Linux Processes

Processes are running instances of programs in Linux.

The ps Command

ps stands for process status and shows the running processes on a system.

To invoke it, run ps and it will display the following information:

  ps -ef

The option -e will display all the processes, and the -f option will display in a full format listing.

  ps -f -u <user>

It displays the process that belongs to user. When you have multiple usernames, separate them using a comma.

  ps -eo comm,etime,user | grep httpd

It shows the command, time, and user(s) related to the "httpd" service. You can replace "httpd" with the service you are looking for.

  ps -eo pmem,pid,cmd | sort -k 1 -nr | head -5

It displays the top five of the output, organized in three columns with the memory that a process is taking, process ID, and the command, sorted by memory usage.

  pstree -np | less

The option -p shows process identification numbers (PIDs) and -n sorts its output in the order of the PIDs.

If you want to see the process tree of any specific user, run pstree <user>. Use your username instead of <user>.

  pmap 1232

It displays the memory usage map of a process 1232. If you need information for multiple processes, you can add the their PID separated by a space.

Killing Processes

All processes in Linux respond to signals. Signals are an OS-level way of telling programs to terminate or modify their behavior.

  kill 1734

This terminates a process with a PID of 1734.

If this fails, the stronger signal 9, called SIGKILL can help by doing kill -9 1734. To see all the options, run kill -l

In case you cannot determine the number of the process, you can use the name of the program to make it stop: kill -9 firefox

  killall firefox

In this case killall is closing a current program(s) that is running a process called firefox.

Prioritizing Processes

Linux schedules the process and allocates CPU time accordingly for each of them, but you can set the priority to get more CPU time by using the nice and renice commands.

The process scheduling priority has a nice value that ranges from -20 to 19. The highest priority will consume a lot of CPU and that is not nice, so we set it as -20. On the other hand, the least priority for a process is represented as nicer because it will not take much CPU resources, and a nice value of 19 then is set.

Only the root user can set a negative value. A nice value of a process can be seen in the column NI after you type top in your terminal.


It displays the main 30 processes on the system sorted by CPU utilization, memory usage, and routine. See more information here.

If you want to sort processes by CPU usage, you can do so with top -o %CPU.

To see a list of processes of any user, use top -u <user>. Remember to replace <user> with your username or root.

  nice -n 10 apt-get upgrade

It sets a positive 10 as a nice value that gives less priority to a process.

  renice 10 -p 2187

It sets a priority of 10 to a process with an ID 2187. If its value was 0, you are lowering the priority.

📝 Note: You can set the default nice value of a particular user or group in the /etc/security/limits.conf file, by using the syntax: [username] [hard|soft] priority [nice value], e.g. backupuser hard priority 1.

Background Processes

A background process executes independently of the shell, without user intervention, leaving the terminal free for other work.

This means that you do not have to wait for a command to finish in the terminal to run another one. For further information, click here.

After using commands to run process in the background, you will immediately be returned to the shell, and you will see the shell prompt.

  ./ &

The file ./ is forked and runs in a separate sub-shell as a job. A process's job number and its PID will be displayed and stored in a special variable $!. This can be seen later with echo $!.

  nohup ./ &

The output generated by ./ will be saved in nohup.out in the current directory. If you logout, your process will not get killed.

To run more scripts at the same and leave them to be finished in background, run ./ & ./ & ./ &.


This creates a new session when you log into another server. A screen ID is displayed after running your command(s).

To create a screen session with a name, run screen -S name. See more screen options on man screen or here.

To detach from the screen session with CTRL+A+D or if you are remotely logged in, you can do it with screen -d [SCREENID].

    [1]+  Running                 ./ &

It displays the list of the current jobs that are running in the background; there is the script ./ with the job number:1.

  dd if=/dev/zero of=myfile bs=1K count=2048000
    [1]+ Stopped dd if=/dev/zero of=file bs=1K count=2048000
    bg %1

The number 1 is the ID of the job as viewed under a job suspended; then, to use it with bg it must be preceded with a %.


Without any argument, fg runs the current job in foreground.

To see the ID of the jobs that are running in the background to bring them to the foreground, type jobs. Then type the ID preceded by a%, e.g. fg %1.

    [1]+  Stopped       
    [1]+ &
    disown %1

The ./ file is executed, then this job is suspended by pressing ctrl+Z, followed by bg to make it run in the background. Then, by typing disown %1, the job won't get the SIGHUP signal to be shut down.

  sleep 3h; mplayer game.mp3

This will wait three hours to play game.mp3.

You might consider using m to set minutes, e.g.sleep 10m ; your_script, or d, for days. If you do not specify anything, the sleepy action will happen in seconds. & & &

The three scripts of "collection" that are running in the background will finish before the process-job-output starts.

wait ensures this process and asks to not exit the containing script until all the execution has finished.