0 Replies Latest reply on Aug 12, 2004 5:46 AM by arisa

    Cache Invalidation (over TCP) intermittent failure

    arisa Newbie

      Hi everyone,

      I have JBoss 3.2.3 in production over two servers. I've been having intermittent cache invalidation problems ever since we deployed. An example is an update to an entity bean in Server A is occasionally not reflected in Server B.

      The situation improved when we optimized the application so that it is lightly loaded (less than 50% utilization). However, it still occurred a couple of times a week.

      Thinking it might be the network switch swallowing multicast packets, I tried to solve the problem by switching to TCP for the transport layer of JGroups but it did not work.

      Two days ago, I added Apache to Server A to do load balancing to Server A and B using mod_jk. The cache invalidation problem has suddenly ballooned and I get 10 or more problem instances per day. (The workaround I use to survive temporarily is to flush the specific entity bean cache when it occurs).

      It would be a tremendous help if someone could help me identify some steps I could take to either diagnose the problem or solve it? Many thanks in advance!

      My JGroups configuration is below.


      <TCP start_port="7800" bind_addr=""/>
       <TCPPING initial_hosts="[7800],[7800]"
       port_range="5" up_thread="true" down_thread="true"
       num_initial_members="2" timeout="5000"/>
       <MERGE2 min_interval="5000" max_interval="10000" />
       <FD shun="true" up_thread="true" down_thread="true"
       timeout="2500" max_tries="5" />
       <VERIFY_SUSPECT timeout="3000" num_msgs="3"
       up_thread="true" down_thread="true" />
       <pbcast.NAKACK gc_lag="50" retransmit_timeout="300,600,1200,2400,4800"
       up_thread="true" down_thread="true" />
       <pbcast.STABLE desired_avg_gossip="20000"
       up_thread="true" down_thread="true" />
       <pbcast.GMS join_timeout="5000" join_retry_timeout="2000"
       shun="true" print_local_addr="true" />
       <pbcast.STATE_TRANSFER up_thread="true" down_thread="true" />