Thursday, October 1, 2009

Tip#26 OMS generates lot of core.xxx files

Recently we had disk full issue on our Grid Control OMS server, firstly we thought it must be just few backup copies which are lying around. After cleaning up the disk of unwanted backups got some breathing space (atleat thats what we thought!) but soon found out it is full again in like an hour.

So digging further found out that OMS was generating HUGE amount of logs (like core.xxx) and at very brisk speed too. Metalink Doc ID 419999.1 says it is due to "the access_log for the http server of the OMS is over 2Gb in size and this is causing the http server to core dump." and to fix it we need to stop all OMS process, remove the log file and startup the OMS process. So basically,

/opmn/bin/opmnctl stopall

rm /Apache/Apache/logs/access_log

/opmn/bin/opmnctl startall

After doing that all seems to be back to normal. To avoid the same in future, as suggested in the metalink, Consider rotating the Apache logs on a regular basis as part of the routine maintenance of the system. This can be done while the OMS is down during monthly maintenance tasks.

No comments: