectures" and Chapter 1 1 , "System
Configuration" .)
SMB
System Main Bus, implemented on Models 850S , 855 S , 860S, 865S , and
870S . Memory and processor boards are plugged into the SMB, which
connects to the Mid-Bus for 1/0 via bus converter. ( See Chapter 10,
"System Architectures" . )
soft limit
In disk quotas, a soft limit (also termed quota) is the amount of file space
allocated for a given user, and can be exceeded temporarily. (See Chapter Glossary
14, "System Accounting" . )
source device
The mass storage device from which HP-UX is installed . The source device
must be a cartridge tape drive or flexible disk drive.
special file
Often called a device file, this is a file associated with an 1/0 device.
Special files are read and written just like ordinary files, but requests
to read or write result in activation of the associated device. These
files norm ally reside in the /dev directory. ( See Chapter 1 1 , "System
Configuration" . )
spinlock
A kind of semaphore used to lock kernel resources very briefly. ( See
Chapter 5, "Process Management" . )

Glossary-33
 stack segment
A memory segment reserved for use by a process's stack. On Series 300 , it
is attached near the top of the logical address space and grows "down" in
memory. On Series 800 , the stack and user data area share one space. ( See
Chapter 7, "Memory Management" . )

standalone
A machine which is not part of an HP-UX cluster.
standard error
By default, all command error messages are displayed to a file known
as standard error. Also, by default , standard error is displayed on the
controlling termin al. ( See Chapter 5, "Process Management" . )
standard input
By default , all command input is taken from a file known as standard input .
Also, by default , standard input is input ( typed ) at the keyboard of the
controlling terminal. ( See Chapter 5, "Process Management" . )
standard output
By default , command output is displayed to a file known as standard
Glossary
output . Also, by default , standard output is displayed on the controlling
terminal. ( See Chapter 5, "Process Management" . )
sticky bit
This is one of the bits on a file's protection mask. It can be set only by
the system administrator via the cbmod command. If set on an executable
program, the program will continue to reside in the swap area until
the system administrator unsets the sticky bit . This can improve the
performance of programs that are often used since they can be loaded
from the swap device, not form the file system. ( See Chapter 8, "HFS File
System" . )
streaming
A design feature which allows a tape drive to move in long continuous
motions rather than jerky motions. This is done by buffering data in
such a way that it can be written smoothly. ( See Chapter 12, "HP-UX
Peripherals" . )

Glossary-34
subfile
Part of a CDF in an HP-UX cluster. The subfiles are under the hidden
directory, and are named for one of the system's context attributes. ( See
Managing Clusters of HP 9000 Systems.)
subject
An active entity, generally in the form of a person, process, or device that
causes information to flow among objects or changes the system state.
Technically a process/domain pair.
superblock
A data structure containing global information about the file system such
as file system size, disk information, and cylinder group parameters. The
superblock is created at the same time as the file system and is replicated
into each cylinder group. Also, HP-UX keeps a copy of the superblock in
memory at all times. The sync command writes the superblock to disk.
(See Chapter 8, "HFS File System" in this manual and sync(lM) in the
HP- UX Reference.)
superuser
The root user who has special privileges. Same as the system administrato:o1ossary

surface
In terms of optical disks, this is one of the disk sides-surface 1 or 2.
swap
Secondary memory (secondary data storage) is commonly termed swap .
In the course of executing a program, data and instructions are swapped
(that is, copied) to and from secondary memory. (See Chapter 7, "Memory
Management" .)
swapping
The means by which entire processes are moved for execution between
RAM and mass storage (typically disk) . (See Chapter 7, "Memory
Management" .)
swap space
Disk space which is reserved for use by the demand-paged virtual memory
system. Processes are swapped to and from swap space. There are two

Glossary-35
 kinds of swap space: primary swap space and dynamic swap space. ( See
Chapter 7, "Memory Management" and Chapter 8 , "HFS File System" . )
synchronization
When using LVM, synchronization keeps mirrored logical volumes
consistent by ensuring that all copies contain the same data. ( See Chapter
9, "Logical Volume Manager" . )
system administrator run-level
A run-level of HP-UX when the system console provides the only
communication mechanism between the system and its users. lnit state s is
the system administrator run-level. ( See Chapter 6, "Run-Levels" . )
system CDFs
System file ( e.g., /etc/re, /etc/ inittab) that are also CDFs
( context-dependent files ) . ( See Managing Clusters of HP 9000 Systems .)
system console
A keyboard and display ( or terminal ) given a unique status by HP-UX
and associated with the special device file /dev/console. All boot ROM
error messages ( messages sent prior to loading HP-UX ) , HP-UX system
Glossary
error messages, and certain system status messages are sent to the system
console. Under certain conditions (for example, the single-user state ) ,
the system console provides the only mechanism for communicating with
HP-UX. ( See Chapter 2, "System Startup" . )
system login script
A system-wide login script, run for every user of a particular shell when he
or she logs in. System login scripts are:
• /etc/profile for Bourne and Korn shell users
• /etc/csh . login for C shell users.
system shutdown
The process of taking HP-UX from a running state to one in which no
processes are running. See also shutdown and reboot . ( See Chapter 3,
"System Shutdown" . )

Glossary-36
system startup
The process of taking HP-UX from a stopped state ( in which no processes
are running ) to a state in which it is ready to accept input from users. ( See
Chapter 2, "System Startup" . )
tape density
A measure of how densely information is stored on a magnetic tape. Bits
per inch is a common measure of tape density. ( See Chapter 12, "HP-UX
Peripherals" . )
TCB
See trusted computing base.
TC SEC
Trusted Computer Systems Evaluation Criteria, also known as the "Orange
Book" . This is the book where evaluation criteria for trusted systems is
documented.
text segment
See code segment .
thrashing Glossary
When the demand-paged virtual memory system spends an inordinate
amount of time paging ( swapping processes to and from swap space ) , so
much so that system performance degrades. ( See Chapter 7, "Memory
Management" . )
time-slice
The amount of time a process can run before the kernel checks to see if
there is an equal-priority process ready to run.
track
One of several concentric circles on the surface of a disk upon which data is
recorded ( refer to Chapter 8, "HFS File System" . )
translation lookaside buffer (TLB)
A page table that matches cache or indexes physical addresses ( depending
on design ) for recently accessed virtual addresses. Once an address is
translated, it can be used by the cache, which uses physical addresses. ( See
Chapter 7, "Memory Management" . )

Glossary-37
 trap door
A hidden software or hardware mechanism that permits system protection
mechanisms to be circumvented. It is activated in some non-apparent
manner (e.g., special "random" key sequences) .
Trojan horse
A computer program with an apparently or actually useful function
that contains additional (hidden) functions that surreptitiously exploit
the legitimate authorizations of the invoking process to the detriment of
security-for example, making a "blind" copy of a sensitive file for the
creator of the Trojan Horse.

trusted computer system

A system that employs sufficient hardware and software integrity measures
to allow its use for processing simultaneously a range of sensitive or
classified information.
trusted computing base
(TCB ) The totality of protection mechanisms within a computer system­
including hardware, firmware, and software-the combination of which is
Glossary responsible for enforcing a security policy. A TCB consists of one or more
components that together enforce a unified security policy over a product
or system. The ability of a TCB to correctly enforce a security policy
depends solely on the mechanisms within the TCB and on the correct input
by system administrative personnel of parameters (e.g. , a user's clearance)
related to the security policy.
trusted process
A process that is restricted and is only run by a user with appropriate
privilege. For HP-UX 6.5 a trusted process is a process with effective user
ID of 0 (superuser) .
trusted software
The software portion of a Trusted Computing Base.
u_area
A data structure containing process characteristics used by the kernel when
a process is executing. When a process is swapped out , its u_area is also
swapped out . ( See Chapter 7, "Memory Management" .)

Glossary-38
unattended mode
The default method by which the boot ROM finds and loads an operating
system. Occurs if the user does not intercede during system startup . (See
Chapter 2, "System Startup" . )
unit number
Part of an address used for devices; a number whose meaning is software­
and device-dependent but which is often used to specify a particular
disk drive in a device with a multi-drive controller. When referring to
single-controller integrated disk/tape or disk/flexible disk drive, a unit is
used to distinguish between disk and cartridge tape drives or hard disk and
:flexible disk drive. (See Chapter 1 1 , "System Configuration" .)
(The unit number also selects a single partition on the 913x series.)

use count
The demand-paged virtual memory system keeps track of the number of
processes accessing shared code. This number is called the use count . (See
Chapter 7, "Memory Management" .)
user
Any person who interacts directly with a computer system. G lossary

user stack segment

Contains the run-time stack when a process is running in user mode. (See
Chapter 7, "Memory Management" . )
UUCP
Unix-to-Unix File Copy Package, mainly used for mail transport .
verification
In a trusted system, the process of comparing two levels of system
specification for proper correspondence (e.g., security policy model with
top-level specification, TLS with source code, or source code with object
code). This process may or may not be automated.
virtual address space
The range of memory locations that the individual process can access,
within the physical limits of swap space and main memory. (See Chapter 7,
"Memory Management" .)

