watchdog kill bugs about nhc HOT 2 CLOSED

Thanks for your report, Walter!

The code is correct as written. The kill command in bash returns 0 on success, yes, but success in this case doesn't mean the process went away. It simply means that the signal was successfully delivered to the process(es) specified. From the bash(1) man page:

kill returns true if at least one signal was successfully sent, or false if an error occurs or an invalid option is encountered.

In particular, "successfully sent" means that there was a process with the specified PID which was able to receive the signal. A non-zero return value, for our purposes, means that the specified process is no longer there to receive the signal -- i.e., it's already been successfully killed.

NHC uses the || return 0 you noted in the code to handle the case where either the process went away on its own, or NHC's previous attempt(s) to terminate it were successful, so it doesn't need to continue trying to kill (i.e., signal) it. If the invocation of kill succeeds, then the process is still there, so NHC might not have successfully gotten it to die yet and must therefore keep trying. Just because kill returns success doesn't mean the process actually died, though. It only means that the signal was delivered.

Make sense?

from nhc.

Yes, so true.

Thanks!

On Wed, Mar 30, 2016 at 8:50 PM, Michael Jennings [email protected]
wrote:

Thanks for your report, Walter!

The code is correct as written. The kill command in bash returns 0 on
success, yes, but success in this case doesn't mean the process went away.
It simply means that the signal was successfully delivered to the
process(es) specified. From the bash(1) man page:

kill returns true if at least one signal was successfully sent, or false if an error occurs or an invalid option is encountered.

In particular, "successfully sent" means that there was a process with the
specified PID which was able to receive the signal. A non-zero return
value, for our purposes, means that the specified process is no longer
there to receive the signal -- i.e., it's already been successfully killed.

NHC uses the || return 0 you noted in the code to handle the case where
either the process went away on its own, or NHC's previous attempt(s) to
terminate it were successful, so it doesn't need to continue trying to
kill (i.e., signal) it. If the invocation of kill succeeds, then the
process is still there, so NHC might not have successfully gotten it to die
yet and must therefore keep trying. Just because kill returns success
doesn't mean the process actually died, though. It only means that the
signal was delivered.

Make sense?

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#13 (comment)

If a woodchuck could chuck wood, and a woodchuck would chuck wood,
how much wood would a woodchuck chuck?

from nhc.