I'm writing a shell script that's supposed to function on various different distributions, some of which are using busybox and some are not.

This script uses the timeout command to figure whether a command it executed ran longer than <time> or not. In the case of distributions which use the coreutils timeout this is no problem because when timeout had to kill a command it always returns 124, so that's simple.

My problem is that when I run the same on Alpine Linux (which uses busybox ) then the exit status is 0 after timeout had to kill the command, but when I use the timeout command from the busybox package in Ubuntu it exits with 143.

Why do the two timeout commands, which are both from busybox, give me a different return value?

Is there any consistency or rule in that which I can use to determine whether the command has timed out or not?


Alpine Linux probably has an earlier version of busybox. A simple solution is to add to your original command another command that has some side effect you can test for, eg writing some output to stdout or a file.

Eg, assuming you want a timeout of 1 for an original command of sleep 2, instead of

timeout 1 sleep 2


ok=$(timeout 1 bash -c 'sleep 2; echo ok')

and test $ok for ok. Obviously, if your command writes to stdout you need to redirect it, eg dup it to fd 3 and redirect to that:

exec 3>&1
ok=$(timeout 1 bash -c 'my command >&3; echo ok')
exec 3>&-

or do the echo ok into a file.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.