G lossary-39
 vnode
A region is usually filled by a vnode, an interface for reading and writing
pages of data between memory and disk. Vnodes are categorized by
file type, such as uf s , nfs , dux, and cdf . (See Chapter 7, "Memory
Management" .)
volume group
In LVM , the combined disk space of one or more LVM disks (physical
volumes) from which logical volumes space are allocated. (See Chapter 9,
"Logical Volume Manager" .)
volume number
Part of an address used for devices; a number whose meaning is software­
and device-dependent but which is often used to specify a particular volume
on a multi-volume disk drive. The volume number is also used to inform
the device driver of special handling semantics (such as printer drivers
skipping over perforations). ( See Chapter 1 1 , "System Configuration" .)
word
A unit of information, typically consisting of four consecutive bytes (32
Glossary bits) .

write protected
Indicates that a storage media is protected from being written to. This is
useful in situations where you have a tape or diskette that you don't want
written to and possibly destroyed. ( See Chapter 12, "HP-UX Peripherals" .)
write ring
A mechanism on a reel-to-reel tape that allows you to write-protect the
tape.

Glossary-40
 Index

I ndex

8 terminals , 12-25
802.4 link, 13-3 addressing, 1 1-21
bus address, 1 0-8 , 10-10
9 definition, Glossary-1
9-track magnetic tape device drivers, 1 1-21
device file format , 1 1-47 format , 1 0-19
hardware path, 1 0-19
A hardware paths , 1 0-41
HP-IB , 12-28
access, Glossary- 1
access control list ( ACL ) , 8-59 , 8-63 HP-PB models , 1 0-2 1
access control mechanism ( ACM ) , logical unit ( lu) specification, 1 1-43
MAP cards, 1 0-22
Glossary- 1
accessing raw data using LVM , 9-9 Mid-Bus, 1 0-27
MUX card, 1 0-10
access permission, 8-58, 8-60
accounting directories, 14-14 select code , 1 0-8
acct , 14-25 Series 300 hardware paths , 10-8
acct ( 4) , 14-9 Series 700 , 1 0-14
acctcms report options , 14-45
Series 800 , 1 0-19
SMB models , 1 0-30
acct cms , to generate command report '
14-42 advisory lock
acctcom, 14-35 definition , Glossary-1
acctcon1 , 14-28 advisory locks, 8-64
acctcon2 , 14-27, 14-30
AFI ( Asynchronous FIFO Interface )
acctdisk, 14-21
management , 1 1-24
acctdusg, 14-17, 14-18
aging, password, 4-9
acctdusg and diskusg compared, 14- 1 9 allo cation policy, 9-47
acctmerg , 14-49 , 14-51
defined, 9-5 , Glossary-2
acctprc1 and acctprc2 , 14-46 alternate superblock, 8-28, 8-50
a . out , 7-28
acct sh, 14-27, 14-53 , 14-61
acctwtmp , 14- 1 1 , 14-24
application layer, 13-2
adding architecture, 10- 1
modems, 1 2-25 CIO , 10-15

lndex- 1
Index

HP-PB , 10-15 backplane and minor number, 1 1-34

Series 300, 10-2 backup kernel, 8-2 1
Series 400 , 10-7 backups
Series 700, 10- 1 1 with mirroring, 9-49
Series 800, 10-15 bad block relocation, Glossary-3
architecture of LVM subsystem, 9-5 1 defined, 9-5
archived libraries , 7-34 bad block relocation policy, 9-35
definition , Glossary-2 bad blocks in inodes , 8-53
vs. shared libraries, 7-38 baud rate, 4- 18, 12-25
ARPA/Berkeley Services , 13-4 bcheckrc , 2-23
ARPA hostname, Glossary-2 bdf
ARPA Networking Services , 8-55 explained, 8- 18
ARPA Services , 13-7 used in accounting, 14-22
as , 7-28 behavioral information and mmor
assigning swap priorities, 7-48 number, 1 1-34
asynchronous vs . synchronous disk Berkeley Networking Services, 8-55
writes , 8-46 BFS file system, 8-55
attended mode bif , 8-58
defined, 2-6 /bin, 4- 16
definition, Glossary-2 binding, Glossary-3
on Series 300/400, 2-9-10 bit , Glossary-3
on Series 700 , 2- 14 bit-mapped display , 2-4
attribute , Glossary-2 bits per inch, Glossary-4
audit ID , 5-5, Glossary-3 bits per inch (bpi) , 12-2
audit trail, Glossary-2 block, Glossary-4
autoboot defined, 9-5
on Series 700 , 2- 13 block device, 1 1-28 , Glossary-4
on Series 800 , 2- 17 file, 8-30 , 1 1-28
autochanger management , 1 1-24 block I/O , 1 1-30
autocreation, Glossary-3 block major number, 1 1-3 1
autosearch block mode, 1 1-27, Glossary-4
on Series 700, 2- 13 block special files, LVM , 9-9
on Series 800, 2- 17, 2-1 9 block vs. character device files, LVM ,
autoxO , 1 1-24 9-9
available memory, 7-3 , 7-5 , Glossary-3 boot
size , 7-3 boot entry in inittab , 2-2 1
bootwait entry in inittab , 2-2 1
B fsck run during bootup , 2-23
background processes, 7-25 loading the operating system, 2-7
background process group , 5-8 , manual boot on Series 800, 2- 18
Glossary-3 primary path, 2- 16

lndex-2
 Index

processes, 2-22 Bourne (/bin/ sh) , 4-5

bootable physical volume layout , 9-29 Bourne shell, 4-30
bootable physical volume ( LYM disk ) , bpi, Glossary-5
9- 18 brc , 2-21 , 2-24
Boot Administration Mode ( Series 700 ) , BSD file system, 8- 1
2- 15 bss, 7-15
boot area, Glossary-4 buffer cache, 8-39 , 8-43
Series 300 /400 implementation, 8-9, dynamic implementation ( Series
8-20 300 /400 / 700 ) , 8-42
Series 700 implementation, 8- 10, 8-22 buffers , 7-2
Series 800 implementation, 8-1 1 , 8-22 bufpages , 8-40
boot area , 8-20 bus , Glossary-5
Boot Console User Interface ( Series defined, 10-2
700 ) , 2- 14 bus address, Glossary-5
Boot Data Reserved Area ( BDRA) , bus converter , 1 0-17, Glossary-5
9-30 byte , Glossary-6
booting, 7-3 byte offset , 7-14, Glossary-6
Boot Mode Selection ( Series 400 ) , 2-9 bytes per inode , Glossary-6
boot or boot-up , Glossary-4
boot process, Glossary-5 c
boot processes, 6-4 cache , 7- 10, Glossary-6
boot ROM , Glossary-5 cache, mirror write consistency, 9-48
Boot ROM cartridge tape, 1 1-27
attended mode on Series 300 /400 , description, 12- 1 2
2-9-10 immediate response mode, 12-12
hardware tests , 2-6 cartridge tape drive
I / 0 configuration, 2-1 1 device file format , 1 1-49
loading HP-UX, 2-6 CDF , Glossary-7
operating system search sequence , CD-ROM File System ( CDFS )
2-7 description, 1 2-13
revisions, 2-8 implementation , 1 2-13-17
search criteria, 2-8 centO , 1 1-24
startup sequence overview, 2-2 central processing unit ( CPU ) , Glossary-6
startup sequence ( Series 300 /400 ) , centronics ( parallel ) interface, 10-5
2-4 certification , Glossary-6
startup sequence ( Series 700 ) , 2-13 changing process accounting files, 14-35
startup sequence ( Series 800 ) , 2- 16 channel adapter, 10-17
system console search sequence, 2-5 channel adapter manager, 1 1-24
unattended mode, 2-6 channel adapter manager ( CAM ) , 1 1-20 ,
boot-up Glossary-7
on MP systems, 5-26 Channel Input / Output ( CIO ) , 10-15

