Bug 18199537 RAC database becomes almost hung when large amount of row cache are used in shared poo

Bug 18199537  RAC database becomes almost hung when large amount of row cache are used in shared pool

 This note gives a brief overview of bug 18199537. 
 The content was last updated on: 15-FEB-2017
 Click here for details of each of the sections below.

Affects:

Product (Component)Oracle Server (Rdbms)Range of versions believed to be affectedVersions BELOW 12.2Versions confirmed as being affected

• 11.2.0.4
• 11.2.0.3
• 11.2.0.2

Platforms affectedGeneric (all / most platforms affected)

Fixed:

The fix for 18199537 is first included in

• 12.2.0.1 (Base Release)
• 12.1.0.2 (Server Patch Set)
• 11.2.0.4.4 (Oct 2014) Database Patch Set Update (DB PSU)
• 11.2.0.4 Bundle Patch 10 for Exadata Database
• 11.2.0.4 Patch 10 on Windows Platforms

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

Symptoms:

Related To:

• Latch Contention
• Performance Affected (General)
• Waits for "row cache cleanup"
• Waits for "latch: row cache objects"
• Waits for "latch: shared pool"
• Stack is likely to include kghfrunp

• RAC (Real Application Clusters) / OPS

Description

This bug is only relevant when using Real Application Clusters (RAC)

A RAC instance may sometimes become unresponsive when the instance startspurging a large amount of row cache objects from shared pool. Rediscovery NotesIf you cannot connect to RAC instance, or see severe performance problem,and AWR reports show symptoms below, you might suspect this problem:- "KQR xxx" occupies large portion in shared pool before problem happens- After problem resolved automatically, "KQR xxx" decreases much and free memory increase a lot If all of these are satisfied, you may be hitting this bug:- LCK0 sits inside 'row cache cleanup' for long time (for example,  more than 30 seconds)- LCK0 is mostly spending time on latch waits.  Specifically it is  alternating on waits 'latch: row cache objects' and   'latch: shared pool' - Other processes are frequently running kghfrunp to create free  memory.  They are also alternating on waits 'latch: row cache objects'  and 'latch: shared pool'  Workaround None.  

Please note: The above is a summary description only. Actual symptoms can vary. Matching to any symptoms here does not confirm that you are encountering this problem. For questions about this bug please consult Oracle Support.

References

Bug:18199537 (This link will only work for PUBLISHED bugs)
Note:245840.1 Information on the sections in this articl