'row cache lock' Causing Slowness On 11.2 Databases

Many Processes Waits Too Short For 'row cache lock' Causing Slowness On 11.2 Databases [ID 1162566.1]




Applies Oracle Server - Enterprise Edition - Version: 11.2.0.1 to 11.2.0.2 - Release: 11.2 to 11.2
Information in this document applies to any platform.

Critical database slows down with 'row cache lock' as the highest event in the "Top 5 timed events" in the AWR report.

Cause

Bug 9875364: ROW CACHE LOCK AT DC_USERS DURING LOGIN
Base Bug 9720182: DUE TO ROW CACHE LOCK WAIT EVENTS IN DATABASE APPLICATION GOT HUNG.

SOLUTION:
This bug will be fixed in the expected 12.1 database version.
Please apply Patch 9720182.


It is not clear which process holds this object in an incompatible mode. This is most probably because it has been released quickly, so when we came to dump the holder process, it was no longer holding it but released it.

Looked into other processes which are waiting for 'row cache lock' and found that "seconds since wait started" is too short. As the above sample process 241 shows, this was only 0.203345 seconds.

The call stack of process 241 includes:
kslwaitctx kqrget kqrLockAndPinPo kqrpre1

This looks like Bug 9875364 which was marked by the development team as duplicate of Bug 9720182.