Simulating

Oracle I/O Workloads to

Configure Storage

James F. Koopmann

NoCOUG
Fall Conference 2007
4:00 – 5:00
Who am I

You fill in the blank _____________________

Who am I

Thanks James,

the world is made a better place because of

the generosity of people like you who are
willing to share knowledge. Great effort !
Introduction

 When you think of storage, what do you think of

– Do I have enough?
 How do you determine what storage to buy?
– Just enough to hold data
 Do you use iostat or another I/O utility?
– Why?
– What do you do to fix I/O issues?
• Tune SQL?
• Purchase more CPU
• Add more Memory
• Move datafiles around
Introduction

 The Storage Solution Challenge

 What is RAID
 What is an Oracle Workload
 What is ORION
– Introduction
– Benefits
 Testing Storage
– RAID 0+1 Test
– Scripts to Use
– RAID 0+1 Results
 Existing Storage System
– Vendor Storage Specifications
– What Can You Use
– Scripts to Monitor Disk Workload
 Configuring Oracle
 Conclusion
The Storage Solution Challenge

 It is difficult to translate database requirements into

a storage solution that takes into account the
abilities of a storage array.
 Tuning an Oracle database is a complex task, and
having an incomplete understanding of the
challenge complicates the task.
 Typical scenario
– The database staff requests additional storage
– Storage administrators allocates disk
– Done!
– BUT neither understands scalability, availability,
manageability or performance implications of the other
– Result is storage / database I/O issues
RAID Basics
RAID Basics

 Redundant Array of Independent Disks

– Provides fault tolerance (redundancy)
– Why Use RAID?
– Improves access rates
• What??
• Does it make my disks faster?
• NO!
• Data is striped (spread) across multiple drives and allows
concurrent I/O operations such that it appears to make disks
operate faster
– RAID Levels
• 0,1,2,3,4,5,10,50,0+1,1+0…..
• RAID Level is a critical component for protecting data and
improving application performance.
• Selecting the proper RAID level for a specific application requires
thoughtful contemplation AND TESTING/BENCHMARKING.
RAID Basics
RAID Level 0 - Striping

http://www.lsi.com/insight_center/industry_standards/storage_connectivity/raid/index.html
RAID Basics
RAID Level 1- Mirroring

http://www.lsi.com/insight_center/industry_standards/storage_connectivity/raid/index.html
RAID Basics
RAID Level 0+1 - Striping and Mirroring

http://www.lsi.com/insight_center/industry_standards/storage_connectivity/raid/index.html
RAID Basics
RAID Level 5 - Striping w/Rotational Parity

http://www.lsi.com/insight_center/industry_standards/storage_connectivity/raid/index.html
Determining the Application Mix

Key to Configuring Storage

Determining the Application Mix
OLTP – Online Transaction Processing
 An OLTP application has these characteristics:
– Read- and write-intensive
– Disk Access is Random
– Transaction-based
– Performance measured by IOPS and average
latency (I/O turn-around time)
– Having typical usage of large bursts of small reads
followed by inserts and updates
– Primarily driven by the database issuing small
random I/Os the size of the defined Oracle database
block size (db_block_size).
– Having a database block size ranging from 2 KB to
32 KB, with 8 KB being average.
Determining the Application Mix
OLAP – Online Analytical Processing
 An OLAP application has these characteristics:
– Data Warehouse or Reporting System
– Throughput-based
– Disk Access is Sequential
– Performance-based on moving large amounts of
data in megabytes per second (MBPS)
– Driven by read only databases except during the
load stage
– Having large multi-block read/write I/O streams
containing multiple 1-MB I/Os
Determining the Application Mix
Hybrid / Changing App
 The following list captures a few of the many sources of
application change.
– A simple OLTP database could be cloned to a different system
and could begin to be used with an OLAP application. The data
is the same, but the I/O pattern is very different.
– A new group of power users could turn an OLTP database into
an OLAP reporting engine without notification.
– The application mix might change throughout the day.
– Database backups that run during the day affect the type of
read activity of the database
– Data loads or extraction, transformation, and load (ETL)
processing could switch an OLAP environment into an OLTP
environment
Determining the Application Mix
Determining the Oracle Workload

Application Mix = Workload

 Determining the Workload for an
Existing Oracle Database
gv$sysstat (name, value)
Total Reads :'physical read total IO requests'
Total Writes:'physical write total IO requests'
Large Reads :'physical read total multi block requests'
Large Writes:'physical write total multi block requests'
Total Bytes Read :'physical read total bytes'
Total Bytes Written :'physical write total bytes'

Therefore:
Small Reads = Total Reads – Large Reads
Small Writes = Total Writes – Large Writes
 Determining the Workload for an
Existing Oracle Database
SQL> SELECT
'Number of Small Reads :'||
sum(decode(name,'physical read total IO requests',value,0)-
decode(name,'physical read total multi block requests',value,0)),
'Number of Small Writes:'||
sum(decode(name,'physical write total IO requests',value,0)-
decode(name,'physical write total multi block requests',value,0)),
'Number of Large Reads :'||
sum(decode(name,'physical read total multi block requests',value,0)),
'Number of Large Writes:'||
sum(decode(name,'physical write total multi block requests',value,0)),
'Total Bytes Read :'||
sum(decode(name,'physical read total bytes',value,0)),
'Total Bytes Written :'||
sum(decode(name,'physical write total bytes',value,0))
FROM gv$sysstat;

Number of Small Reads :205903

Number of Small Writes:106883
Number of Large Reads :40298
Number of Large Writes:2791
Total Bytes Read :4188587008
Total Bytes Written :2009381888
 Determining the Workload for an
Existing Oracle Database
Wait 10 Minutes
 Determining the Workload for an
Existing Oracle Database
SQL> SELECT
'Number of Small Reads :'||
sum(decode(name,'physical read total IO requests',value,0)-
decode(name,'physical read total multi block requests',value,0)),
'Number of Small Writes:'||
sum(decode(name,'physical write total IO requests',value,0)-
decode(name,'physical write total multi block requests',value,0)),
'Number of Large Reads :'||
sum(decode(name,'physical read total multi block requests',value,0)),
'Number of Large Writes:'||
sum(decode(name,'physical write total multi block requests',value,0)),
'Total Bytes Read :'||
sum(decode(name,'physical read total bytes',value,0)),
'Total Bytes Written :'||
sum(decode(name,'physical write total bytes',value,0))
FROM gv$sysstat;
Number of Small Reads :500211
Number of Small Writes:123474
Number of Large Reads :142981
Number of Large Writes:8010
Total Bytes Read :22232604961
Total Bytes Written :5586081648
 Determining the Workload for an
Existing Oracle Database
Calculate the IOPS for small I/O,
the IOPS for large I/O,
the percentage of reads to writes,
the MBPS throughput
Small Read IOPS =(500211-205903)/(10*60) = 490 IOPS
Small Write IOPS =(123474-106883)/(10*60) = 27 IOPS
Total Small IOPS =(294308-16591)/(10*60) = 517 IOPS
I/O Percentage of Reads to Writes = 94:6
Large Read IOPS =(142981-40298)/(10*60) = 171 IOPS
Large Write IOPS =(8010-2791) /(10*60) = 8 IOPS
Total Large IOPS =(102683+5219) /(10*60) = 179 IOPS
I/O Percentage of Reads to Writes = 96:4
Total MBPS Read =((22232604961-4188587008)/(10*60))/1048576 =28 MBPS
Total MBPS Written =((5586081648-2009381888) /(10*60))/1048576= 5 MBPS
Total MBPS =((18044017953+3576699760)/(10*60))/1048576 =34 MBPS

What type of application do we have?

 Introducing ORION

Let the Benchmarking Begin

Introducing ORION
Oracle I/O Numbers
 ORION is a configurable workload generating
tool that mimics Oracle I/O without requiring
installation of an Oracle database.
 The ORION workload tool has a variety of
options that enable storage architects to design
an OLTP or OLAP workload that matches the
expected transaction mixes.
 Stresses a storage array and gives results that
allows us to better understand storage
configurations so that we can design storage to
meet performance requirements.
Introducing ORION
Possible Tasks
 Create benchmarks for specific application mixes
 Single-node tests
 Multi-node (RAC) tests
 Balancing disk drive layout against system workload
 Verifying performance levels for various I/O scenarios
 Revealing specific I/O benefits and deficiencies
 Configuring storage array before use
 Configuring storage before installing Oracle database
 Timely fixing of hardware flaws
 Comparing benchmarks across different storage
products and configurations
Introducing ORION
ORION Workload Tool Options
Three Run Levels

Simple Small and large random I/O are

tested individually
Normal Small and large random I/O are
tested individually and
combinations of small and random
I/Os together

Advanced A wide variety of options are

available to fine tune the workload
and will be user defined
Introducing ORION
ORION Workload Tool Options
Advanced Run Levels

