MQSeries.net :: View topic

chintu · Posted: Thu Jun 21, 2007 4:02 am Post subject:

We were on
AIX 5.2
MQ 5.3.CSD12
WAS 5.0
WF 3.6( but using 3.4 classic jar files)

Last friday, we migrated our Production environment to WAS 6.0.2.5 and updated two J2EE applications that were using the 3.4 jars to use the WF 3.6 jar files. Since then we are seeing an increase in the number of bindings to some of the queues for the WF QMGR.

Here is a sample of what we are seeing:
echo 'display qstatus(EXCINPUTQ) type (handle) APPLTAG CONNAME'|runmqsc|grep -c EXCINPUTQ
10875
echo 'display qstatus(ADMINPUTQ) type (handle) APPLTAG CONNAME'|runmqsc|grep -c ADMINPUTQ
5437
echo 'display qstatus(RTSINPUTQ) type (handle) APPLTAG CONNAME'|runmqsc|grep -c RTSINPUTQ
5440

a)Is it ok to have such a hig OPPROCS count?

All of them point to "java". The processIds point to either of the 2 J2EE applications that are now using the 3.6 native jar files.
Both J2EE applications are custom webclients( wirtten using the WF API). There are no interactions to any queues or databases. Hence there is no chance that the code is not closing connections...etc.

We have not updated the MQ_INSTALL_ROOT websphere environment variable in WAS6 to point to the /usr/mqm. It is currently pointing to the MQ libraries that were installed with WAS.

b)Could this be an mq jar mismatch issue?

Thanks.