Process Management

• We studied process management in chapter 4

– Here, we examine managing processes from a user’s
and system administrator’s perspective
• how to start a process
• moving processes between foreground and background
• how to monitor processes
• how to change process priority
• scheduling processes
• terminating processes
– We then look at special processes known as services
• how to configure them
• how to control them
• how to alter when they start
 Starting a Process
• To user issues a command requesting that the operating
system start a new process
– Through the GUI
• Double clicking on a short cut icon
• Selecting the process name from a menu
– From the command line
• Typing the name of the process (including the proper path if necessary
to the process’ executable)

A Windows shortcut contains the

executable instruction as if you entered
it from the command line
 What the OS does Now
• No matter how the user requests a process, the OS
now takes over
– the OS interprets the command
– if from the command line, it includes the location of the
executable
– if from the GUI, the shortcut icon or menu property for
the program includes the location of the executable
– the OS then finds the executable, loads it into swap
space and then loads initial pages into memory
– the OS creates a data structure that stores the process’
status information
– if room exists, the OS moves the process into the ready
 Types of Processes in Linux
• In Linux, a process is created by a parent process
which “spawns” the child process
– For instance, the bash shell might be the parent or the
Windows environment might be the parent
• Commonly, child processes will report their status
to the parent process for book keeping
– If a parent process dies it causes the child process to
become an “orphan” in which case the child is usually
adopted by the first process, init
– A child process is not allowed to exit the system until
the parent acknowledges that it can leave, the parent
may not be ready to do so if the parent is asleep
• this causes the child to become a zombie process – defunct
but still in the system
 Foreground vs Background
• The active process(es) is in the foreground
• Background processes are those that are not
– receiving user interaction (minimized, not at the top on the desktop)
– currently part of the ready queue so that they do not receive CPU
attention
• You can move processes between foreground and background
• In Windows, you move a process from foreground to background
by
– Minimizing it
– Moving another process “on top” of it in the desktop
• In Linux, start a process in the background by adding & to the
command
– e.g., find ~ -name core* &
– Look at available processes in the shell by typing jobs
– Move jobs between foreground and background using bg # and fg #
 Example Linux Interaction
• Below is a user’s interaction in bash, moving processes between
foreground and background – comments listed after //
$ top // issue top, an interactive program, in the foreground
control+z // suspend top, prompt returns
$ jobs // ask for jobs listing
[1]+ Stopped top // + indicates the most recent fg process
fg // resumes top
control+z
$ vi // start vi
control+z // suspend vi
$ jobs
[1]- Stopped top
[2]+ Stopped vi // vi is the most recent fg process, top is next most recent
$ fg 1 // resumes top rather than vi
control+c // exit top
$ top & // launch top directly into the bg
$ // returns the prompt because there are no fg jobs
 Changing Process Priority
• Aside from moving processes between fg and bg,
the user can also impact a process by changing its
priority
– In Windows, go to task manager, in Processes tab,
right click on process name and select Set Priority
• realtime/high/above normal/normal/below normal/low
– In Linux
• You control priority by changing the process’ niceness value
– a nicer process is willing to give away some of its CPU time, so it
has lower priority (higher niceness = lower priority)
– niceness values range from -20 (least nice, highest priority) to +19
– nice process-name –n value
• NOTE: only root can lower a niceness value (raise the
priority), users are only allowed to raise the niceness value
 Process Status - Windows
• Use the task manager
– Tabs display
• Applications – start, stop, move between fg and bg
• Processes – end process (useful when a process hangs), end
process tree (end process and child processes), change
priority, set affinity (which processors or cores this process
can run on), obtain properties of process
– see the next slide
• Services – start and stop them, start the services GUI
• Performance – shown in two slides
• Networking – specifically, network performance and usage
• Users – those users logged in
Process Status – Windows Continued
 Process Status - Linux
• Like Windows, Linux has a GUI called the System
Monitor
– processes tab, resources tab, file systems information
 Process Status – Linux top
• Interactive, text-based, runs in a terminal window

