ksim generic wait event

Bug 16921832  Very high "ksim generic wait event" from parallel query in a RAC environment

 This note gives a brief overview of bug 16921832. 
 The content was last updated on: 08-FEB-2015
 Click here for details of each of the sections below.

Affects:

Product (Component)	Oracle Server (Rdbms)
Range of versions believed to be affected	Versions BELOW 12.2
Versions confirmed as being affected	11.2.0.3
Platforms affected	Generic (all / most platforms affected)

Fixed:

The fix for 16921832 is first included in

12.2 (Future Release) 12.1.0.2 (Server Patch Set) 11.2.0.4 (Server Patch Set) 11.2.0.3 Patch 24 on Windows Platforms

Interim patches may be available for earlier versions - click here to check.

Symptoms:	Related To:
Process May Dump (ORA-7445) / Abend / Abort Performance Affected (General) Dump in or under kjuinc Waits for "ksim generic wait event" Stack is likely to include kxfpActiveInstList	Parallel Query (Auto DOP) RAC (Real Application Clusters) / OPS Parallel Query (PQO) parallel_force_local

Description

This bug is only relevant when using Real Application Clusters (RAC) and Parallel Query (PQO)

  
Very high "ksim generic wait event" may be seen when using parallel query in a RAC 
environment with auto DOP. Also OCSSD may use high CPU.
 
Rediscovery Notes
  If the system is RAC
  AND
  parallel query is used with automatic degree of parallelism
  AND
  many sessions are seen to wait for "ksim generic wait event" 
  then this bug may be suspected.
  AND
  OCSSD keeps reporting 'clssgmGroupState: requested group state of unknown group'
 
  If an interim patch for 16554552 has been installed then an ORA-7445 [kjuinc] 
  might be seen with kxfpRefreshInst on the stack. This fix should be used in place of that fix.
 
  
Workaround
  Setting parallel_force_local=true can help avoid this issue.