sorry this cause unit test failure :)

Is just a lambda with the logging call would handle that SIGTERM correctly?!

No, the lambda is what happens when the sigterm is receive. https://docs.python.org/2/library/signal.html
On Fri, 2 Aug 2019 at 15:02 Islam Mohamed> wrote: Is just a lambda with the logging call would handle that SIGTERM correctly?! — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#5423?email_source=notifications&email_token=AEBNDBMMVL24BWZC725TEG3QCQO4ZA5CNFSM4HC4V3P2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3NRMLQ#issuecomment-517674542>, or mute the thread https://github.com/notifications/unsubscribe-auth/AEBNDBIMKHVHYC6CEOGBWZLQCQO4ZANCNFSM4HC4V3PQ .

Understood, Maybe my comment wasn't clear. I'm wondering how the signal handler which is doing nothing but logging would help to handle the SIGTERM for tasks correctly?

No, the lambda is only printing the log for the user
signal.signal(signal.SIGTERM, handle_sigterm)
This is the code such handling the SIGTERM

No, the lambda is only printing the log for the user
signal.signal(signal.SIGTERM, handle_sigterm)
This is the code such handling the SIGTERM

I think he's asking why you're not doing anything besides logging. I would expect the worker to exit after finishing the current task but it looks like it would just log this and continue running indefinitely.

No, the lambda is only printing the log for the user
signal.signal(signal.SIGTERM, handle_sigterm)
This is the code such handling the SIGTERM

I think he's asking why you're not doing anything besides logging. I would expect the worker to exit after finishing the current task but it looks like it would just log this and continue running indefinitely.

Exactly, logging wouldn't be enough and considered as a solution for handling SIGTERM properly

Let's breathe...

1. Worker currently supports warm shutdown on SIGTERM
2. The issue was that when the worker received SIGTERM he sent this SIGTERM to current running tasks.
3. When the task received SIGTERM no one handled it and the task stopped immediately.

Now after my fix 1 and 2 happens as it happens previously.
My fix handles the SIGTERM and not killing the running task and the worker is waiting till the task finished. After that worker goes to shutdown.

Let's breathe...

1. Worker currently supports warm shutdown on SIGTERM
2. The issue was that when the worker received SIGTERM he sent this SIGTERM to current running tasks.
3. When the task received SIGTERM no one handled it and the task stopped immediately.

Now after my fix 1 and 2 happens as it happens previously.
My fix handles the SIGTERM and not killing the running task and the worker is waiting till the task finished. After that worker goes to shutdown.

Thank you so much for explaining it, Makes total sense 👍