Displays
system
stats at top
and individual
process stats
at the bottom
 Process Status – Linux ps
• Command line instruction to provide a “snapshot” of process
information
– Many different options available
ps vs ps aux
 Process Scheduling - Windows
• Windows task
scheduler
– Schedule the task
(action)
• start a program or
script
• send an email
• display a text
message
– The time
• daily
• weekly
• monthly
• one-time occurrence
with starting time
 Process Scheduling – Linux at
• at –f filename time [date]
• at –f filename now + value
– if you omit –f filename, you are placed at a prompt to input
one at a time the scheduled task(s)
– Once issued, you can inspect scheduled tasks using atq
– Delete scheduled tasks using atrm
– time specified as HH:MM[am|pm]
• uses military time if am or pm is not supplied
• executes within 24 hours at the next occurrence of that time
• Also available are noon, midnight and teateam (4pm)
– date specified using mmddyy, mm/dd/yy, dd.mm.yy, today
or tomorrow
–
 Process Scheduling – Linux cron
• at schedules a 1 time task, crontab schedules recurring tasks
– Create a file that contains the recurrence and task
– Issue the command crontab filename
– Inspect scheduled tasks using crontab –l
– Recurrence is indicated by 5 values: minute, hour (military time),
date, month, day of week (0 to 7, 0 and 7 for Sunday, 1 for
Monday)
• Examples:
– 15 3 1 * * -first day of every month at 3:15 am
– 0 14 * * 0 – every Sunday at 2 pm
– 0 0 12 31 * - every December 31 at midnight
• You can list multiple times using either by separating the
times by commas or using */number where number is degree
of recurrence
– */10 * * * * - every 10 minutes of every hour of every day
 Terminating Processes
• Processes usually terminate normally
– User ends the program or the program reaches its ending state
• Some processes will terminate abnormally
– Reach a run-time error such as a memory violation or
arithmetic error
– In Linux, such a process may leave behind a core dump
• the image of the process in memory so that a programmer can debug
it
• Other processes hang requiring that the user terminate
them
– Through Windows task manager’s Processes tab
– In Linux, use the kill command
• kill –s signal pid
• the signal indicates what level to kill the process, with 9 meaning
 Services
• A service is a special type of process
– Owned by the OS
– Runs in the background
• this keeps services from negatively impacting processor
performance
– Handles services from various agents (users, software, the
OS, network clients)
– In Linux, services are often called daemons
• A service will read its configuration file to determine
how to act
– We can edit configuration files to alter the service’s
behavior
• In both Windows and Linux we can establish if and
 Types of Services
• Logging
– Create log file entries for running software and operating system
operations
– Useful for determining if an event happened as planned or for
troubleshooting
• Scheduling
– Services that determine if a scheduled task should be executed
during this minute or hour
• Network
– Handling communication (incoming, outgoing), dealing with IP
addressing, dealing with specific applications (e.g., ssh, ftp, http,
email), firewall, DNS caching, etc
• File system
– Mounting remote file systems, dealing with RAID storage,
Windows Services GUI
 Windows Logging Service
• Windows event viewer displays the results of different
events
– Events are logged at one of the five levels: critical, error,
warning, information, audit success
– Below are examples of warnings and one specific event warning
 Automating Services
• In Windows, specify
– when services will run
• at system boot time
• at system boot but delayed until the rest of the system is
running
• disabled (not running)
– what to do when a service stops running (how long to wait
before restarting, how many attempts at restarting)
• In Linux, based on the runlevel, services are either
scheduled to start or stop at initialization time
– Runlevels discussed in the next two slides
• The system administrator can start or stop any service at
any time
 Linux Runlevels
• 7 Runlevel as shown below
– Each runlevel starts/stops different services
– The directory /etc/rc.d/rc#.d has symbolic links to
indicate which services start (S) and which stop
(K) (see the next slide)
 rc5.d (Runlevel 5) Directory

Structure of rc#.d directories

 Configuring Services in Linux
• Alter configuration file and restart service
– /sbin/service servicename command
• command is one of start, stop, restart, status
• Configuration files stored in /etc
– /etc/syslog.conf – to configure syslogd logging
daemon
– /etc/sysconfig/iptables – to configure Linux firewall
(ip6tables to configure for IPv6)
– /etc/resolv.conf – configure IP resolver (map IP aliases
to IP addresses by calling upon DNS servers listed in
this file)
–
 syslog.conf Entries
• This file stores entries that describe for each type
of system software, what types of events should be
logged
– Entries are of the form source:priority action
• where source is the software type, priority is the type of
message, and action is usually a log file
 Firewall Configuration Rules
• We specify firewall rules that indicate
– The direction of the message (input, output, forward)
– The result of the message (accept, reject, drop)
• reject causes the computer to respond to the sender that the
message was received but dropped, drop just drops the message
with no response
• any parameters such as port, protocol, IP source address, etc
• -A RH-Firewall-1-INPUT –p udp --dport 5353 –d
224.0.0.251 –j ACCEPT
– accept UDP packets coming in over port 5353 from IP
address 224.0.0.251
• -A RH-Firewall-1-INPUT –j REJECT
– reject all remaining messages (a backstop rule)