lndex-3
Index

Channel Input/Output (CIO) (Series company holiday lines, Glossary-8

800) , 10-16 compiled object code ( a . out ) , 7-28
channel 1/0 adapter, Glossary-7 conf ig, 1 1-22
character device, 1 1-28 , Glossary-7 configurable parameters, 1 1-16
file, 8-30 , 1 1-28 configuration
character 1/0 , 1 1-30 configuration file, 1 1-13
character maj or number, 1 1-31 defined, 1 1-1
character mode , 1 1-27 init , 1 1-17
characters per inch ( cpi) , 12-2 kernel drivers, 1 1-25
chargefe e , 14-6, 14-8 machine-dependent initialization,
charging fees, 14-48 1 1-17
chatr, 7-28, 7-33 Series 800 auto-configuration, 1 1-18
checklist , 7-46 swap space, 7-52
chgrp, 5-7 terminal, 12-25
CIO architecture, 10-15 configuration control menu (Series 400) ,
cio_caO , 1 1-22, 1 1-24 2-9
ckpacct , 14-8 configuration, kernel, 1 1-17
ckpacct , to check size of file , 14-34 configuration maintenance, LVM , 9-37
clean byte, 2-23 , Glossary-7 configuring, Glossary-8
cluster, Glossary-7 configuring device drivers, 1 1-23
cluster client , Glossary-8 connectivity checks by fsck , 8-55
cluster node, Glossary-8 connect session, Glossary-8
cluster server, Glossary-8 connect session accounting, 14-23 , 14-24
cluster server process (CSP), Glossary-8 console , 1 1-14
code segment , 7- 1 5 , 7-44, Glossary-8 console management , 1 1-24
command context , Glossary-9
fsck, 8-28, 8-44, 8-49 context-dependent file, Glossary-9
f sdb, 8-28, 8-55 contiguous vs. non-contiguous LVM
commands disk space, 9-18
chgrp , 5-7 continuation inode , 8-51
iostat , 5-1 6 controlling process, 5-8 , Glossary-9
kill , 5-14 controlling terminal, 5-8 , Glossary-9
monitor , 5- 16 copy-on-access , 5- 10
nice, 5- 15 copy-on-write, 5-10, 7-24, Glossary-9
ps, 5- 13 core input/output function number ,
renic e , 5- 15 1 1-38
sar, 5- 16 correcting file system corruption, 8-48
top, 5- 16 cpio, 8-58
vmstat , 5- 16 CPU use, 5- 1 8
comment lines, Glossary-8 CRC , 12-5
communication protocol, 2-28 creating a file system, 8-2

lndex-4
 I ndex

creating system run-levels, 6-5 demand-loaded code, 7-28 , 7-33, 7-36

cron, 14-8 demand loading, 7-3 1
used for system accounting, 14-53 DEMAND_MAGIC, 7-3 1
cron, Glossary-10 demand-paged virtual memory,
crt O , 7-36 Glossary- 1 1
CS/80, Glossary- 10 defined , 7-1
csBO kernel driver, 1 1-23 origin, 7-9
C shell (/bin/csh) , 4-5 demand paging, 7-24, Glossary- 1 1
ct , 1 1-27 desfree , 7-26
customization file, 2-25 destination device, Glossary- 1 1
cyclic redundancy check, 12-5 detecting file system corruption, 8-48
cylinder , Glossary- 10 determining amount of swap space, 7-5 1
cylinder group , 8-9, 8-26 , 8-30, 8-38, /dev , 1 1-27
Glossary- 10 DEV_BSIZE, 9-26
cylinder group information, 8-28, 8-44, /dev/conf ig , 1 1-25
Glossary- 10 device adapter, 10-17, Glossary- 1 1
device adapter manager (DAM) , 1 1-20 ,
D Glossary-12
daemon , 7-25, Glossary- 10 device driver, 8- 1 , 1 1-20 , 1 1-22,
daemons startup , 2-25 Glossary-12
daily command summary, Glossary- 10 maj or number, 1 1-31
daily line usage report, Glossary-10 device drivers, 1 1-20
daily resource usage report , Glossary-1 1 addressing, 1 1-21
DASS, Glossary- 1 1 and major numbers, 1 1-31
databases, 9-45 configuration, 1 1-23
data block count , 8-54 device file, 1 1-26
data blocks, 8-28, 8-30, 8-44 block, 8-30, 1 1-28
data-link layer, 13-2 block vs. character, 1 1-28
data segment , 7- 15, 7-44, Glossary- 1 1 character, 8-30 , 1 1-28
data storage, 7- 1 , 8-35 description, 1 1-26
DDS-format tape LVM block vs . character, 9-9
compared to nine-track magnetic tape, naming convention, 1 1-27
12-9 network, 1 1-28
defined, 12-9 path name , 1 1-27
organization, 12-9 device file directories, 1 1-27
resources , 12- 1 1 device files, 1 1-26
tar , 12-10 9-track magnetic tape drive, 1 1-47
DDS-format tape drive, 1 1-23 cartridge tape drive, 1 1-49
device file format , 1 1-48 DDS-format tape drive , 1 1-48
default local login script, 4-24 disks, 1 1-45
demand loadable, Glossary- 1 1 HP-IB devices, 1 1-49

lndex-5
Index

QIC-format tape drive , 1 1-48 inetd, 14-70

required for file systems, 8-5 overview and planning, 14-62
role of, 1 1-26 quota, 14-69
device manager (DM) , 1 1-20 , Glossary-12 quotact l , 14-66
device mapping, 7-40 quotaoff , 14-66
devices quotaon , 14-66
in clusters , 1 1-27 quota output , 14-68
devices , character or raw, 1 1-28 quotas and limits defined , 14-63
devices , drivers, 1 1-14 quot check , 14-67
device swap space, 7-42, 7-43, Glossary- 12 remote enforcement , 14-70
/dev/kmem, 1 1-25 repquota, 14-70
/dev/null , 1 1-25 rquotad, 14-70
df ile , 1 1-3, 1 1-4 , 1 1-13, 1 1-16, 1 1-17 setting time limits , 14-68
lists configured device drivers, 8-4 summarizing disk usage, 14-70
DIO bus, 10-2 user perspective, 14-65
DI0-11 bus, 10-2 user warnings, 14-68
implementation , 10-3 disks
Direct Access Secondary Storage , 12-19 device filenames , 1 1-45
directories disk sectioning, Glossary- 13
setting permission bits, 8-60 disk sections , Series 800, 8- 12
directory structure, Glossary- 18 disk space allocation, 8-37
dis c1 , 1 1-24 policies, 8-38
disc2 , 1 1-24 using LVM , 9-24
disc3 , 1 1-24 disk space usage accounting, 14-17
Discretionary Access Control (DAC), disk spanning
Glossary-12 introduced , 9-1
disk, Glossary- 13 disk special files , 1 1-27
disk block descriptor ( dbd ) , 7-21 disk usage reported with du, 8- 19
disk-drive manager, 1 1-24 diskusg, 14-17, 14-18
disk layout , 8-8 disk writes
Series 300/400, 8-9 f s_async , 8-46
Series 700 , 8- 10 synchronous vs. asynchronous, 8-46
Series 800, 8- 1 1 displayO , 1 1-24
diskless workstation, Glossary-13 displaying disk-quota status , 14-69
disk quotas , Glossary- 13 dmesg, 2- 1 , 10-38
and /etc/checklist , 14-66 , 14-67 dodisk , 14-21
commands , 14-66 domain, Glossary- 13
displaying status, 14-69 DOMAIN operating system (Series 400) ,
edquota, 14-67 2-9
file-system guidelines, 14-62 driver, Glossary- 13
how disk quotas work , 14-63 driver number, Glossary- 13

