MySQL CPU Usage

[wdimmit]

This problem started occurring sometime while I had 4.0.2 and was not corrected by upgrading to 4.0.5. The host OS is CentOS 4.3 running 2.6.9-42.0.3.

The problem is that the MySQL predominantly uses all available CPU resources, resulting in a server load level of 2.5+ even when all else is quiet. I've not been able to find anything conclusive on the forums or wiki. The mysqld.slow log is huge, with a wide variety of queries included. It is acting almost as though it has lost its indexes and is having to do scan lookups for every action.

Any suggestions would be greatly welcomed.

Thanks,
Will

[bburtin]

Which MySQL instance is the problematic one, the one used by the mailbox server or by logger? You can find out by running ps -ef and grepping for the process id in question. If it's logger, try disabling the logger task in the zimbra user's crontab.

[bobby]

Comment out this line to prevent the log processing that populates the logger db:

crontab -l | grep logproc
00,10,20,30,40,50 * * * * /opt/zimbra/libexec/zmlogprocess > /tmp/logprocess.out 2>&1

To completely disable the logger service:

zmprov ms `zmhostname ` -- -zimbraServiceEnabled logger
zmloggerctl stop