The following two examples represent a pair of symmetrical stacks, used to configure a cluster of two TCP members.

*Note that if you enable Windows XP's built-in firewall, even when you use exceptions for the

program you start, there will be problems with members discovering each other with TCPPING and with

merging (MERGE2)*

polaris stack

  <config>
    <TCP bind_addr="polaris" start_port="7800" loopback="true"></TCP>
    <TCPPING initial_hosts="gatecrasher[7800]"
             port_range="3"
             timeout="3500"
             num_initial_members="3"
             up_thread="true"
             down_thread="true"></TCPPING>
    <MERGE2 min_interval="5000" max_interval="10000"></MERGE2>
    <FD shun="true" timeout="2500" max_tries="5" up_thread="true" down_thread="true" ></FD>
    <VERIFY_SUSPECT timeout="1500" down_thread="false" up_thread="false" ></VERIFY_SUSPECT>
    <pbcast.NAKACK down_thread="true" up_thread="true" gc_lag="100" retransmit_timeout="3000" ></pbcast>
    <pbcast.STABLE desired_avg_gossip="20000" down_thread="false" up_thread="false" ></pbcast>
    <pbcast.GMS join_timeout="5000"
                join_retry_timeout="2000"
                shun="false"
                print_local_addr="false"
                down_thread="true"
                up_thread="true"></pbcast>
  </config>

gatecrasher stack

  <config>
    <TCP bind_addr="gatecrasher " start_port="7800" loopback="true"></TCP>
    <TCPPING initial_hosts="polaris[7800]"
             port_range="3"
             timeout="3500"
             num_initial_members="3"
             up_thread="true"
             down_thread="true"></TCPPING>
    <MERGE2 min_interval="5000" max_interval="10000"></MERGE2>
    <FD shun="true" timeout="2500" max_tries="5" up_thread="true" down_thread="true" ></FD>
    <VERIFY_SUSPECT timeout="1500" down_thread="false" up_thread="false" ></VERIFY_SUSPECT>
    <pbcast.NAKACK down_thread="true" up_thread="true" gc_lag="100" retransmit_timeout="3000" ></pbcast>
    <pbcast.STABLE desired_avg_gossip="20000" down_thread="false" up_thread="false" ></pbcast>
    <pbcast.GMS join_timeout="5000"
                join_retry_timeout="2000"
                shun="false"
                print_local_addr="false"
                down_thread="true"
                up_thread="true"></pbcast>
  </config>

Note

If you want to use the above stack to configure a JBoss partition, add

      <pbcast.STATE_TRANSFER up_thread="true" down_thread="true" ></pbcast>

at the top of the stack.

Back to JGroups