lndex-6
 Index

drivers,configuring, 1 1-25 /et c/brc , 2-21 , 2-24

dsk, 1 1-27 /etc/btmp , 4-3
dskles s_f sbuf s , 8-40 used by login, 4- 12
du /etc/checklist , 7-46
explained, 8-19 and swap, 7-47
used in accounting, 14-22 lists mountable file systems, 8-8
dumps , 1 1-14 /etc/csh . login script, 4-22
duplicate blocks, 8-52 /et c/diskinfo
dynamic loader ( /lib/dld . sl ) output explained, 8- 16
as a shared library, 7-34 used when managing disk space, 8-16
how it works , 7-36 used with LVM , 8- 16
dynamic swap space , Glossary- 1 3 /etc/disktab , 7-5 1
allocation, 7-9 explained, 8- 14
file-system swap , 7-45 purpose, 8-6
use as disk and file-system tool, 8- 14
E used by newfs , 8- 15
edquota, 14-64, 14-67 /et c/d . login , 4-26
setting time limits, 14-68 /et c/dmesg, 2- 1 , 10-8, 10-38
EEPROM, 2- 1 1 /et c/getty, 2-27, 12-26
effective group ID, 5-4 , Glossary- 14 /et c/gettydefs , 12-27
effective user ID, 5-4, Glossary- 14 /etc/init , 2-20, 6-2
EISA bus, 10-13 /etc/inittab , 2-21-28, 6-2
EISA bus adapter (Series 700) , 10-13 action , 2-2 1
eisa_config, 2-23 , 10-13, 1 1-23 boot , 2-21
end of tape (EOT) mark, 12-3 bootwait , 2-2 1
enforcement locking mode, 8-6 1 , 8-65 default run levels, 2-22
enforcement mode, Glossary- 14 /et c/bcheckrc , 2-23
ENOMEM , 7-44 /et c/brc, 2-24
ENOMEM, 7-16, 7-44 /et c/recoversl , 2-24
environment file initdefault , 2-2 1
/etc/inittab, 6-2, 6-5 process, 2-21
environment variables respawn, 2-22
described, 4- 13 run-levels , 2-2 1
LOGNAME , 4-8 syntax, 2-2 1
TERM, 4- 16 sysinit , 2-2 1
EOT, 12-3 /etc/ioconf ig, 1 1-17
erase key /etc/ioinit , 2-22
setting, 4-7, 4- 18 /et c/ios can, 10-38 , 10-41
/et c , 4-16 /et c/lvmpvg , 9-45
/et c/bcheckrc , 2-2 1 , 2-23 /et c/lvmrc, 2-23
runs fsck , 8-48 / etc / lvmtab , 9-37

lndex-7
Index

/ etc/master, 1 1-20, 1 1-22 file access from inode to data blocks ,

/et c/mirrorrc , 2-23 8-32
/et c/pas swd file locking, Glossary- 14
syntax, 4-8 file mark, 12-3, Glossary- 14
/et c/powerfail , 2-27 file name conventions, device, 1 1-27
/et c/pre_init_rc, 2-20 file owner, 8-58, 8-6 1
/et c/profile script, 4-20 files
/et c/re , 2-24, 2-25, 7-39 , 14-32 access control list ( ACL) , 8-63
use in accounting, 14- 10 advisory locks, 8-64
/et c/recoversl , 2-24 block and character device, 1 1-28
/et c/wtmp , 4-3 , 14- 1 1 device, 1 1-26
used b y login , 4- 12 f cntl , 8-64
I etc/wtmp and connect session format and compatibility, 8-55
accounting, 14-23 lockf , 8-64
exchange , Glossary- 14 locking, 8-6 1 , 8-64, 8-66
exec, 7-52 protection, 8-58
EXEC_MAGIC , 7-30 sharing, 8-64
executable code special, 1 1-26
handling , 7-28-34 transferring, 8-55
types, 7-28 file size inconsistencies, 8-53
execute permission, 8-58 , 8-60 file system, Glossary- 15
exit , 5- 12 adding to /etc/checklist , 8-3
Extended Industry Standard alternate superblock locations, 8-28
Architecture (E/ISA) bus , 10-13 associating file system with disk, 8-2
extendfs(lM) boot area, 8-20
device file for LVM , 9-9 buffer cache , 8-39, 8-43
extent checking, 2-23
defined, 9-5, Glossary- 14 connectivity checks by f s ck, 8-55
introduced, 9-2 corruption, 8-47, 8-48
extent size introduced, 9-2 creation, 8-2
external buses cylinder group, 8-26
defined, 10-2 cylinder group updated, 8-44
data blocks, 8-30
F data blocks released, 8-43
factory-floor communication, 10-28 data storage, 8-35
factory-floor devices , 13-7 debugger ( fsdb ) , 8-28
f cntl, 8-64 device files required, 8-5
fiber-link interface , 1 1-24 disk layout , 8-8
FIFO , 8-30 , 10-38 disk sections , 8- 12
file, Glossary- 14 displaying free blocks with bdf , 8- 18
du reports disk usage, 8- 19

l ndex-8
 Index

file cabinet analogy, 8-6 free space threshold, Glossary- 16

free space, 8-31 fs, 7-39
HFS, 8- 1 fs_async , 8-46
how files are modified, 8-43 fs_bsize, Glossary- 16
immediate reporting, 8-45 f s ck , 8-47
inode creation, 8-30 alternate superblock locations, 8-28 ,
inode error messages , 8-30 8-50
inodes written, 8-43 checks data block count , 8-54
layout , 8-24 checks file size inconsistencies, 8-53
logical volumes, 8- 13 checks file-system connectivity, 8-55
mapping inode to file data blocks, checks file system size, 8-50
8-32 checks free blocks, 8-50
permission bits, 8-58 checks inode consistency, 8-5 1
purpose of mounting and unmounting, checks inode count , 8-5 1
8-7 description, 8-49
relationship to disk, 8- 1 mounted file system, 8-44
reserving free space, 8-36 preen mode, 2-23
root , 2-20 , 2-23 run during HP-UX startup , 2-20
size , 8-50 use of /lost+found, 8-48
superblock, 8-25 use to detect , correct file-system
superblock written, 8-43 corruption, 8-48
two meanings, 8-2 f s clean
use of f s ck , 8-44, 8-48 run by /et c/bcheckrc, 2-23
use of fsdbas debugger, 8-55 fsdb
when not to unmount , 8-8 file-system debugger, 8-55
file-system guidelines, for disk quotas, FTAM, 13-3
14-62 FTAM/9000 , 13-7
file-system swap , 7-45 ft io , 8-55
file-system swap space, Glossary-15 function number, 1 1-38
file transfer fwtmp to display accounting records,
utilities and services, 8-55 14-25
file type , 1 1-22
file types, Glossary- 15 G
foreground process group , 5-8 , gateway page , 7-15, Glossary-16
Glossary- 16 gathering data, for system accounting,
format and type of inodes, 8-52 14-8
fragment , 8-31 , 8-34, Glossary- 16 general-purpose parallel interface , 1 1-24
frame, 12-2 , Glossary-16 generating accounting reports, 14-9
free block, 8-50 generating reports, for system accounting,
free blocks, 8-18 14-9
free space, 8-3 1 , 8-37 generic device file , Glossary- 16

lndex-9
Index

getty, 2-27, 1 2-26 homogeneous cluster, Glossary- 17

gpioO , 1 1-24 hostname, Glossary- 18
gpio1 , 1 1-24 how disk quotas work , 14-63
graphO , 1 1-24 HP Apollo 9000 Series 700 bus
graph2 , 1 1-24 architecture, 10-12
graphics , 7-40 HP-FL , 10-18
graphics interface, 10-6 hpflO , 1 1-24
graphics management , 1 1-24 HP FTAM / 9000 , 8-55
graphics tablet interface, 1 2-29 HP-HIL ( Human Interface Link ) , 10-5
group HP-IB , 10-18
access list , 5-7 limitation in volume group, 9-7
ID, 8-58, 8-6 1 not supported on HP MirrorDisk / UX,
group access list, Glossary- 16 9-40
group file hpibO , 1 1-24
minor number, 9- 1 1 hpib1 ' 1 1-24
groups HP-IB connectivity, 12-28
changing with chgrp , 5-7 HP-IB device management , 1 1-24
HP-IB devices
H file format , 1 1-49
halting, Glossary- 17 HP-IB guidelines, 1 2-28
halting the system, 3- 1 electrical, 12-29
hard disk drive HP-IB ( IEEE-488 ) interface, 10-5
system ( root ) , 12-29 HP-IB instrumentation management ,
hard limit , Glossary- 1 7 1 1-24
hardware address, 1 1-2 1 , 1 1-42 HP-IB limitations , 9- 12
hardware failure, 8-47 HP MirrorDisk / UX , 9-1 , 9-40-50
hardware modules defined, 9-6, Glossary-22
defined , 10-17 HP OSI Express MAP , 13-3
hardware path, 10-19, 1 1-18, Glossary-1 7 HP-PB
hardware path and minor number, 1 1-34 definition, Glossary- 17
hardware paths , 10-41 HP-PB adapter , Glossary- 17
hardwired , 12-27 HP-PB architecture, 10-15
Hewlett-Packard Precision Bus ( HP-PB ) HP-PB bus architecture, 10-20
( Series 800 ) , 10-16 /hp-ux , 8-2 1
HFS file system, 8-1 , 8-55, Glossary- 17 hpux, 2- 14, 1 1-16
hidden directory , Glossary- 17 HP-UX cluster
HOG FACTOR, 14-37 booting the cluster, 2-7
holidays file , 14-4 system accounting, 14-14, 14-18,
$HOME/bin, 4- 16 14-22
home directory, Glossary- 17 system CDF, 1 1-28
$HOME/ . login script, 4-32 HP-UX Startup Sequence, 2-20