Num_disks Defines the number of spindles in the

storage array to be tested
Size_small Defines the size for small random I/O

Size_large Defines the size for large random or

sequential I/O
Type Defines the type of large I/O (random or
sequential)
Write Defines the percentage of writes in the
workload
Matrix Defines the mixture of workload to run
(random, sequential, or both together)
Introducing ORION
Using the Orion Workload Tool
 Very Easy to Use the Orion Workload Tool.
 Download the Orion Workload tool.
http://www.oracle.com/technology/software/tech/orion/index.html
 Read Orion User Guide (at least twice)
 Install the Orion Workload tool by unzipping the file.
 Create a file that contains a list of raw volumes or
files to test.
 Run the Orion Workload tool with workload options.
 View the tabular output.
 Graph the tabular output

***Warning ****
DO NOT run ORION on a production Machine or any
machine that has data you are concerned about
KEEPING!
ORION Benchmarking

In Practice
ORION Benchmarking

 The Storage System

 Test Methodology
 Test Scenarios
 RAID 0+1 Example
 Volume Group Layout
 What to Test
– 64, 128, 192, & 224 Disks
 Scripts Used
 Understanding the Results
– IOPS Overview
– IOPS Graphical
ORION Benchmarking
The Storage System

Drive Tray
14
Drive
Trays

16 Drives / Drive Tray

SAN
ORION Benchmarking
Test Scenarios / Methodology
 Prove Performance
– Test various RAID Levels
• 0,1,0+1,5
– Test Various Read/Write Percentages
• 0/100, 75/25, 0/100
 Prove Scalability
– Assume Entry Level
• 4 Trays (8TB)
– Expand Storage
• 8, 12, 14 Trays (17TB, 26TB, 30TB)
 Automate All Tests
– Pre-Build a RAID System
– Present all volumes to Host
– Run Scripts
– 48 Tests / 250 Hours
 ORION Benchmarking
Volume Group Layout – RAID 0+1
slot 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Tray 1 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 2 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 3 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 4 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 5 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 6 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 7 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 8 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 9 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 10 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48

Tray 11 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 12 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 13 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 14 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48
 ORION Benchmarking
What to Test : 4-Tray Test : 64 Disks
slot 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Tray 1 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 2 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 3 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 4 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 5 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 6 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 7 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 8 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 9 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 10 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48

Tray 11 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 12 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 13 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 14 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48
 ORION Benchmarking
What to Test : 8-Tray Test : 128 Disks
slot 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Tray 1 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 2 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 3 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 4 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 5 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 6 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 7 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 8 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 9 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 10 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48

Tray 11 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 12 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 13 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 14 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48
 ORION Benchmarking
What to Test : 12-Tray Test : 192 Disks
slot 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Tray 1 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 2 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 3 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 4 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 5 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 6 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 7 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 8 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 9 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 10 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48

Tray 11 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 12 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 13 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 14 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48
 ORION Benchmarking
What to Test : 14-Tray Test : 224 Disks
slot 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Tray 1 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 2 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 3 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 4 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 5 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 6 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 7 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 8 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 9 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 10 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48

Tray 11 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 12 VG 9 VG 9 VG 10 VG 10 VG 11 VG 11 VG 12 VG 12 VG 13 VG 13 VG 14 VG 14 VG 15 VG 15 VG 16 VG 16

