Kill Linux Processes with pkill

One of the best features in Linux is the way you can control processes from the command line, so if you have an application that locks up your GUI, you can always SSH over from another machine and just kill the offending process.

The problem is that if you are killing the same process repeatedly, it’s very tedious to have to figure out the process ID every single time so that you can kill it… so here’s the easier way to do it.

The Old Way

The classic way of killing processes meant you’d first need to use the ps command piped through grep to find the process you are trying to kill:

$ ps -ef | grep swiftfox
geek 7206 22694 0 Dec04 ? 00:00:00 /bin/sh /opt/swiftfox/swiftfox
geek 7209 7206 0 Dec04 ? 00:00:00 /bin/sh /opt/swiftfox/run-mozilla.sh /opt/swiftfox/swiftfox-bin
geek 7213 7209 0 Dec04 ? 00:04:29 /opt/swiftfox/swiftfox-bin
geek 14863 14224 0 18:19 pts/4 00:00:00 grep swiftfox

Then to kill the process, you’d have to use the kill command:

$ kill 7206

The New Way

Instead of going through all of that, you can simply use the pkill command if you already know the process name or part of it.

$ pkill swiftfox

It’s as simple as that. You should note that pkill will kill all processes matching the search text, in this case swiftfox

If you want to see what process names are matched before using the pkill command, you can use the pgrep command. Passing the -l switch tells pgrep to show the process name as well.

$ pgrep -l swiftfox
7206 swiftfox
7213 swiftfox-bin

Monitoring Processes with Kill

If you have a process ID but aren’t sure whether it’s valid, you can use the most unlikely of candidates to test it: the kill command. If you don’t see any reference to this on the kill(1) man page, check the info pages. The man/info page states that signal 0 is special and that the exit code from kill tells whether a signal could be sent to the specified process (or processes).

So kill -0 will not terminate the process, and the return status can be used to determine whether a process is running. For example:

$ echo $$ # show our process id
12833
$ /bin/bash # create new process
$ echo $$ # show new process id
12902
$ kill -0 12902
$ echo $? # exists, exit code is 0
0
$ exit # return to previous shell
$ kill -0 12902
bash: kill: (12902) - No such process
$ echo $? # doesn’t exist, exit code is 1
1

Many UNIX dæmons store their process IDs in a file in /var/run when they are started. Using kill -0 to test the pid is a lot easier than parsing ps output. For example, to test whether cron is running, do the following:

# kill -0 $(cat /var/run/cron.pid)
# echo $?
0