lndex- 1 0
 Index

hshpib kernel driver, 1 1-23 interrupts, 5-17, 5-24

1/0 adapter, Glossary-18
1/0 channel separation, 9-43-46,
identify burst , 12-3 Glossary-18
idle process state, 5-22 defined, 9-5
IDs, 5-4 physical volume groups, 9-45
immediate report , 12-8 1/0 configuration menu, 2-1 1
immediate reporting, 8-45 ioinit , 2-22, 1 1-17, 1 1-18
immediate response mode, 12-7, 1/0 interface, 10-18
Glossary- 18 1/0 manager, 1 1-20 , 1 1-22
include statements, 1 1-10, 1 1- 14 iomap, 7-40
inetd and disk quotas, 14-70 1/0 mappings, 7-1 5 , Glossary-1 9
init , 2-20, 6-2 , 1 1-17 ioscan, 10-38, 10-41 , 1 1-23
initdefault, 6-2 , Glossary-18 1/0 scheduling policy
inittab, 12-26 parallel, 9-47
inode, 8-28, 8-29, 8-30, 8-32, 8-44, 8-51 , sequential, 9-47
Glossary-18 iostat , 5-16
access to data blocks, 8-32 io statement , 1 1-7, 1 1-23
bad blocks, 8-53 CIO configuration, 1 1-10
consistency, 8-5 1 HP-PB configuration , 1 1-12
duplicate blocks, 8-52 syntax, 1 1-9
format and type , 8-52 IRG, 12-5
link count , 8-52 ISL , 1 1-16
inode and minor numbers, 1 1-34 ITE, Glossary- 19
inode creation, 8-30
inode error messages, 8-30 J
inode header files, 8-29 j ob , Glossary- 19
inodes and accounting, 14-18 j ob control, Glossary-19
insf , 1 1-18, 1 1-26, 1 1-43 j obs
instrO , 1 1-24 and job control, 5-6
interactive search, 2- 19
interface card K
Boot ROM search sequence, 2-5 KCORE Mii , 14-37
HP-IB , 12-28 kermi t , 8-55
select code, 10-8 kernel
interface driver, 1 1-20, Glossary-18 and demand paging, 7-24
interleaving swap devices, 7-44 backup kernel, 8-2 1
Internet address, Glossary- 18 configuration, 1 1-2
interprocess communication, 13-7 configuration file, 1 1-3, 1 1-13
and shared memory, 7-40 configuring drivers, 1 1-25
inter-record gap , 12-5 configuring swap space, 7-52

lndex-1 1
Index

definition, Glossary- 19 lo ckf , 8-64

kernel devices , 1 1-14 locking kernel resources ( MP ) , 5-27
kernel stack segment , Glossary-1 9 locking, process, 7-40
kernel configuration, 1 1-17 logical device manager ( LDM ) , 1 1-20 ,
kernel mode, 5- 17 Glossary-20
kernel stack segment , 7-15 logical extent
keyboard management , 1 1-24 defined , 9-5
Key shell (/usr/bin/keysh) , 4-5 Logic al Interchange Format ( LIF ) , 2-3 ,
kill 8-20 , Glossary-19
described, 5-14 logical unit ( see lu) , Glossary-21
kill key logical volume
setting, 4-7, 4-18 allocating disk space for , 9-23
Korn shell (/bin/ksh) , 4-5 defined, 9-6, Glossary- 19
file name, 9-8
L introduced, 9-2
LABEL file, 9-31 mapping extents, 9-24
LAN, 13-3 organization, 9-21
lanO , 1 1-24 Logical Volume Manager ( LVM )
lan1 I 1 1-24 accessing raw data, 9-9
LAN / 9000, 13-7 and newfs, 9- 14
LAN card for MAP networking , 1 1-24 backing up mirrored data, 9-49
LAN ( networking) management , 1 1-24 capabilities, 9-1
ld, 7-28 , 7-33 character vs. block device files, 9-8
/lib , 7-34 configuration maintenance, 9-37
l if , 8-58 defined , 9-6, Glossary-20
LIF , 8-20, Glossary- 19 DEV_BSIZE, 9-26
LIF header, 9-3 1 device driver, 9-5 1
limit, Glossary-20 device file names, 9-8
line control switch pack settings, 2-5 device files, 9-9
link count in inodes , 8-52 / etc / lvmtab , 9-37
link level address, Glossary-20 extent size , 9-2
links, 13-3 group file, 9-9
802.4, 13-3 hpuxboot options , 9-39
loading HP-UX , 2-6 , 2-7 intern al representation, 9-51
load point , 12-3 1/0 channel separation , 9-43
local login script, Glossary-20 LVM disk device file, 9- 10
local login script, default , 4-24 maintenance mode, 9-39
local login scripts, 4-6 major and minor numbers, 9- 1 0 , 9- 1 1
localrc shell function , 2-25 migration introduced , 9-4
local swapping, Glossary-20 mirror consistency cache mechanisms,
lockable memory, 7-5 , Glossary-21 9-48

lndex- 1 2
 Index

mirroring, 9-40 lv device driver, 9-5 1

mirroring commands, 9-49 lvdisplay ( lM ) , 9-6 , 9-24, 9-50
paradigm, 9-2 lvextend ( lM ) , 9-50
preparation, 9- 1 , 9-2 lvlnboot ( lM ) , 9-30
recovery of mirrored data, 9-48 LVM disk
removable media, 9-39 defined, 9-6 , Glossary-20
scheduling policies, 9-47 LVM disk space
sector size, 9-26 contiguous vs . non-contiguous , 9- 18
synchronization of mirrored data, for root volume group , 9- 18
9-48 lvmerge ( lM ) , 9-50
synchronizing mirrors, 9-49 lvmpvg , 9-45
terminology, 9-5-7 LVM root volume group
used for file systems, 8- 13 allocation of disk space, 9-18
uses, 9-1 LVM subsystem architecture, 9-5 1
volume group minor numbers, 9-10 lvrmboot ( lM ) , 9-30
logical volumes lvsplit ( l M ) , 9-50
compared to sections, 8- 13 lvsync ( lM ) , 9-50
logical volumes , Series 800, 8- 13
. login, Glossary-20 M
login magic, 7-28
and I etc/passwd, 4-3 magnetic tape, 12-2
and /etc/wtmp , 4-3 capabilities , 1 2-4
correcting mistakes during, 4-7 characteristics, 12-2
how HP-UX uses, 4-2 coding, 12-5
invokes shells, 4-3 density, 12-2
. login script, 4-32 errors, 12-6
logname , 4-8 format , 12-4
logout script, 4-27 organization, 12-3
long gap , 12-6 preventive maintenance, 1 2-6
/lost+found, 8-48 , 8-52, 8-55 record checks, 12-5
lprO , 1 1-24 magneto-optical ( MO ) , Glossary-21
lpr 1 , 1 1-24 main memory, 7- 1 , 7-3, 7-24
lpr2 , 1 1-24 maintenance mode, LVM , 9-39
lsdev , 1 1-23 , 1 1-3 1 , 1 1-33 major number , 1 1-23 , 1 1-30 , 1 1-3 1 ,
lu, 1 1-43 Glossary-2 1
and /etc/ioconf ig, 1 1-43 and device drivers, 1 1-31
and ioscan, 1 1-43 defined, 1 1-31
definition, Glossary-21 maj or number, LVM , 9-10
how assigned, 1 1-43 manual boot on Series 800, 2- 18
lvchange ( lM ) , 9-5 , 9-50 manufacturing automation protocol
lvcreate ( lM ) , 9-50 ( MAP ) , Glossary-21

