Hi,

I have big problem with uWSGI. From time to time (in rush hours) uWSGI
just hangs and throws such messages:

Wed Sep 14 19:01:51 2011 - *** uWSGI listen queue of socket 3 full !!!
(129/128) ***
Wed Sep 14 19:01:52 2011 - *** uWSGI listen queue of socket 3 full !!!
(128/128) ***
Wed Sep 14 19:01:55 2011 - *** uWSGI listen queue of socket 3 full !!!
(129/128) ***
Wed Sep 14 19:01:56 2011 - *** uWSGI listen queue of socket 3 full !!!
(129/128) ***
timeout. skip request.

After restart - everything is fine. We have two servers-workers, both
are stalled in same time. How can I track what is causing this
problem?

Our config.

<uwsgi>
  <socket>192.168.1.100:9001</socket>
  <module>wsgi</module>
  <processes>48</processes>
  <listen>128</listen>
  <master/>
  <pidfile>/tmp/uwsgi.pid</pidfile>
  <uid>1000</uid>
  <gid>1002</gid>
  <disable-logging/>
  <binary-path>/usr/bin/uwsgi</binary-path>
  <limit-post>10485760</limit-post>
  <max-requests>1000</max-requests>
  <harakiri>90</harakiri>
</uwsgi>

We are running Django application. I thought that maybe PostgreSQL is
blocking, but there is no big load on DB server, and there are no "not
granted locks". We are using NFS share too, for static (generated by
uWSGI -> ImageMagick via subprocess.call()) files.

uWSGI 0.9.9.1, but problem exists on LTS version too.

-- 
Best regards,
Szymon
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi

Reply via email to