Tray 13 VG 25 VG 25 VG 26 VG 26 VG 27 VG 27 VG 28 VG 28 VG 29 VG 29 VG 30 VG 30 VG 31 VG 31 VG 32 VG 32

Tray 14 VG 41 VG 41 VG 42 VG 42 VG 43 VG 43 VG 44 VG 44 VG 45 VG 45 VG 46 VG 46 VG 47 VG 47 VG 48 VG 48
ORION Benchmarking
Scripts : crontab Entry
35 14 22 3 * /orion/Host/bin/orion_run.ksh
 ORION Benchmarking
Scripts : orion_run.ksh
#!/usr/bin/ksh
cd /orion/Host/bin
cat /orion/Host/bin/runnames.txt |\
while read NAME
do
DISKS=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $2}'`
SMALL=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $3}'`
LARGE=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $4}'`
TYPE=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $5}'`
SIMULATE=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $6}'`
DURATION=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $7}'`
WRITE=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $8}'`
MATRIX=`grep $NAME /orion/Host/bin/rundata.txt | awk '{print $9}'`
/orion/Host/bin/tray01.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray02.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray03.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray04.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray05.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray06.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray07.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray08.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray09.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray10.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray11.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray12.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray13.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
/orion/Host/bin/tray14.ksh $NAME $DISKS $SMALL $LARGE $TYPE $SIMULATE $DURATION $WRITE $MATRIX &
sleep 300
count=`ls -1 /orion/Host/bin/*.flag | wc -l`
while [ $count -gt 0 ]
do
sleep 600
count=`ls -1 /orion/Host/bin/*.flag | wc -l`
done
done
 ORION Benchmarking
Scripts : runnames.txt
R10DTC4SIO8randomLIO1024randomRPCT100
R10DTC4SIO8randomLIO1024randomRPCT75
R10DTC4SIO8randomLIO1024randomRPCT0
R10DTC4SIO8randomLIO1024sequentialRPCT0
R10DTC8SIO8randomLIO1024randomRPCT100
R10DTC8SIO8randomLIO1024randomRPCT75
R10DTC8SIO8randomLIO1024randomRPCT0
R10DTC8SIO8randomLIO1024sequentialRPCT0
R10DTC12SIO8randomLIO1024randomRPCT100
R10DTC12SIO8randomLIO1024randomRPCT75
R10DTC12SIO8randomLIO1024randomRPCT0
R10DTC12SIO8randomLIO1024sequentialRPCT0
R10DTC14SIO8randomLIO1024randomRPCT100
R10DTC14SIO8randomLIO1024randomRPCT75
R10DTC14SIO8randomLIO1024randomRPCT0
R10DTC14SIO8randomLIO1024sequentialRPCT0
 ORION Benchmarking
Scripts : rundata.txt
testname num_disks size_small size_large type simulate duration write matrix
R10DTC4SIO8randomLIO1024randomRPCT100 16 8 1024 rand concat 120 0 basic
R10DTC4SIO8randomLIO1024randomRPCT75 16 8 1024 rand concat 120 25 basic
R10DTC4SIO8randomLIO1024randomRPCT0 16 8 1024 rand concat 120 100 basic
R10DTC8SIO8randomLIO1024randomRPCT100 16 8 1024 rand concat 120 0 basic
R10DTC8SIO8randomLIO1024randomRPCT75 16 8 1024 rand concat 120 25 basic
R10DTC8SIO8randomLIO1024randomRPCT0 16 8 1024 rand concat 120 100 basic
R10DTC12SIO8randomLIO1024randomRPCT100 16 8 1024 rand concat 120 0 basic
R10DTC12SIO8randomLIO1024randomRPCT75 16 8 1024 rand concat 120 25 basic
R10DTC12SIO8randomLIO1024randomRPCT0 16 8 1024 rand concat 120 100 basic
R10DTC14SIO8randomLIO1024randomRPCT100 16 8 1024 rand concat 120 0 basic
R10DTC14SIO8randomLIO1024randomRPCT75 16 8 1024 rand concat 120 25 basic
R10DTC14SIO8randomLIO1024randomRPCT0 16 8 1024 rand concat 120 100 basic
ORION Benchmarking
Scripts : tray01.ksh
#!/usr/bin/ksh
cd /orion/Host/bin

if [ -f 01_$2.lun ]
then
touch /orion/Host/bin/01.flag
cp /orion/Host/bin/01_$2.lun /orion/Host/bin/01_$1.lun

/orion/Host/bin/orion_linux
-run advanced -testname 01_$1 -num_disks $2
-size_small $3 -size_large $4 -type $5
-simulate $6 -duration $7 -write $8 -matrix $9

rm /orion/Host/bin/01.flag
fi
ORION Benchmarking
Scripts : 01_16.lun
/dev/sdb
/dev/sddl 4 Volume Groups (Luns)
/dev/sdh 16 Drives
/dev/sddq
slot 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Tray 1 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