lndex- 1 3
Index

MAP applications, 10-16 process-to-page , 7- 12

MAP card addressing, 10-22 memory-resident resources , 7-2
MAP cards , 10-28 Mid-Bus, 10-15
maxbpg, 8-38 Mid-Bus addressing, 1 0-27
maxds iz , 7- 14 Mid-Bus architecture , 10-27
maxf ile s , 5- 1 1 Mid-Bus ( Series 800 ) , 10-16
maxf iles_lim, 5-1 1 minfree , 8-36
maxlvs , 9-33 minfree , Glossary-22
maxpvs , 9-33 minor number, 1 1-30, 1 1-38, 1 1-42 ,
maxpxs , 9-33 1 1-43, Glossary-22
maxs s iz , 7- 14 and backplane, 1 1-34
maxt s iz , 7-14, 7-15 and behavioral information, 1 1-34
McKusick file system, 8- 1 and bus address, 1 1-34
MEAN SIZE, 14-37 and hardware characteristics, 1 1-34
mechanical changer, Glossary-21 and inodes, 1 1-34
media, Glossary-2 1 and select codes , 1 1-35
mediainit and SPU, 1 1-34
device file required, 8-5 creating, 1 1-39
when to use, 8-5 defined, 1 1-34
memory generic format , 1 1-38
cache , 7- 1 0 minor numbers
locations, 7-6 and insf , 1 1-42
Stable Storage, 2- 16 and Series 800, 1 1-42
transactions, 7-1 0 minor numbers, LYM , 9- 10
types, 7-2 Mirror Consistency Record ( MCR) ,
volatility, 7-2 9-34
memory address, 7- 13 mirroring
memory management backups , 9-49
demand-loaded code, 7-33 commands , 9-49
dynamic swap space, 7-45 defined, 9-6, Glossary-22
page size, 7- 13 introduced , 9- 1
physical memory, 7-2 logical to physical extents, 9-43
primary swap space , 7-43 Mirror Write Record, 9-48
shared code , 7-33 recovering data, 9-48
static swap space , 7-43 synchronization of data, 9-48-49
swapper , 7-27 mirroring, LYM , 9-40-50
swap space requirements, 7-52 mirroring, three-way, 9-40
vhand, 7-27 mirroring, two-way, 9-40
memory management unit ( MMU ) , mirror write consistency cache, 9-48
7- 10, Glossary-22 mkf s , 7-44
memory mapping reserves free space , 8-36

lndex- 1 4
 Index

mklost+found , 8-48 MUX card addressing, 10-10

mknod, 1 1-26 , 1 1-35
mkrs , 2-1 N
mksf , 1 1-26 named object , Glossary-23
MMS/9000 , 13-7 nbuf , 8-40
Model 375 workstation, 10-3 NCSC, Glossary-23
modem device files , 1 1-50 NetIPC , 13-7
modem installation , 12-25 network architecture , 1 3-2
modes, process , 5-17 network device file , 1 1-28
modifying files, 8-43 Network File System (NFS ) , 13-4
module number , 10-19, 1 1-38 networking
monacct , 14-49 , 14-61 application layer, 13-2
monarch and serfs , 5-26 connectivity , 13-2
monitor , 5- 16 data-link layer , 13-2
monthly total command summary, factory-floor devices, 13-7
Glossary-22 interprocess communication, 13-7
mount links , 13-3
device file required, 8-5 network layer, 13-2
moving between run-levels , 6-5 OSI model, 13-2
mp . h, 5-26 physical layer, 13-2
mt , 1 1-27 presentation layer, 13-2
multiprocessing (MP) , 5-26-31 services , 13-3
boot-up , 5-26 session layer, 13-2
bootup , 5-30 transport layer, 13-2
characterization tools , 5-30 transports, 13-3
concurrent processing example, 5-28 network layer , 13-2
defined, Glossary-22 Network Services (NS ) , 13-3 , 13-4
header files , 5-26 newf s
monarch and serf processors, 5-30 device file required, 8-5
monarch and serfs , 5-26 reserves free space , 8-36
processor affinity, 5-30 when to use, 8-6
scheduling, 5-27 newfs(lM)
semaphores and spinlocks, 5-27 device file for LVM , 9-9
symmetry, 5-26 newf s and LVM , 9- 14
timing hazards, 5-30 nf ile , 5- 1 1
uniprocessor emulation, 5-30 NFS/9000 , 13-7
vs. uniprocessing, 5-26 NFS Remote File Access , 8-55
multi-user run-level, Glossary-23 nice
MUX, 10-18, Glossary-23 shell differences described , 5- 15
muxO , 1 1-24 ninode , 5- 1 1
mux 1 , 1 1-24 non return t o zero immediate, 12-5

lndex- 1 5
Index

NS_ARPA Services, Glossary-23 OTS/800 , 13-7

NS nodename, Glossary-23
p
0 pacct , 14-31
objects , Glossary-23 pacct , managing size of file, 14-34
open files ( see processes ) , 5-1 1 page, Glossary-24
open pro.cesses, 5- 1 1 page alignment , 7-3 1
Open Systems Interconnection (OSI ) , page boundaries , 7-28
13-2 page fault , 7-24, Glossary-24
operating-system parameters, 1 1-16 pageout daemon, 7-25
argdevnblks , 7-52 pages, 7-24
bufpages , 8-40 page size, virtual memory, 7-13
dskless_f sbuf s , 8-40 page tables , 7-2
maxds iz , 7- 14 paging, 7- 1 , 7-24, Glossary-24
maxf iles , 5- 1 1 PAM shell ( /bin/pam ) , 4-5
maxf iles_lim, 5- 1 1 parallel interface , 10-5
maxl vs , 9-33 parallel I/O scheduling policy, 9-47
maxpvs , 9-33 parallel-printer management , 1 1-24
maxpxs , 9-33 parallel SCSI interface, 1 1-24
max s s iz , 7- 14, 7- 15 parent process, 5-6, 5-9 , Glossary-24
maxts iz , 7- 14, 7- 1 5 parent process ID , 5-3 , Glossary-24
nbuf , 8-40 PA-RISC 1 . 1 chip set , 10- 1 1
nf ile , 5- 1 1 password, 4-8 , Glossary-24
ninode , 5- 1 1 password aging, 4-9
shmmn i , 7- 14 path name, Glossary-25
shms eg , 7- 14 pdc , 2- 13
swchunk , 7-45 pdnO , 1 1-24
unlockable_mem, 7-5 peripheral device, Glossary-25
optical autochanger, Glossary-24 peripheral location, Glossary-25
Optical technology peripherals
implementation, 12-18 description , 1 1-26
orphaned process group , 5-8, Glossary-24 permission bits, file, 8-58
OSF Mirror Write Consistency Cache , per-process region (pregion ) , 7- 13,
9-48 7-14-15 , 7-24, Glossary-25
os iO , 1 1-24 pfdat data structure, 7-1 3 , 7-24
OSI FTAM ( File Transfer and Access phantom record, Glossary-25
Management ) , 8-55 phase encoding, 12-5
OSI model, 1 3-2 physical extent
connectivity, 13-2 defined, 9-5
OSI Transport Services, 13-7 physical layer, 13-2
OTS , 1 3-3

lndex- 1 6
 Index

physical memory, 7-1 , 7-2 , 7-5, priority, Glossary-26

Glossary-25 priority, swapping, 7-48
at boot time, 7-3 process, Glossary-26
physical volume group process accounting, 14-31 , 14-34, 14-35
defined, 9-6 processes
illustrated, with I / O channel as handled by kernel, 5- 17
separation, 9-45 background process group , 5-8
physical volume ( LVM disk ) controlling process, 5-8
bootable volume layout , 9-29 controlling terminal, 5-8
contents for booting, 9-1 8 creation, 5-9
defined, 9-6 , Glossary-25 defined, 5-2
file name, 9-8 foreground process group , 5-8
layout , 9-29 ID, 5-3
organization, 9-25 open files , 5- 1 1
reserved areas, 9-29 open processes allowed, 5- 1 1
sector size, 9-26 orphaned process group , 5-8
physical volume ( LVM disk ) , bootable, ownership, 5-4
9- 18 parent process ID, 5-3
Physical Volume Reserved Area ( PVRA ) , PID , 5-3
9-32 PPID, 5-3
PID , 5-3, Glossary-26 process group ID, 5-6
pipes, 8-30 process group leader, 5-6
plock ( 2 ) , 7-5 , 7- 13, 7-40 process group lifetime, 5-7
plotter management, 1 1-24 process groups, 5-6
port , Glossary-26 process ID, 5-4 , 5- 1 0
powerfail, 2-27 relative priority , 5- 15
powerfail routines, 7-2 sessions, 5-7
powering up the system, 2-6 terminal affiliation, 5-8
PPID , 5-3 , Glossary-26 termination, 5- 12
prctmp , 14-30 process group , Glossary-26
prctmp script, 14-27 process group ID, 5-6, Glossary-27
prdaily , 14-8 , 14-49 , 14-60 process group leader, 5-6 , Glossary-27
preen mode, f s ck , 2-23 process group lifetime, 5-7 , Glossary-27
presentation layer, 13-2 process ID, Glossary-27
primary boot path, 2-16, 1 1-14 process lifetime, Glossary-27
primary inode, 8-5 1 process locking, 7-40
primary storage, Glossary-26 process modes, 5- 17
primary superblock, 2-23, 8-25 , 8-50 processor affinity, multiprocessing ( MP ) ,
primary swap space, 7-43, Glossary-26 5-30
prime time, 14-4 processor-1 / 0 board, 10-5
printer,interface, 12-29 process states, 5-22

