Job stuck after receiving the SIGTERM and not killed within the max run time

[jasonlkh-wa]

[2025/01/23 12:35:01] <stdout of my job>
Caught SIGTERM, killing child: 75973

# Job failed at 2025/01/23 12:44:43 (GMT+0).
# Error: Job Aborted: Exceeded maximum run time (1 minute)
# End of log.

There are 3 things confusing me in this log,

1. the process id is 75972 in the cronicle UI, but it's sending a SIGTERM to 75973
2. after getting a SIGTERM, the job stucks there
3. it does not get killed within 1 mins which is the maximum run time I configured, but only after 10mins.

Note that I could not consistently replicate the issue since the job works fine 98% of the time. So it's a bit random when it happens and it also consumes 39.9% of CPU while on average this job consumes 0%.

[jhuckaby]

Well, I think I can answer your first question at least. When using the Shell Plugin, which I assume you are doing, there is a second process in play. Cronicle spawns a Shell Plugin controller process first (that's the PID shown in the UI), and then the plugin launches your process from there. When a job is aborted, the Shell Plugin sends a SIGTERM to the next child process, so that's why you're seeing that second PID in the log.

I cannot explain your other issues here, as I don't know anything about your setup, server hardware / software, or versions of anything. If a process doesn't respond to a term signal, Cronicle should only wait for the child_kill_timeout before sending a SIGKILL. This timeout defaults to 10 seconds.

Your third item makes no sense to me at all.

I'm sorry I could not be of much help. This is a very confusing problem, and not one I have ever encountered before.