Tray 2 VG 17 VG 17 VG 18 VG 18 VG 19 VG 19 VG 20 VG 20 VG 21 VG 21 VG 22 VG 22 VG 23 VG 23 VG 24 VG 24

Tray 3 VG 33 VG 33 VG 34 VG 34 VG 35 VG 35 VG 36 VG 36 VG 37 VG 37 VG 38 VG 38 VG 39 VG 39 VG 40 VG 40

Tray 4 VG 49 VG 49 VG 50 VG 50 VG 51 VG 51 VG 52 VG 52 VG 53 VG 53 VG 54 VG 54 VG 55 VG 55 VG 56 VG 56

Tray 5 VG 1 VG 1 VG 2 VG 2 VG 3 VG 3 VG 4 VG 4 VG 5 VG 5 VG 6 VG 6 VG 7 VG 7 VG 8 VG 8

tray01.ksh tray05.ksh
(01_16.lun) (05_16.lun)
ORION Benchmarking
Understanding the Test Results
 ORION Outputs
– IOPS
– MBPS
– Latencies
– Summary
 IOPS Overview
 Graphical Results
– 0, 75, 100 percent Reads
ORION Benchmarking
IOPS Results
 Input/outputs per second (IOPS) are primarily a
concern in an online transaction processing
(OLTP) application environment. OLTP
application environments have these
characteristics:
– High intensity of reads and writes
– Highly transaction-based applications
– Large bursts of small reads followed by inserts and
updates
– Small random I/Os, typically the size of the database
block size
– Read-to-write ratios, typically 70:30
ORION Benchmarking
IOPS Results - Overview
RAID Drive Tray Drive Read-to-Write Ratios
Level Count Count 0:100 75:25 100:0
0 4 64 16329 18759 20121
0 8 128 30960 35086 37500
0 12 192 42685 47302 50342
0 14 224 48290 52917 56288
1 4 64 9043 15536 20885
1 8 128 17414 29305 38845
1 12 192 24481 39376 51189
1 14 224 27885 43788 56467
10 4 64 8513 14720 19365
10 8 128 16577 27605 36186
10 12 192 23131 36766 47360
10 14 224 26104 40614 52034
5 4 64 4258 9997 18149
5 8 128 8180 18958 34283
5 12 192 11182 25128 44831
5 14 224 12564 27797 49295
ORION Benchmarking
IOPS Results - Overview
Read-to-Write Ratios IOPS / Drive
RAID Drive Tray Drive
Level Count Count 0:100 75:25 100:0 100:0
0 4 64 16329 18759 20121 314
0 8 128 30960 35086 37500 292
0 12 192 42685 47302 50342 254
0 14 224 48290 52917 56288 251
1 4 64 9043 15536 20885 326
1 8 128 17414 29305 38845 303
1 12 192 24481 39376 51189 266
1 14 224 27885 43788 56467 252
10 4 64 8513 14720 19365 302
10 8 128 16577 27605 36186 282
10 12 192 23131 36766 47360 246
10 14 224 26104 40614 52034 232
5 4 64 4258 9997 18149 283
5 8 128 8180 18958 34283 267
5 12 192 11182 25128 44831 233
5 14 224 12564 27797 49295 220
ORION Benchmarking
IOPS - RAID Comparison - Read=0%
I/O Throughput (RAID Comparison, Read=0% ) I/O Throughput (RAID Comparison, Read=0% )

60000 60000

50000 50000

40000 40000
64 Drives RAID 0 128 Drives RAID 0
IOPS

IOPS
64 Drives RAID 1 128 Drives RAID 1
30000 30000
64 Drives RAID 10 128 Drives RAID 10
64 Drives RAID 5 128 Drives RAID 5
20000 20000

10000 10000

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Outstanding Asynchronous I/O (Load) Outstanding Asynchronous I/O (Load)

I/O Throughput (RAID Comparison, Read=0% ) I/O Throughput (RAID Comparison, Read=0% )

60000 60000

50000 50000

40000 40000
192 Drives RAID 0 224 Drives RAID 0
IOPS

IOPS

192 Drives RAID 1 224 Drives RAID 1

30000 30000
192 Drives RAID 10 224 Drives RAID 10
192 Drives RAID 5 224 Drives RAID 5
20000 20000

10000 10000

0 0
0 20 40 60 80 100 0 20 40 60 80 100 120
Outstanding Asynchronous I/O (Load) Outstanding Asynchronous I/O (Load)
ORION Benchmarking
IOPS - RAID Comparison - Read=75%
I/O Throughput (RAID Comparison, Read=75% ) I/O Throughput (RAID Comparison, Read=75% )

60000 60000

50000 50000

40000 40000
64 Drives RAID 0 128 Drives RAID 0
IOPS

IOPS
64 Drives RAID 1 128 Drives RAID 1
30000 30000
64 Drives RAID 10 128 Drives RAID 10
64 Drives RAID 5 128 Drives RAID 5
20000 20000

10000 10000

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Outstanding Asynchronous I/O (Load) Outstanding Asynchronous I/O (Load)

I/O Throughput (RAID Comparison, Read=75% ) I/O Throughput (RAID Comparison, Read=75% )