lndex- 1 7
Index

defined in proc . h, 5-22 random access memory ( RAM ) , 7-2,

process state transitions , 5-24 Glossary-28
proc . h, 5-22 raw devices , 1 1-28
proc table entry ( PTE ) , 7-24 raw mode, Glossary-28
. prof ile script , 4-30 raw special files , LVM , 9-9
programmable serial interface re command, Glossary-28
( networking ) management , 1 1-24 rct , 1 1-27
programs rdsk , 1 1-27
as running processes , 5-2 read-only memory ( ROM ) , Glossary-29
protocol, communication , 2-28 read permission , 8-58, 8-60
prtacct , 14-49, 14-50 real group ID, 5-4, Glossary-28
ps , 7-25 real mem, 7-2
options described , 5- 13 real-time interface , 1 1-24
pseudo-drivers, 1 1-25 real user ID, 5-4, Glossary-28
pseudo-swap reservation reboot , Glossary-28
defined, Glossary-27 reboot command, 3-3
pseudo-terminal, 1 1-25 rebooting the system, 3- 1
psiO , 1 1-24 record, Glossary-29
pty, 1 1-25 recovering mirrored data, 9-48
Public data network management , 1 1-24 recovering shared libraries, 2-24
pvcreate ( lM ) , 9-6 recoversl , 2-2 1 , 2-24, 7-39
-B option , 9- 18 recovery system
device file, 9-9 defined , 2- 1
PVG-strict allocation , 9-43 region, 7-2 1 , Glossary-29
relative process priority, 5- 15
Q remote swapping, Glossary-29
QIC-format tape drive rendezvous code, 5-26
device file format , 1 1-48 renice , 5- 15
quiescent state, 8-49 repquota, 14-65, 14-70
quorum resource, Glossary-29
defined, 9-7, Glossary-27 respawn , 2-28
quota, 14-65 restricted shells, 4-5
quota, Glossary-28 rewritable optical, Glossary-29
quotacheck , 14-65 rewritable optical storage, 12-18-24
quotactl, 14-70 rmt , 1 1-27
quotas file , 14-64 root , 1 1-14
quotas . h, 14-65 root , Glossary-29
root device, 12-29
R root file system, 2-20, 2-23, Glossary-29
RAM , 10-2 root logical volume, 9-7, Glossary-29
at boot time , 7-3 Root Volume Group , 9- 16 .

lndex- 1 8
 Index

rquotad , 14-70 role of select codes , 2-7

rt iO , 1 1-24 search sequence , Boot ROM , 2-7
runacct , 14-8, 14-49 search sequence (Series 800) , 2- 17
runacct , using cron to invoke , 14-53-58 secondary loader , 2-3, 8-22, Glossary-30
run-level, Glossary-30 secondary memory, Glossary-30
0, 6-2 secondary storage , 7-6, Glossary-30
creating, 6-5 section, Glossary-30
moving between, 6-5 sector size of physical volumes (LVM
s, 8-49 disks) , 9-26
s, 6-2 security and system accounting, 14-22
run-levels, 2-2 1 security policy, Glossary-30
run process state, 5-22 select code, 10-8 , Glossary-30
run queues, 5- 19 select codes
role in Boot ROM search sequence,
s 2-7
ssoo , 1 1-3, 1 1-7, 1 1-13, 1 1-14, 1 1-16, select codes and minor number, 1 1-35
1 1-17 self-test , 2- 16
io statement , 1 1-7 semaphore, Glossary-3 1
lists configured device drivers, 8-4 semaphores , multiprocessing (MP), 5-27
SSOO file , 1 1-25 sensitive information , Glossary-31
sar , 5- 16 sequential I/O scheduling policy, 9-47
sc c 1 , 1 1-24 serfs and monarch , 5-26
scheduler , 5-20 serial interface card, 2-5
scheduling disk writes, 9-47 serial-printer management , 1 1-24
scheduling, multiprocessing (MP ) , 5-27 Series 300 architecture , 10-2
SCSI, 10-18 Series 400 architecture , 10-7
immediate reporting capability, 8-45 Series 700 addressing, 10-14
s cs i 1 , 1 1-24 Series 700 architecture , 10- 1 1
s c s i2 , 1 1-24 Series 700 bus architecture , 10-12
SCSI addressing, 10-18 Series 700 modules , 10-13
SCSI, differential, 12-30 Series 800 addressing, 10-19
SCSI (differential) , 10-13 Series 800 architecture , 10-15
SCSI guidelines, 12-30 services , 13-3
SCSI-interface management , 1 1-24 session, Glossary-3 1
SCSI, single-ended, 12-30 session layer, 13-2
SCSI (Small Computer System Interface) , session leader, Glossary-3 1
10-5 session lifetime, Glossary-3 1
sdf , 8-58 sessions , 5-7
SDF file system, 8-55 set group ID bit , Glossary-31
search, interactive, 2- 19 set group ID bit (sgid) , 8-65
search sequence set group ID bit (sgid or s etgid) , 8-6 1

l ndex- 1 9
Index

s etpgid, 5-7 HP-UX interfaces, 5-25

set user ID bit , Glossary-31 SIGCONT , 5-24
set user ID bit (suid or setuid) , 8-6 1 SIGKILL , 5-14
shared code , 7-28 , 7-3 1 , 7-33, Glossary-3 1 SIGSTOP , 5-24
i n a cluster , 7-34 SIGTERM , 5-14, 5-25
shared libraries, 7-34-39 SIGTTIN , 5-8
administering, 7-39 SIGTTOU, 5-8
defined , Glossary-32 SIGTTIN signal, 5-8
demand-loaded code , 7-36 SIMM cards, 10-13
recovering, 7-39 sleep process state, 5-22
shared library segment , 7- 15 slot, 1 0-19, Glossary-33
shared library segment defined, SMB
Glossary-32 definition , Glossary-33
vs. archived libraries, 7-38 SMB addressing, 10-30
shared libraries , recovering, 2-24 SMB bus architecture, 10-30
shared memory soft limit , Glossary-33
and interprocess communication , 7-40 software module paths, 10-41
shared memory segment , 7- 15 source device, Glossary-33
shared memory segment defined , special file, Glossary-33
Glossary-32 special files
shared segment , 7-44 block, 1 1-28
SHARE_MAGIC, 7-30 character , 1 1-28
shell, Glossary-32 description, 1 1-26
shell environments spinlock, Glossary-33
initialization, 4-5 spinlock . h, 5-26
shmat , 7-40 spinlocks, multiprocessing ( MP ) , 5-27
shmctl , 7-40 SPU and minor number, 1 1-34
shmctl ( 2 ) , 7-5 Stable Storage, 2- 16
shmget , 7-40 stack segment , 7-44, Glossary-34
shmmn i, 7- 14 stale vs. fresh, 9-7
shmseg, 7- 14 standalone, Glossary-34
shutdown, Glossary-32 standard error, 5- 1 1 , Glossary-34
description, 3- 1 Standard Graphics Connect ( SGC )
halting the system, 3- 1 system bus ( Series 700 ) , 10-13
improper, 2-23 standard input, 5-1 1 , Glossary-34
reboot command, 3-3 standard output , 5- 1 1 , Glossary-34
rebooting the system, 3-1 sticky bit , Glossary-34
shutdown command, 3-2 stopped process state, 5-22
shutdown command, Glossary-33 streaming, Glossary-34
s ignal . h, 5-25 stty , 4-18
signals, 5-25 stty settings, 4- 7

