PQ62564: WEBSPHERE TRANSACTION COORDINATOR CAN'T RECOVER THE IN-DOUBT TRANSACTION WHEN THE RESOURCE MANAGER IS NOT AVAILABLE

APAR status
Closed as program error.

Error description
Application uses MQSeries 5.2 Queue and Oracle 8.1.7.4 JDBC
Driver.
  The application takes out messages from MQSeries Queue and
inserts records into Oracle database tables.
  WAS and MQSeries are on the same machine and Oracle is on
another machine.
  After all the components (i.e. MQSeries, WebSphere
Application Sever & Oracle) of the system are started up and
running, abruptly crash the machine on which the WAS
application server is running by powering it off. The Oracle
database shows pending transactions in table DBA_2PC_PENDING
and DBA_2PC_COMMIT. Then bring the Application Server back
up. MQSeries is still down when Application Server is brought
back up.
  Once the Application Server is back up all the pending
transactions are cleared automatically by WebSphere.
  MQSeries is brought back up.
  Problem: The sum of messages in the MQSeries Queue and the
count of number of records in the Oracle database tables does
not match with the number of messages that they started.
.
WebSphere Transaction Coordinator has problem with recovery
when the Resource Manager (MQ Series in this case) is not
available.
Local fix Problem summary
****************************************************************
* USERS AFFECTED: Recovering WebSphere Application Servers     *
*                 where the Resource Managers are not yet      *
*                 running.                                     *
*                 This applies to 3.5 and 4.0 systems.         *
****************************************************************
* PROBLEM DESCRIPTION: Possible data inconsistencies when      *
*                      recovery is performed when a resource   *
*                      manager is not yet started.             *
****************************************************************
* RECOMMENDATION: It is reccomended that all WebSphere         *
*                 installations apply this e-fix.              *
****************************************************************
When a WebSphere Application Server is attempting to recover
Resource Managers such as DB2, Oracle, MQSeries, etc.  If a
Resource Manager has not yet been started when WebSphere does
its recovery it will issue an event message, but will
continue with recovering other Resource Managers.  This is
incorrect operation and can lead to data inconsistencies.
Problem conclusion
WebSphere will try to contact the Resource Managers during
recovery. If any are unavailable then WebSphere will issue
event messages stating that is is waiting for a particular
Resource Manager. If all the Resource Managers are available
within 4 minutes then recovery will proceed. If a Resource
Manager is not available within this time then the Application
Server will terminate.  Please note that the event messages
do not appear immediatley in the WebSphere Console window
when it is trying to start a server, but do get written to
the log.
Temporary fix Comments
APAR information
APAR number PQ62564
Reported component name WEBSPHERE AE NT
Reported component ID 5630A2201
Reported release 400
Status CLOSED PER
PE NoPE
HIPER NoHIPER
Submitted date 2002-06-24
Closed date 2002-07-17
Last modified date 2002-07-17

APAR is sysrouted FROM one or more of the following:

APAR is sysrouted TO one or more of the following:
PQ40320

Modules/Macros
JTSXA XARM        

Fix information
Fixed component name WEBSPHERE AE NT
Fixed component ID 5630A2201

Applicable component levels
R400 PSY    UP


Document Information


Product categories: Software > Application Servers > Distributed Application & Web Servers > WebSphere Application Server > General
Operating system(s):
Software version: 400
Software edition:
Reference #: PQ62564
IBM Group: Software Group
Modified date: Jul 17, 2002