MQSeries.net :: View topic

schroederms · Posted: Fri Aug 15, 2003 2:59 pm Post subject:

We had a production problem last night and I want to run a question by someone to see if this has happened to them.

We have three queue managers in our cluster, for simplicity I'll call them Qmgr0, Qmgr1, and Qmgr2. The clients connect to Qmgr0 and the WMQI brokers run on Qmgr1 and Qmgr2. The queue EAI.REQUEST resides on Qmgr1 and Qmgr2.

Okay, saying that. What happened last night was that messages were being round robin between Qmgr1 and Qmgr2 which is fine, but the flow running on Qmgr1 was not pulling the messages off EAI.REQUEST queue, as the exact flow was on Qmgr2. I stopped and started the brokers on Qmgr1 thinking it would correct the problem and start processing the messages, however it did not.

I tried taking EAI.REQUEST out of the cluster, thinking Qmgr0 would not see it, however there must be a refresh needed to make sure all Qmgrs in the cluster recognize a queue being pulled from the cluster. Anyway I ended up deleting EAI.REQUEST on Qmgr1, defining it again, and adding it back to the cluster and then messages in EAI.REQUEST on both Qmgrs were getting processed by each WMQI broker.

My question is.... Has anyone ever seen a broker just stop processing messages from a queue when visually all looks OK (No RED) ? If so what do you do to correct.

WMQI is farly new to me and I'm not a MQ Admin by any means.

Thanks,
Mike