lndex-20
 Index

subfile , Glossary-35 SYSHPUX, 8-2 1

subject , Glossary-35 system accounting
summarizing and reporting accounting acct , 14-25
information, 14-49 acctcms for report of commands,
summary information area, 8-25 14-42
superblock, 8-9 , 8-25, 8-28 , 8-44, 8-50, acct cms report options , 14-45
Glossary-35 acct com, 14-35
superblock, alternate, 8-28 acctcon1, 14-28
superuser, 5-5 , Glossary-35 acctcon2 , 14-27, 14-30
surface, Glossary-35 acctdisk, 14-21
swap , 1 1- 14 acctdusg, 14-17, 14-18
swap , Glossary-35 acctmerg, 14-49 , 14-5 1
swapon, 7-8 , 7-43 , 7-48 acctprc1 and acctprc2 , 14-46
swapper , 7-25 , 7-27 acctsh, 14-27, 14-53, 14-61
swapping, 7-1 , Glossary-35 acctwtmp, 14-24
swapping priority , 7-48 and shutdown, 14-12
swap space , Glossary-35 and startup , 14- 1 1
allocating, 7-9 changing process accounting files,
cluster, 7-52 14-35
configuring, 7-52 charef ee, 14-48
defined, 7-7 chargefee, 14-6, 14-8
determining requirements, 7-5 1 ckpacct , 14-8, 14-34
device swap space, 7-43 cluster concerns, 14-10, 14-22
dynamic, 7-9 cron, 14-8
dynamic swap space, 7-45 daily usage, 14-53
file-system allocations, 7-50 directory structure, 14- 14
listed in /etc/checklist , 7-47 disk quotas, 14-62
swap device , 7-43 diskusg, 14- 17, 14-18
swap-device allocations, 7-48 displaying process accounting records ,
swap space management , 7-41-52 14-35
swchunk , 7-45 dodisk, 14-21
switch settings , 10-8 errors , 14-57
symmetry, 5-26 /et c/wtmp , 14-23
sync , 7-2 file cleanup , 14- 1 1
synchronization files , 14-13
defined , 9- 7, Glossary-36 fwtmp , 14-25
synchronization of mirrored data, 9-48-49 gathering data, 14-8
synchronous vs. asynchronous disk generating reports, 14-9
writes, 8-46 generating summary reports, 14-9
SYSBCKU P , 8-2 1 HOG FACTOR, 14-37
SYSDEBUG, 8-2 1 holidays, 14-4

lndex-2 1
Index

HP-UX cluster , 14-18, 14-19 System Main Bus ( SMB ) , 10-15

installation , 14-10 System Main Bus ( SMB ) ( Series 800 ) ,
KCORE MIN , 14-37 10-16
logging in, 14-12 system performance , 1 1-16
managing size of pacct , 14-34 assigning swap priorities , 7-48
MEAN SIZE, 14-37 degraded by thrashing, 7-27
monacct , 14-49 , 14-61 interleaving swap devices , 7-44
monthly usage , 14-61 using lockable memory to improve ,
overview, 14-1 , 14-2, 14-7 7-5
pacct , 14-31 using swchunk, 7-45
planning and billing, 14-3 system processing unit ( SPU ) , Glossary-6
prctmp , 14-30 system shutdown, 2-23, 3- 1 , Glossary-36
prctmp script , 14-27 system shutdown guidelines , 8-47
prdaily , 14-8 , 14-49, 14-60 system startup , Glossary-37
prime time , 14-4 attended mode, 2- 14
process accounting, 14-31 Boot Administration Mode ( Series
prtacct , 14-49 , 14-50 700 ) , 2- 15
runacct , 14-8 , 14-49, 14-53-58 Boot ROM overview , 2-2
setprivgrp , 14-22 Boot ROM sequence ( Series 300 / 400 ) ,
time considerations, 14-4 2-4
total accounting records, 14-9 Boot ROM sequence ( Series 700 ) ,
use of bdf , 14-22 2- 13
use of du, 14-22 Boot ROM sequence ( Series 800 ) ,
using /etc/re , 14- 1 0 2- 16
wtmpf ix , 14-26 defined, 2-1
system administration mode, 6-2 HP-UX sequence , 2- 1 , 2-20
system administrator run-level, initdefault , 2- 1
Glossary-36 init process, 2-20
system bus module number, 1 1-38 interrupting the boot sequence , 2- 14
system calls loading HP-UX, 2-6
exe c , 5-1 1 physical memory, 7-2
exit , 5- 12 recovery system, 2-1
fork , 5-9, 5- 1 1 system startup guidelines, 8-48
setpgid, 5-7
setrlimit , 5- 1 1 T
vfork , 5-9 , 5- 10 tacct . h , 14-9
system CDFs, Glossary-36 tape 1 , 1 1-24
system console, Glossary-36 tape2 , 1 1-24
system console search sequence , 2-5 tape density, Glossary-37
system interface , LVM , 9-51 tape density and minor number , 1 1-34
system login script, 4- 19, Glossary-36 tape-drive management , 1 1-24

lndex-22
 Index

tape-drive special files, 1 1-27 two-way mirroring, 9-40

tar, 8-58, 12-10
tcio , 8-58 u
TCP /IP transport layer, 13-3 u_area, 5- 18, 7- 15
templates, 1 1-7 unattended mode, 2-6, Glossary-39
TERM environment variable, 4- 16 uncorrectable file-system corruption,
TERM = (hp) prompt, 4- 16 using fsdb for, 8-55
terminal unit number, Glossary-39
adding, 12-25 unload tape, 12-4
configuration, 12-25 unlockable_mem, 7-5
terminal, configuration, 12-25 use count , 7-3 1 , Glossary-39
terminal management , 1 1-24 user, Glossary-39
t erminfo database , 4-16 user data area, 9-35
text segment , 7- 15, Glossary-37 user ID, 8-61
t_gpgshi , 7-26 user mode , 5- 17
t_gpgslo , 7-26 username, 4-8
thrashing, 7-27, Glossary-37 user perspective, on disk quotas, 14-65
three-way mirroring, 9-40 user processes, 7-1
time-out option, 12-26 user stack segment , 7- 15
time-slice, Glossary-37 user warnings , for disk quotas, 14-68
used to manage CPU time, 5- 18 /usr/adm/acct/sum directory, 14-11
top , 5- 16 /usr/adm/pacct , 14- 1 1
total accounting records, 14-9, 14-2 1 , /usr/contrib/bin, 4- 16
14-27, 14-46, 14-49 /usr/lib, 4- 16
track, Glossary-37 /usr/lib/libdld . sl , 7-34
transferring files /usr/lib/terminfo database, 4- 16
methods (table) , 8-55 /usr/local , 7-39
translation lookaside buffer (TLB) , 7- 10, /usr/local/bin, 4-16
Glossary-37 uucp , 8-55
transport layer, 13-2 UUCP , 13-4, Glossary-39
transports, 13-3 uxgen, 1 1-22 , 1 1-25
trap door, Glossary-38
trap instruction, 5- 17 v
Trojan horse , Glossary-38 vas . h, 7- 13
trusted computer system, Glossary-38 verification, Glossary-39
trusted computing base, Glossary-38 vgcfgbackup( lM) , 9-38
trusted process , Glossary-38 vgcfgrestore( lM) , 9-38
trusted software, Glossary-38 vgchange(lM) , 9-7, 9-50
t s et , 4-17 vgcreate(lM) , 9-7, 9-24
tunefs vgdisplay(lM) , 9-24
changes reserved free space, 8-36 vgexport( lM) , 9-39

lndex-23
Index

vgimport ( lM ) , 9-39 organization, 9- 12

vhand, 7-25-27 root group contents, 9- 16
virtual address Volume Group Descriptor Area ( VGDA ) ,
introduced, 7-1 9-33
translation, 7-10 Volume Group Reserved Area ( VGRA ) ,
virtual address space, 7-13-21 , 9-33
Glossary-39 volume group , root , 9-16
configuring, 7-14 Volume Group Status Area ( VGSA ) ,
introduced, 7-9 9-34
of Series 300 /400 , 7-17 volume header, 8-9, 8-2 1
of Series 800, 7-20 volume number, Glossary-40
virtual file descriptor (vtd) , 7-2 1
virtual memory, 7-9 w
virtual node ( vnode ) , 7-2 1 word, Glossary-40
vmstat , 5- 16 write permission, 8-58, 8-60
vnode , 7-2 1 , Glossary-40 write protected, Glossary-40
volatility, 7-2 write protection, 12-2
volume directory information, 8-2 1 write ring, Glossary-40
volume group wtmpf ix to fix wtmp errors, 14-26
defined, 9-7, Glossary-40
file name, 9-8 x
group file , 9-9 X.25, 13-3, 13-7
group file , 9-1 1 X.400, 13-7
HP-IB limitations, 9- 12
introduced, 9-2 z
minor numbers, 9- 10 zombie process state, 5-22

lndex-24