2 Replies Latest reply on Dec 7, 2016 3:12 AM by cennore

    Cluster + Infinispan Issues

    cennore
    cennore Feb 10, 2016 5:29 AM

      Hello,

       

      I have Wildfly 8.2.0 running in domain mode and I've managed to set up a cluster using mod_cluster and httpd successfully.

       

      I currently have 5 cluster groups with 2 nodes each running without any challenge. WIldfly domain management is sweet I must confess!

       

      I however have a problem with Infinispan. I noticed once I add more than 2 nodes in any given cluster group I get the following execptions:

      1. "Failed to start rebalance for cache dist TimeoutException One of the nodes null timed out". This is thrown on one of any of the existing 2 nodes which are working fine

      2. "Initial state transfer timed out for cache dist". This is thrown on the new 3rd node I'm trying to add to the cluster after the state transfer timeout has elapsed.

       

      Are Wildfly clusters limited to just 2 nodes per cluster? Please advise.

       

      Thanks

        • 1. Re: Cluster + Infinispan Issues
          wdfink
          wdfink Feb 12, 2016 5:02 PM (in response to cennore)

          No WildFly clusters are not limited to 2 nodes

          But there are some pitfalls.

           

          Did you use infinispan for your application as a cache? I don't mean session replication.

          If not did you changed the infinispan config?

          What JGroups configutation do you use?

            • Actions
          • 2. Re: Cluster + Infinispan Issues
            cennore
            cennore Dec 7, 2016 3:12 AM (in response to wdfink)

            Hi Fink,

             

            First of all thanks for taking time out to reply my question.

             

            I noticed the following however, not exactly sure if or how they are related to the exceptions observed.

             

            1. I shifted from using Wildfly inbuilt hornetq for messaging to an external ActiveMQ broker using the Wildfly RA. This drastically reduced any issue I had with Jgroups immediately.

            2. I also noticed I had to start up my hosts in sequence i.e I start up all the hosts on a particular node before starting up other hosts within a different node for the same server group. It also seems more stable when I start up with the node which was first registered on the Domain.

             

            What do you think?

             

            Thanks a lot!

              • Actions