SYMPTOMS
Database hangs with excessive waits for 'cursor mutex S'
AWR report shows:
Top wait event: 'cursor: mutex S' and 'library cache lock'
high version count for an insert into sys.aud$ ....
Trace file from event errorstack of session that was waiting shows a call stack similar to the following:
kksMutexWait kgxWait kgxShared kkshGetNextChild kkscsSearchChildLis kksfbc kkspsc0 kksParseCursor ....
current sql:
insert into sys.aud$( sessionid,entryid,statement,ntimestamp#, userid,userhost,terminal,action#,returncode, obj$creator,obj$name,auth$privileges,auth$grantee, new$owner,new$name,ses$actions,ses$tid,logoff$pread, logoff$lwrite,logoff$dead,comment$text,spare1,spare2, priv$used,clientid,sessioncpu,proxy$sid,user$guid, instance#,process#,xid,scn,auditid, sqlbind,sqltext,obj$edition,dbid) values(:1,:2,:3,SYS_EXTRACT_UTC(SYSTIMESTAMP), :4,:5,:6,:7,:8, :9,:10,:11,:12, :13,:14,:15,:16,:17, :18,:19,:20,:21,:22, :23,:24,:25,:26,:27, :28,:29,:30,:31,:32, :33,:34,:35,:36)
CHANGES
CAUSE
Issue is similar to that reported in:
Bug 11936699 Abstract: WAIT TIME OF LIBRARY CACHE LOCK INCREASES DUE TO MANY CHILD CURSORS OF SYS.AUD$
SOLUTION
Apply patches: 10151017, 9944129, and 10636231 if available for version / platform
OR
Apply 11.2.0.3 patchset which includes all 3 fixes
|