[Supervisor-users] PIDs, HUPs and restarts

[Matt Black]

Hey list,

I've been toying around with restarting "celery worker" with a HUP signal,
which will shut down all current child processes, and restart with the same
args as the celery process was started.

In doing this, Supervisor sees celery is down and tries to restart it - and
fails since it restarted itself.

So really I'd like to better understand what's happening, if anyone can
assist:

- How does supervisor track processes?
- Can I not provide a "pidfile" path to supervisor to watch/use?
- Would the inclusion of this pull
request<https://github.com/Supervisor/supervisor/issues/179>enable
supervisor to restart celery with a HUP and keep track of it?

For now, I'll avoid HUP and use supervisor directly to restart celery :)

Cheers

[Matt Black]

Thanks Jonathan. I've already read the most of the docs - admittedly I
missed that part of the introduction though.

In the last few days my testing with HUPs got me nowhere, so I've already
ported my code to use supervisorctl to restart celery (and several others)
when necessary.

For anyone else pondering celery, this is useful reading from the author -
seems restarting the process with supervisor is correct:
http://serverfault.com/a/471377/89669


On 27 August 2013 07:52, Jonathan Vanasco <***@2xlp.com> wrote: