Skip to main content Link Search Menu Expand Document (external link)

Managing Processes

Jobs

  • Groups of processes, created from a command (from user f.ex)
  • Can be run in:
    • Foreground: run directly from the shell, and when one job is running, other jobs need to wait for shell access
    • Background: Attaching a &. Gets a lower priority, releases the shell to other uses.
    • By default, jobs run in foreground. bg and fg can be used to run jobs after suspending them (Ctrl + z).
    • Background jobs can be shown with jobs

Monitoring

To list and monitor processes, two main tools are used.

ps

  • Provides information about processes
  • Has several options to control its output. For example 
      ps #Shows processes in current shell
  ps -ef #Shows all processes in the system with full format
  ps -eLf #Same as above + threads and thread info
  ps aux #BSD Argument style, shows all processes of all users
  ps aox pid,ppcpu #BSD Argument style, shows only PID and %CPU
  • A variant is pstree which shows all the processes in the system, and the threads and their relation (parent-child) in tree form.

top

  • Interactive tool to monitor and control processes

  • Has “summary” lines at the beginning, showing:
    • Uptime and Load Average
    • Process count and their state
    • %CPU for user processes (us), for system processes (si).
    • The percentage of user jobs running at a lower priority (ni), Idle mode (id), the percentage of jobs waiting (wa) for I/O is listed, the percentage of hardware (hi) vs. software interrupts (si) and steal time (st) (generally used with virtual machines)
    • Memory information (RAM and Swap, which is used when RAM is full)
  • Shows processes with the following categories as default:
    • PID, USER, Priority (PR), Niceness (N), Virtual (VIRT), physical (RES), and shared memory (SHR), Status (S), %CPU, %MEM, Execution time (TIME+) and COMMAND.
  • Using interactive commands, more information can be obtained, processes can be sorted, prioritized or killed:
    • f: Change what is displayed and the sort parameter
    • r: Renice a process
    • k: Kill a process
    • 1: Show all CPUs in summary
    • Many others

Load Average

  • Is the average system load on a PC for a defined period of time
  • Takes into consideration process RUNNING, in run queue, or in SLEEP state
  • Can be read in the output from w, uptime and top
  • Interpreted like (load average: 0,20, 0,58, 0,67, for a 1 CPU PC):
    • For the last minute, 20% of the CPU has been utilized in avg.
    • For the last 5 minutes, 58% of the CPU has been utilized in avg.
    • For the last 15 minutes, 67% of the CPU has been utilized in avg.
  • For more CPU all numbers are divided by the number of CPUs

Kill

To kill processes (only the ones from the current user, otherwise admin acesss is needed):

kill -SIGKILL <PID>
kill -9 <PID>

Scheduling processes to execute

at

  • Can be used to schedule when in the future a job should run
  • For example:
at now + 2 days # Time to execute
at> cat file1.txt # Tasks to perform
at> <EOT> #Press Ctrl+D to submit

cron

  • Tool for scheduling background jobs (at specific times or periodic)
  • Driven by a configuration file at /etc/crontab (cron table)
  • There are both system-wide crontab files and individual user-based ones
  • Can be edited with crontab -e
    • Each line has 6 parameters: <MIN> <HOUR> <DOM> <MON> <DOW> <CMD>
    • For example * * * * * /usr/local/bin/execute/this/script.sh will schedule a job to execute script.sh every minute of every hour of every day of the month, and every month and every day in the week.
    • 30 08 10 06 * /home/sysadmin/full-backup will schedule a full-backup at 8.30 a.m., 10-June, irrespective of the day of the week.

sleep

  • Used to delay a command (make it wait before finish)
  • By default accepts seconds but with suffixes can accept anything:
sleep 2 # 2 seconds
sleep 2s # 2 seconds
sleep 2m # 2 minutes
sleep 2h # 2 hours
sleep 2d # 2 days