Announcement Announcement Module
Collapse
No announcement yet.
Spring Integration, WebSphere MQ, Tomcat, & Availability Page Title Module
Move Remove Collapse
X
Conversation Detail Module
Collapse
  • Filter
  • Time
  • Show
Clear All
new posts

  • Spring Integration, WebSphere MQ, Tomcat, & Availability

    Hello,

    How can I configure Spring Integration, Tomcat (JNDI), and WebSphere MQ technologies to dynamically handle Availability?

    Availability = when the primary MQ server is down, configure the application to dynamically use the secondary MQ server. The primary and secondary MQ servers are exactly identical apart from their host name.

    JNDI Definition:
    <Resource CHAN="SOME.CHANNEL" HOST="SOME.HOST" PORT="1414" QMGR="THEQMGR" auth="Container" factory="com.ibm.mq.jms.MQQueueConnectionFactoryFa ctory" name="jms/someQCF" type="com.ibm.mq.jms.MQQueueConnectionFactory"/>

  • #2
    Websphere MQ has provision for clustering and high availibility. So you dont have to do anything specific in you application to ensure high availability.

    Gary has got one project up here which demonstrates high availability in spring integration. Have a look at it too.

    Comment


    • #3
      If you are running active-active MQ queue managers, this is not a trivial problem. MQ clustering is the area you should be looking at.

      We run active-passive. If MQ fails on one server, we fail over to another server, but also take a floating IP across as well. So an IP address is always assigned to a server that is running the active MQ node. Then we create a DNS entry that points at that floating IP and thus we always connect to the correct node that is running the queue manager.

      Hope that makes sense.

      Comment


      • #4
        The link @amol provided will not help for this situation.

        It's been a while since I worked with WebsphereMQ, but IIRC, you configure a local connection definition table (or something named similarly) that knows about the HA broker locations and the URL you provide to the connection factory points to the local client and it deals with failover transparently.

        I suggest you dig around in the IBM documentation. I'll see if I can find any more info in my email archive.

        Comment


        • #5
          Thanks for the responses.

          @GaryRussell - yes, I will explore CDDTs (client channel definition table). Hopefully, implementation's straightforward and that this solves the issue. If you find more info, please send it over.

          @DPR - Floating IP is unfortunately not part of our infrastructure.

          Comment

          Working...
          X