60000 60000

50000 50000

40000 40000
192 Drives RAID 0 224 Drives RAID 0
IOPS

IOPS

192 Drives RAID 1 224 Drives RAID 1

30000 30000
192 Drives RAID 10 224 Drives RAID 10
192 Drives RAID 5 224 Drives RAID 5
20000 20000

10000 10000

0 0
0 20 40 60 80 100 0 20 40 60 80 100 120
Outstanding Asynchronous I/O (Load) Outstanding Asynchronous I/O (Load)
ORION Benchmarking
IOPS - RAID Comparison - Read=100%
I/O Throughput (RAID Comparison, Read=100% ) I/O Throughput (RAID Comparison, Read=100% )

60000 60000

50000 50000

40000 40000
64 Drives RAID 0 128 Drives RAID 0
IOPS

IOPS
64 Drives RAID 1 128 Drives RAID 1
30000 30000
64 Drives RAID 10 128 Drives RAID 10
64 Drives RAID 5 128 Drives RAID 5
20000 20000

10000 10000

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Outstanding Asynchronous I/O (Load) Outstanding Asynchronous I/O (Load)

I/O Throughput (RAID Comparison, Read=100% ) I/O Throughput (RAID Comparison, Read=100% )

60000 60000

50000 50000

40000 40000
192 Drives RAID 0 224 Drives RAID 0
IOPS

IOPS

192 Drives RAID 1 224 Drives RAID 1

30000 30000
192 Drives RAID 10 224 Drives RAID 10
192 Drives RAID 5 224 Drives RAID 5
20000 20000

10000 10000

0 0
0 20 40 60 80 100 0 20 40 60 80 100 120
Outstanding Asynchronous I/O (Load) Outstanding Asynchronous I/O (Load)
Existing System

Can I Validate?
Existing System
What Can You Do
 Disk Characteristics
– Check Profile of SAN
 Calculating IOPS from Vendor Data
 Calculating MBPS from Vendor Data
 Monitoring Your System
– Database Statistics
– Sample Scripts
– Validating IOPS & MBPS
 Re-Configuration
Existing System
Disk Characteristics - What a Vendor Gives You
RPM
Revolutions Per Minute. The
number of rotations completed
in one minute.

•Seek Time The

time required to position the
head over the target track.
(less for read than write)

•Average Latency
(Rotational Latency) Average
time after head is over the
track for the target sector to
rotate under the head before a
read or write.
 Existing System
Disk Characteristics - Calculate IOPS for Disk
RPM Rotations Rotations Full Rotation Rotational Latency Average IO IOPS
Rotations Per Per (Half Rotation) Seek Time
Per Second Mili-second Time
Minute

(x) (x/60) (x/60,000) (1/ [x/60000] ) (1/ [x/60000] ) / 2

Y Z (Y+Z) (1/[Y+Z])*1000

15,000 15,000/60 15,000/ 4ms 2ms 4ms 6ms 167

60,000

10,000 10,000/60 10,000/ 6ms 3ms 5.15ms 8.15ms 122

60,000

10ms 5ms 9ms 14ms 71

7,200 7200/60 7,200/ 8.4ms 4.2ms 9.9ms 14.1ms 71

60,000
Existing System
Disk Characteristics - Calculate MBPS for Disk
IOPS Number of Segment MBPS
Disks Size

X Y Z (X*Y*Z)/1024

167 1 128K 20MBPS

122 1 128K 15MBPS

71 1 128K 9MBPS

71 1 128K 9MBPS
ORION Benchmarking
MBPS - RAID Comparison-Read=various%
Data Transfer Rate (RAID Comparison, Read=0%) Data Transfer Rate (RAID Comparison, Read=75%)

1800 1800

1600 1600

1400 1400

1200 1200
224 Drives RAID 0 224 Drives RAID 0
MBPS

MBPS
1000 224 Drives RAID 1 1000 224 Drives RAID 1
800 224 Drives RAID 10 800 224 Drives RAID 10
224 Drives RAID 5 224 Drives RAID 5
600 600

400 400

200 200

0 0
0 10 20 30 40 50 0 10 20 30 40 50
Outstanding Asynchronous I/O (Load) Outstanding Asynchronous I/O (Load)

Data Transfer Rate (RAID Comparison, Read=100%)

