Celery jobs throw exceptions about django-postgrespool HOT 7 CLOSED

@ask any ideas?

from django-postgrespool.

I actually had the same issue when running an SQLAlchemy app under uwsgi. The problem in that case was that uwsgi forked the workers which resulted in sqlalchemy using the same sockets between the workers. In uwsgi this was solved with the -lazy flag. See http://lists.unbit.it/pipermail/uwsgi/2011-May/002079.html

I'm pretty sure this is the same kind of problem.

from django-postgrespool.

What @mvantellingen says sounds about right! You should use multiprocessing.util.register_after_fork or something.

(though celery uses the billiard fork of multiprocessing which now has its own after fork registry (billiard.util.register_after_fork, I totally forgot about that, and it should really use the multiprocessing one for library consistency)

from django-postgrespool.

Interesting, we're using gunicorn. I don't see any flags that may help with that but it's a start.

from django-postgrespool.

@brockhaywood you solved this problem? I'm get this errors in my sentry just now :-(
I'm also use gunicorn.

from django-postgrespool.

No, I was unable to resolve.

On Wed, Feb 13, 2013 at 12:08 PM, Denis Veselov [email protected]:

@brockhaywood https://github.com/brockhaywood you solved this problem?
I'm get this errors in my sentry just now :-(

—
Reply to this email directly or view it on GitHubhttps://github.com/kennethreitz/django-postgrespool/issues/3#issuecomment-13515303.

• bh

from django-postgrespool.

I'm about to push this to production, our celery tasks are intermittent and non-critical, so I'll add any information we get out of this here.

from django-postgrespool.