DB2 FOR LUW / DB2FORLUW / 970 - DB2

Deactivate takes a long time to complete on HADR standby if one 
of the redo workers is working on a queue having the minLSN 
record, it never leaves the queue to check if an interrupt has 
arrived. 
 
When the redo master thread gets the interrupt, sets a flag for 
the redo workers to see, but the redo workers can see it only if 
they are trying to find a queue to work on. If the redo worker 
is already working on a queue and happens to have the minLSN 
record on the queue, it does not check for the flag hence does 
not know that the interrupt has arrived.

**************************************************************** 
* USERS AFFECTED:                                              * 
* ALL                                                          * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* Deactivate takes a long time to complete on HADR standby.The * 
* typical symptom is that one of the db2redow                  * 
* workerthread/process will keep on working and the deactivate * 
* willwaitfor that to complete.If a db2trc is taken of that    * 
* process/thread, it'll show thatitis working and processing   * 
* records.At this point, the only way to come out of this      * 
* situationquickly is to kill the instance using db2_kill or   * 
* wait tilldb2redow thread has completed its work.LOCAL        * 
* FIX:db2_kill to kill the instance. Please note, that restart * 
* oftheinstance will now take time as recovery will start      * 
* rightfromthe point where it got interrupted by the kill      * 
* signal.                                                      * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Upgrade to DB2 version 9.7 Fix Pak 3.                        * 
****************************************************************

db2_kill to kill the instance. Please note, that restart of the 
instance will now take time as recovery will start right from 
the point where it got interrupted by the kill signal.

DB2 Version 9.7 Fix Pack 3 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 3a for Linux, UNIX, and Windows
 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
 

Problem was first fixed in Version 9.7 Fix Pack 3.

See LOCAL FIX

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

11.05.2010
30.09.2010
30.09.2010

9.7.,
9.7.FP3