1800

1600

1400

1200
224 Drives RAID 0

MBPS
1000 224 Drives RAID 1
800 224 Drives RAID 10
224 Drives RAID 5
600

400

200

0
0 10 20 30 40 50
Outstanding Asynchronous I/O (Load)
Existing System
Scripts – Validate IOPS & MBPS
 Set of scripts to run against your existing
database
 Check / Validate what the database sees
against what you expect from hardware
 Run against current statistics
 Run against Snapshots
Existing System
Report IOPS & MBPS over time
 Purpose
– Report IOPS & MBPS
• For small and large I/O
• Percentage of Reads to Writes
• User Defined Sample Period

 Importance
– Matching Expected I/O rates on the Disk Subsystem
to What is Actually Seen in the Database

 Script
– vsysstat_ioworkload.sql
 Existing System
vsysstat_ioworkload.sql - Output
First Sample
Number of Small Reads : 6773457
Number of Small Writes: 8150855
Number of Large Reads : 2114186
Number of Large Writes: 5860450
Total Bytes Read : 3.1256E+11
Total Bytes Written : 2.8618E+11
Enter the amount of time (in seconds) you would like this process to sleep for sampling data
Enter value for sleeptime: 300
Sleep Time (secs): 300
Second Sample
Number of Small Reads : 6778045
Number of Small Writes: 8162002
Number of Large Reads : 2114206
Number of Large Writes: 5867811
Total Bytes Read : 3.1260E+11
Total Bytes Written : 2.8651E+11
Results :
Small Read IOPS = 15.293 IOPS
Small Write IOPS = 37.157 IOPS
Total Small IOPS = 52.45 IOPS
Small Read I/O % = 29.158 %
Small Write I/O % = 70.842 %
Large Read IOPS = .067 IOPS
Large Write IOPS = 24.537 IOPS
Total Large IOPS = 24.603 IOPS
Large Read I/O % = .271 %
Large Write I/O % = 99.729 %
Total Read = .127 MBPS
Total Written = 1.049 MBPS
Total MBPS = 1.176 MBPS
Existing System
Report IOPS & MBPS over Snapshots
 Purpose
– Report IOPS & MBPS
• Uses Snapshots ; Workload Repository History (WRH)
• For small and large I/O
• Percentage of Reads to Writes
• User Defined Sample Period

 Importance
– Matching Expected I/O rates on the Disk Subsystem
to What is Actually Seen in the Database

 Warning
– Performance Tuning Pack Requirement

 Script
– wrh_sysstat_ioworkload.sql
 Existing System
wrh_vsysstat_ioworkload.sql - Output
 First Sample
 Number of Small Reads : 24352077
 Number of Small Writes: 21296716
 Number of Large Reads : 3305589
 Number of Large Writes: 14464153
 Total Bytes Read : 5.6736E+11
 Total Bytes Written : 5.6482E+11
 Second Sample
 Number of Small Reads : 24759830
 Number of Small Writes: 21652738
 Number of Large Reads : 3314771
 Number of Large Writes: 14664751
 Total Bytes Read : 5.7180E+11
 Total Bytes Written : 5.7518E+11
 Results :
 Small Read IOPS = 113.265 IOPS
 Small Write IOPS = 98.895 IOPS
 Total Small IOPS = 212.16 IOPS
 Small Read I/O % = 53.387 %
 Small Write I/O % = 46.613 %
 Large Read IOPS = 2.551 IOPS
 Large Write IOPS = 55.722 IOPS
 Total Large IOPS = 58.272 IOPS
 Large Read I/O % = 4.377 %
 Large Write I/O % = 95.623 %
 Total Read = 1.176 MBPS
 Total Written = 2.744 MBPS
 Total MBPS = 3.921 MBPS
Existing System
Report IOPS & MBPS over Snapshots
 Purpose
– Report IOPS & MBPS
• Uses Snapshots ; Workload Repository History (WRH)
• For small and large I/O
• Percentage of Reads to Writes
• EVERYTHING

 Importance
– Matching Expected I/O rates on the Disk Subsystem
to What is Actually Seen in the Database

