MQSeries.net :: View topic

Getson · Posted: Mon Dec 09, 2002 1:07 pm Post subject:

I am building specifications for a new production installation of MQSeries and I am trying to determine an effective method for monitoring the health of the queue manager. This will likely be running within an active HACMP environment and so the results of this status report will be used to trigger HACMP roll-over. Do you have any thoughts as too what the "best" method would be to checking the status of a Queue Manager? I can think of three ways to do this but I am not sure of the relative merits of the three. All options are run from an AIX shell:

1)
echo "ping qmgr" | runmqsc myqmgr
ping succeeds or fails. However it seems to fail if QM is starting, and will hang for about 20 seconds if awaiting a response when a QM ends (endmqm -i)

2)
ps -fu mqm|grep "amqzxma0 -m myqmgr"|grep -v grep|awk '{print $2}' -`
returns a PID if queue manager is running, nothing if not. however this doesn't impart any knowledge of the health of the particular process...(i.e. perhaps it is hung and still running, starting/stopping and thus not processing messages)

3)
dspmq myqmgr
displays status of the qmgr. Reports starting, running, ending, ended (the later two including the type of stop). This is my preferred method as the command seems to be intended for just this purpose. Any gotcha's on this that I am missing?

I am avoiding an obvious fourth option (MQ event messages) as receiving those depends on the queue manager being available, and involves more complicated programming then I am willing to get into at this time.