1 Reply Latest reply on Aug 22, 2013 10:52 AM by Stanley Tso

    Can I configure JMS XA Recovery for use with Amazon EC2 auto-scaling in EAP 5.1.2?

    Stanley Tso Newbie

      I am attempting to configure a cluster of JBoss EAP 5.1.2 nodes at Amazon EC2.  My goal is to set up a template image so that I can use the nearly the same configuration when I add new nodes with Amazon auto-scaling.  I am following the documentation at https://access.redhat.com/site/documentation/en-US/JBoss_Enterprise_Application_Platform/5/html-single/Messaging_User_Guide/index.html#recovery and I need clarification on what it is saying.


      Let say that I add a new JBoss node using Amazon auto-scaling and I wish for this node to join the existing JBoss cluster, in particular JMS.  Is it even possible for a new JMS node to dynamically appear in the cluster and participate in the JMS XA recovery process?  My understanding of what the documentation says is that it is possible for a new JMS provider to be defined and added dynamically in "server/all/deploy/messaging/hajndi-jms-ds.xml" because this file is located in a directory watched by the hot deployment scanner.  I would have to add the new provider (JMS node) to all of the existing JMS nodes in this particular configuration file.  However, for new JMS nodes to participate in JMS XA recovery, would I have to also modify "server/all/conf/jbossts-properties.xml" to tell the recovery manager that there are new JMS nodes from which we must recover?  Isn't this file statically loaded by each JBoss node and if so, does that mean I have to restart the existing JBoss JMS nodes to pick up this new configuration?  That would defeat the purpose of auto-scaling in the first place, wouldn't it?


      Am I on the right track here?  Thank you in advance for your help!

      Stan Tso

        • 1. Re: Can I configure JMS XA Recovery for use with Amazon EC2 auto-scaling in EAP 5.1.2?
          Stanley Tso Newbie

          After playing with the configuration files and testing, I have decided not to use EC2 auto-scaling with my JMS tier under JBoss EAP 5.1.2.


          Let's say that a new JBoss JMS node comes online using auto-scaling.  What does it take to get this new node visible to the rest of the JMS cluster?  Based on the documentation for JMS, it appears that I would have to modify several configuration files across all nodes to make this happen.  I would then have to restart those nodes because the file 'jbossts-properties.xml' is statically loaded at startup and is not hot-deployable.  That would defeat the purpose of bringing a new node online which is to increase capacity.  It looks like my team is going to establish a separately managed JMS tier running only queues, topics, connection factories, etc., and put any application logic on auto-scaled JBoss app server nodes hosted outside of the queues, topics, etc.

          1 of 1 people found this helpful