3 Replies Latest reply on Nov 14, 2006 12:52 PM by marlig

    Help needed with InvalidationManager and InvalidationBridge

    marlig

      Hi everyone,

      I've been searching the forum for a solution to my problem but have not found any. So maybe someone here can help me.

      I'm having problems when using the invalidation service in a 2 - nodes cluster environment. It appears that the last started node does not get to know that there are already other nodes with the invalidation manager running. Btw, I tried this up to 4 nodes, it's always the last one to be unaware of the others.

      What I'm doing:

      1. default InvalidationManager (cache-invalidation-service.xml) - and InvalidationBridge (cluster-service.xml) - settings at every node.

      2. defining and registering of invalidation groups at the InvalidationManager (dynamic via MBean) during startup of every node.

      3. startup node 1

      4. startup node 2

      5. invalidation test of defined invalidation groups at node 1 via jmx-console --> invalidation is done at node1 and node2

      6. invalidation test of defined invalidation groups at node 2 via jmx-console --> invalidation is done at node2 only

      7. refreshing the InvalidationBridge - MBean at node1 (stop - start) via jmx-console

      8. invalidation test of defined invalidation groups at node 2 via jmx-console --> invalidation is now done at node1 and node2


      Am I missing something for the setup? Or is this the desired behaviour (which I can hardly imagine)? In the latter case: is it possible to avoid step 7 and automate this process?

      Any help is highly appreciated.

      Thanks
      Martin