Den 10/10/2010 kl. 17.55 skrev Massimiliano della Rovere: > > * Which process(es) on the server is using 100% CPU time? > directly from htop: > PID USER PRI NI VIRT RES SHR S CPU% MEM% TIME+ Command > 544 mader 20 0 969M 953M 1444 R 96.0 47.4 50:09.85 python > ./manage.py celeryd -l warning
This, and the fact that you have no errors or seriously long queries in your slow query log, indicates that the MySQL server is operating fine. Something seems to be timing out on the MySQL side. I'm not sure if you mentioned it, but are these InnoDB tables? The stack trace you posted is in the rollback of a transaction. It's possible that everything is running within the same transaction which MySQL is shutting down at some point. Try this: SQL> show variables where variable_name LIKE '%timeout%'; to see the values in your current server instance. If nothing suspicious turns up, I guess it's time to follow the stack trace into the celery source code or where your own code calls celery and add some debugging. Maybe it's a specific query that trips the code. Maybe it's a specific number of queries. Maybe it's at a specific timespan after connecting to MySQL. Thanks, Erik
smime.p7s
Description: S/MIME cryptographic signature
