Confirmed.
As a workaround we watch the queue and restart celery if it grows too
big ... or if nobody is listening.
I found evidence that this is a known problem with Celery 2. The
solution should be to upgrade to Celery 3. The API did however change so
it requires some amount of work.
So far the workaround has been good enough for us and it hasn't made it
to the top of my todo list.
/Mads
On 07/25/2016 02:01 PM, Thomas De Schampheleire wrote:
Hi,
We are running Kallithea (some older version, 0.2.1 based) with
Celery, and sometimes the Celery workers just hang. This is observed
by e-mails not being received anymore. Restarting celery, or sending
'kill -SIGUSR1' to one of them, resolves the problem.
There are several reports on the internet about such problems, but
could be due to various reasons. Before digging deeper, I would like
to hear if some of you have or are experiencing the same thing, and
whether you have solutions.
This is celery 2.2.10, librabbitmq 1.6.1 and RabbitMQ 3.5.0.
Thanks,
Thomas
_______________________________________________
kallithea-general mailing list
[email protected]
http://lists.sfconservancy.org/mailman/listinfo/kallithea-general
_______________________________________________
kallithea-general mailing list
[email protected]
http://lists.sfconservancy.org/mailman/listinfo/kallithea-general
