3 Replies Latest reply on Apr 15, 2011 11:24 AM by Nigel Deakin-Oracle

    Broker cluster

    hipjim
      Hi,

      I have a web service running on a glassfish cluster with two nodes. The service is looking up a jms queue (jndi lookup) that is deployed with target cluster. Depending on which node the lookup is done, queue state is different.
      Is this a normal behaviour? What can i do to have consistent state for all brokers in the cluster?

      Thanks.
        • 1. Re: Broker cluster
          Nigel Deakin-Oracle
          The two Queue objects should be the same. In any case, a Queue object in JNDI doesn't contain very much information apart from the name of the queue.

          Can you be more specific about what you mean by "queue state is different"?

          Nigel
          • 2. Re: Broker cluster
            hipjim
            I am doing a jndi look-up for the connection factory and the queue and using a queue browser i'm iterating over the messages. Dependent on which node i am, the messages in the queue are not the same.

            The broker cluster looks like this:

            ----------------------------
            Host Primary Port
            ----------------------------
            xxx.xxx.xxx.xxx xxxx

            Cluster ID
            Cluster is Highly Available false

            ----------------------------------------------
            Address State
            ----------------------------------------------
            xxx.xxx.xxx.xxx:xxxx OPERATING
            xxx.xxx.xxx.xxx:xxxx OPERATING
            • 3. Re: Broker cluster
              Nigel Deakin-Oracle
              The Message Queue 4.5 release note states:

              "A client can only browse the contents of queues that are located on its home broker. The client can still send messages to any queue or consume messages from any queue in the cluster; the limitation only affects queue browsing."

              Nigel