DB2 FOR LUW / DB2FORLUW / 970 - DB2

A memory leak occurs in Database Heap on HADR standby database. 
This memory leak happens at a data structure used by log file 
management on standby database. This data structure is allocated 
when a new log file is created. But, it is not freed after 
using. Fortunately, this data structure will be reused if its 
corresponding log file is reused. 
 
So, this memory leak can be triggered by LOGARCHMETH1 = 
LOGRETAIN, since log files will not be reused. 
 
Database Heap memory usage can be monitored by one of the 
following methods: 
- db2pd -db <database> -mempools   (monitor the "dbh" pool) 
- database snapshot 
- select POOL_ID, POOL_CUR_SIZE, POOL_WATERMARK from 
sysibmadm.snapdb_memory_pool where POOL_ID = 'DATABASE' 
 
If this memory leak exists, you can observe a memory usage 
increase by 2096 bytes for every new log file. You can use db2pd 
to monitor memory usage of this data structure. 
 
  db2pd -d <standby db name> -memb sort 2 
 
Here is a sample output: 
======================== 
Memory blocks sorted by size for dbh pool: 
PoolID   PoolName TotalSize(Bytes) TotalCount LOC File 
...... 
2        dbh      16768            8          74  418710486 
 
Please note that the file 418710486 is your target to monitor. 
"TotalCount" means how many data structures allocated, and still 
not released. 
"TotalSize(Bytes)" means how many bytes occupied by this data 
structure. 
You can see that "TotalSize(Bytes) / TotalCount = 2096" and both 
of them keep rising along with new log files being created.

**************************************************************** 
* USERS AFFECTED:                                              * 
* ALL                                                          * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* A memory leak occurs in Database Heap on HADR standby        * 
* database.                                                    * 
* This memory leak happens at a data structure used by log     * 
* file management on standby database. This data structure is  * 
* allocated when a new log file is created. But, it is not     * 
* freed after using. Fortunately, this data structure will be  * 
* reused if its corresponding log file is reused.              * 
*                                                              * 
* So, this memory leak can be triggered by                     * 
* LOGARCHMETH1=LOGRETAIN, since log files will not be reused.  * 
*                                                              * 
* Database Heap memory usage can be monitored by one of the    * 
* following methods:                                           * 
* - db2pd -db <database> -mempools  (monitor the "dbh" pool)   * 
* - database snapshot                                          * 
* - select POOL_ID, POOL_CUR_SIZE, POOL_WATERMARK from         * 
* sysibmadm.snapdb_memory_pool where POOL_ID = 'DATABASE'      * 
*                                                              * 
* If this memory leak exists, you can observe a memory usage   * 
* increase by 2096 bytes for every new log file. You can use   * 
* db2pd to monitor memory usage of this data structure.        * 
*                                                              * 
* db2pd -d <standby db name> -memb sort 2                      * 
*                                                              * 
* Here is a sample output:                                     * 
* ========================                                     * 
* Memory blocks sorted by size for dbh pool:                   * 
* PoolID  PoolName TotalSize(Bytes) TotalCount LOC File        * 
* ......                                                       * 
* 2        dbh      16768            8          74  418710486  * 
*                                                              * 
* Please note that the file 418710486 is your target to        * 
* monitor.                                                     * 
* "TotalCount" means how many data structures allocated, and   * 
* still not released.                                          * 
* "TotalSize(Bytes)" means how many bytes occupied by this     * 
* data structure.                                              * 
* You can see that "TotalSize(Bytes) / TotalCount = 2096" and  * 
* both of them keep rising along with new log files being      * 
* created.                                                     * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Update to version 9.7 fixpack 4 or later.                    * 
****************************************************************

Set log archive method to other methods, instead of LOGRETAIN. 
Alternatively, if you have chance to restart your standby 
database, the memory will be freed.

DB2 Version 9.7 Fix Pack 4 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 5 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 6 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 7 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 8 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 9 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 9a for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 10 for Linux, UNIX, and Windows
 

This problem is first fixed in version 9.7 fixpack 4.

not known / see Local fix

forerunner  : APAR is sysrouted TO one or more of the following: IC70802 IC70803 
follow-up : 

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

18.08.2010
28.04.2011
28.04.2011

9.7.FP4