DB2 FOR LUW / DB2FORLUW / A10 - DB2

When an alter bufferpool command fails trying to increase the 
size of the bufferpool, not all memory allocated to the point of 
failure is freed back to the database memory set.  The 
bufferpool heap (specific to the bufferpool being altered) 
caches the memory that was used for the additional bufferpool 
descriptors (metadata).  In the case of constrained memory (eg. 
fixed database memory, fixed instance memory, or operating 
system limits), it may be that database memory overflow has been 
exhausted and database memory can't grow any further. Areas that 
rely on using additional database memory overflow (such as 
package cache, database heap growing beyond its initial size to 
load additional table information) may no longer be able to 
allocate memory.  This may result in SQL0973 and other failures 
related to out-of-memory conditions. 
 
After the bufferpool increase failure, db2pd -db <database> 
-mempools may show that overflow is exhausted: 
 
Database Partition 0 -- Database SAMPLE -- Active 
 
Memory Sets: 
Name     Size(Kb)  DBP  Type Unrsv(Kb)  Used(Kb) HWM(Kb) 
Cmt(Kb) 
SAMPLE  126848    0      1       0              70784      71808 
71808 
 
note the largest portion of the memory required for bufferpools, 
that of the bufferpool memory itself, is released, and it is 
only rare cases that the smaller unreleased cached memory for 
bufferpool metadata causes a problem. (typically it requires a 
fixed database_memory setting resulting in a relatively 
constrained configuration, and a large error in the size 
specified in the alter bufferpool command). 
 
A workaround is to issue a second "alter bufferpool" which will 
release any leftover cached memory upon success.  It is enough 
just to raise or lower the number of pages by 1 from the current 
(pre-failure) size.

**************************************************************** 
* USERS AFFECTED:                                              * 
* All systems                                                  * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* See Error Description                                        * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Use workaround (Local Fix) as the APAR is a secondary        * 
* problem resulting from an expected failure                   * 
****************************************************************

Issue an additional alter bufferpool command, increasing the 
number of pages by one from the current size.  This will release 
any leftover retained memory from the failing bufferpool 
increase operation.  The bufferpool can then be altered back to 
the original size.

DB2 Version 10.1 Fix Pack 1 for Linux, UNIX, and Windows
 DB2 Version 10.1 Fix Pack 2 for Linux, UNIX, and Windows
 DB2 Version 10.1 Fix Pack 3 for Linux, UNIX, and Windows
 DB2 Version 10.1 Fix Pack 4 for Linux, UNIX, and Windows
 DB2 Version 10.1 Fix Pack 3a for Linux, UNIX, and Windows
 DB2 Version 10.1 Fix Pack 6 for Linux, UNIX, and Windows
 

First fixed in DB2 Version 10.1 Fix Pack 1

See Local Fix

Date  - problem reported    :
Date  - problem closed      :
Date  - last modified       :

13.06.2012
02.11.2012
02.11.2012