I configured my watchdog.conf, and i want try if my computer restart when it crash. so, I wanted kill my Watchdog to try if my computer will restart.

screenshot of pid of watchdogd

But I try to killed it with the command "kill -9 90" and "killall -s SIGKILL watchdog" but neither of two successfulled.

If someone have an idea to kill the watchdogd ?

Best regards.

  • Please don't post pictures of text. Jun 24, 2022 at 7:54
  • You need to check what module it is using, if any, then rmmod it. This is the only way to stop it from userland.
    – Bib
    Jun 24, 2022 at 11:18

2 Answers 2


In general, tasks that ps shows enclosed by [brackets] are kernel threads.

There is just no way to kill a kernel thread from userland. Period.

The only way I know is to make it crash writing to /dev/mem with more or less predictable consequences. (Care! going that way, the probability of crashing something else than what you precisely want is almost… certain)

In the particular case of the watchdog daemon, its behaviour can nevertheless be directed using some options, via some configuration file.
It can even be tested, see the manpage (under Test directory paragraph).

  • Unless you rmmod the module.
    – Bib
    Jun 24, 2022 at 11:19

A better way to test the watchdog demon is by suspending it. You can achieve that using the kill* command.

killall -STOP watchdog
Thiswill suspend the watchdog process unless it gets SIGCONT and the watchdog timer eventually reset the computer after the timeout.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .