Processes and Jobs

Introduction

Every running program is a process. Each has an ID number (PID), an owner and a state. Knowing how to see and control processes is vital when something hangs or uses too much memory.

Seeing processes

ps                 # processes in your current terminal
ps aux             # every process, with owner, CPU and memory
ps aux | grep node # find a particular program
top                # live view, press q to quit

In ps aux, useful columns are USER, PID, %CPU, %MEM and the COMMAND. Many systems also offer htop, a friendlier version of top.

To find a process ID by name:

pgrep -l firefox

Stopping processes

kill does not always "kill". It sends a signal.

kill 1234          # polite request to stop (SIGTERM)
kill -9 1234       # forceful stop (SIGKILL)
pkill firefox      # by name

Always try the normal kill first. It lets the program clean up files and save data. Use kill -9 only when it will not stop, because the program gets no chance to tidy up.

Foreground, background and jobs

A command that takes a long time blocks your terminal. Add & to run it in the background:

sleep 300 &
jobs               # list background jobs in this shell
fg %1              # bring job 1 to the foreground

To pause a running command, press Ctrl+Z. Then continue it in the background with bg, or bring it back with fg. Ctrl+C interrupts and stops it.

Closing the terminal normally ends its jobs. To keep a program running after you log out, use nohup:

nohup ./long-task.sh > output.log 2>&1 &

2>&1 sends error messages to the same place as normal output. For serious long-lived work, tools such as tmux or a system service are better.

Exit codes

When a command finishes, it returns a number. 0 means success, anything else means a problem. Check the last one with:

ls /nowhere
echo $?

On most Linux systems the output of echo $? is 2 here, because the folder does not exist. Scripts depend on exit codes, as you will see in the advanced course.

Chaining commands

mkdir backup && cp notes.txt backup/     # second runs only if the first succeeded
command1 || command2                      # second runs only if the first failed
command1 ; command2                       # run both regardless

Try it yourself

Start sleep 300 &, find it with ps aux | grep sleep, then stop it with kill and its PID. Check with jobs that it is gone.

Summary

• ps, top and pgrep show processes, and each has a PID.
• kill sends signals. Try the polite one first, and -9 as a last resort.
• &, jobs, fg, bg and nohup manage background work, and exit code 0 means success.