Il giorno 01/ott/2012, alle ore 19:40, Ryan Showalter <[email protected]> ha scritto:
> Hello, > > One of my @cron decorated functions with target='mule' has not been > executing the last few days and I think I've traced it down to an > issue where mules don't like low-memory environments. > > Workers and spool jobs still function nominally, and if I remove the > target parameter, things seem to work again. While debugging this > issue, the issue disappeared when I reduced the number of > workers/mules.. and then afterwards, the issue still did not appear > when I brought the levels of mules/workers to their typical level. > Only after I put some load on the server did the @cron mule begin to > fail with the message below. > > Mon Sep 24 08:45:13 2012 - uWSGI mule 3 braying: my master died, i > will follow him... > OOOPS mule 3 (pid: 3869) crippled...trying respawn... > spawned uWSGI mule 3 (pid: 4546) > > I'm wondering if this might somehow be avoided? > The master is dying, this is pretty ugly, and spoolers as well as workers should follow him pretty soon. Can you check if the master is really dead ? Is a traceback available when the mule dies (or when the master dies, if it happens) in the logs ? -- Roberto De Ioris http://unbit.it JID: [email protected] _______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
