Kill Command

Quick Reference

 # kill process
 $ kill 
 # hangup process
 $ kill -1 
 # terminate process
 $ kill -15 

Run in Warp

Determining which version of kill your shell is using

When working with kill, the first step is to make sure your shell is running the code you think it is. Both Bash and Zsh shells have their own built-in version of kill that may take precedence over the Linux binary kill. So you may need to run /bin/kill instead of kill in order to utilize the standard-library Linux binary kill; however, the functionality of the shell built-in kill will likely be largely the same.

For the rest of this article we will use the syntax kill …, and assume that we are referring to Linux kill. To see which one will be run when you run a kill  command, you can run which kill as follows:

Structure of a kill command

The Linux kill command has the following structure:

 kill [OPTIONS] [PID]...

Run in Warp

kill signals available

“Options” in the structure above has a number of components, with the most commonly used option being the signal flag. To see a list of the signals available we can run the following in our shell:

 kill -l

Run in Warp

Each of these signals is also associated with a given number, as again defined in the Linux docs. Depending on the configuration of your shell, you may also be able to run kill -L to see both the signal names and numbers.

Either the signal name or the signal number can be combined with the kill command as follows:

 $ kill -9 
 $ kill -s SIGKILL 
 $ kill -SIGKILL 
 $ kill 

Run in Warp

The signal here (SIGKILL) happens to be the default signal, and will also be used if no specific signal is passed.

kill to check user access to a specific process

SIGNULL (0), checks user access to or validity of a pid without sending an action signal.

 $ kill -0 
 $ kill -s SIGNULL 
 $ kill -SIGNULL 

Run in Warp

kill to terminate a process gently

SIGTERM is the “normal” kill signal. The application will be given time to shut down cleanly (save its state, free resources such as temporary files, etc.), and an application that is programmed to not immediately terminate upon a SIGTERM signal may take a moment to be terminated.

 $ kill -15 
 $ kill -s SIGTERM 
 $ kill -SIGTERM 

Run in Warp

kill to force termination

SIGKILL can be used to force termination and the application is not given a chance to respond.

 $ kill -9 
 $ kill -s SIGKILL 
 $ kill -SIGKILL 

Run in Warp

Determining the PID of a process

The second part of the kill command is the process ID, or “pid”. There are a number of ways we can determine which processes are running and their pid’s, including ps, pidof, and pgrep and top. Which to use will depend on exactly what you are trying to do.

For example, to get the pid of all processes related to Google Chrome you can run the following:

 $ pidof chrome

Run in Warp

Alternatively, to list all processes running on your machine, consider ps aux or top:

kill -9 -1 terminates all running processes

However, this should be done very carefully; this can cause damage to your kernel depending on what is running on your machine at the time.

Combining kill with pidof

kill is most powerful when combined with other terminal commands. For example, to identify and kill all processes related to Google Chrome we can run the following command:

 kill $(pidof chrome)

Run in Warp

Overcoming kill permission issues with sudo

If any of kill commands fail one reason might be lack of permissions. Simply re-run the command using sudo and enter your password in order to proceed:

 # Careful! Using `sudo` can be dangerous when 
 # dealing with system processes
 sudo kill $(pidof chrome)

Run in Warp