 Warning
– Performance Tuning Pack Requirement

 Script
– wrh_sysstat_ioworkload_ALL.sql
 Existing System
wrh_vsysstat_ioworkload_ALL.sql - Output
Small Small Total Small Small
Read Write Small Read Write
END_TIME IOPS IOPS IOPS I/O% I/O%
------------------------- ---------- ---------- ---------- ---------- ----------
13-AUG-07 07.00.54.540 AM 43.117 53.969 97.087 44.411 55.589
13-AUG-07 08.00.56.844 AM 41.677 53.642 95.318 43.724 56.276
13-AUG-07 09.00.00.297 AM 83.613 66.803 150.417 55.588 44.412
13-AUG-07 10.00.04.307 AM 143.974 122.113 266.087 54.108 45.892
13-AUG-07 11.00.07.604 AM 178.109 149.302 327.41 54.399 45.601
13-AUG-07 12.00.10.494 PM 100.027 86.229 186.256 53.704 46.296
13-AUG-07 01.00.13.213 PM 119.235 94.052 213.288 55.904 44.096
13-AUG-07 02.00.17.176 PM 239.61 198.644 438.254 54.674 45.326
13-AUG-07 03.00.20.998 PM 139.792 108.863 248.655 56.219 43.781
13-AUG-07 04.00.26.942 PM 457.685 277.04 734.725 62.293 37.707
13-AUG-07 05.00.30.527 PM 150.167 118.779 268.946 55.835 44.165
13-AUG-07 06.00.33.084 PM 30.141 24.554 54.695 55.108 44.892
13-AUG-07 07.00.36.907 PM 76.618 70.219 146.837 52.179 47.821
13-AUG-07 08.00.39.735 PM 31.321 45.13 76.451 40.968 59.032
13-AUG-07 09.00.42.588 PM 197.372 150.143 347.515 56.795 43.205
13-AUG-07 10.00.45.432 PM 37.311 32.712 70.022 53.284 46.716
13-AUG-07 11.00.48.299 PM 265.166 69.311 334.477 79.278 20.722
14-AUG-07 12.00.50.723 AM 186.941 67.7 254.641 73.413 26.587
14-AUG-07 01.00.52.585 AM 118.794 94.839 213.633 55.607 44.393
14-AUG-07 02.00.55.051 AM 199.069 159.653 358.721 55.494 44.506
14-AUG-07 03.00.56.688 AM 38.899 38.201 77.1 50.452 49.548

… And MUCH MORE data

Existing System
wrh_sysstat_ioworkload_ALL.sql - Graph
Existing System
Report I/O Percentages
 Purpose
– Report I/O Percentages
• Across mount points
• Current database statistics

 Importance
– Investigate which mount point are experiencing the
most read or writes
– Use to see if any mount point is being used more
than others

 Script
– vfilestat_iopercent_mount.sql
 Existing System
vfilestat_iopercent_mount.sql - Output
 Percentage
 Mount of Total Percentage Percentage
 Point I/O of Reads of Writes
 ------------------------------ ---------- ---------- ----------
 /u00/app/oracle/oradata/prd .81 .59 1.18
 /u02/app/oracle/oradata/prd .87 .94 .76
 /u03/app/oracle/oradata/prd 7.20 .22 19.07
 /u04/app/oracle/oradata/prd 46.31 45.90 47.01
 /u05/app/oracle/oradata/prd 26.85 27.69 25.41
 /u06/app/oracle/oradata/prd 5.12 7.49 1.08
 /u07/app/oracle/oradata/prd .00 .00 .00
 /u09/app/oracle/oradata/prd 4.03 4.45 3.33
 /u10/app/oracle/oradata/prd .54 .71 .25
 /u11/app/oracle/oradata/prd 8.28 12.02 1.90
 10 rows selected.
Existing System
Report I/O Percentages
 Purpose
– Report I/O Percentages
• Across mount points
• For small and large I/O
• Uses Snapshots ; Workload Repository History (WRH)

 Importance
– Investigate which mount point are experiencing the most read
or writes
– Use to see if any mount point is being used more than others

 Script
– wrh_vfilestat_iopercent_mount.sql
 Existing System
wrh_vfilestat_iopercent.sql - Output
Percentage
Mount of Total Percentage Percentage
Point I/O of Reads of Writes
------------------------------ ---------- ---------- ----------
/u02/app/oracle/oradata/prd .93 .96 .89
/u03/app/oracle/oradata/prd 7.17 .22 19.32
/u04/app/oracle/oradata/prd 45.99 45.13 47.48
/u05/app/oracle/oradata/prd 27.99 29.32 25.67
/u06/app/oracle/oradata/prd 5.10 7.39 1.09
/u07/app/oracle/oradata/prd .00 .00 .00
/u09/app/oracle/oradata/prd 4.01 4.38 3.36
/u10/app/oracle/oradata/prd .57 .74 .25
/u11/app/oracle/oradata/prd 8.24 11.86 1.92
Conclusion / Questions
Simulating
Oracle I/O Workloads to
Configure Storage

James F. Koopmann

NoCOUG
Fall Conference 2007
4:00 – 5:00