0 Replies Latest reply on Dec 11, 2017 2:47 AM by Benjamin Buehlmann

    Some questions about the <cluster-connection> for a simple static Artemis cluster in Wildfly 11

    Benjamin Buehlmann Newbie

      Hello everybody,


      What I would like to setup with two Wildfly 11 instances:

      • Artemis broker pair with one each live/backup node
      • HA strategy should be replication
      • Not using discovery of other brokers over Multicast, just wire the two brokers statically over cluster-connection
      • No load balancing between the two brokers
      • SSL encrypted traffic between brokers and between clients and brokers
      • Support for "old" clients using HornetQ Core protocol and for new Clients using Artemis Core multiplexed over http


      Configuration of the cluster-connection looks somehow like this:


      <replication-master cluster-name="artemis-cluster" group-name="artemis-group-1" check-for-live-server="true"/>  
      <cluster-connection name="artemis-cluster" address="jms" connector-name="http-connector" static-connectors="adjacent-broker-connector"/>  
      <http-connector name="adjacent-broker-connector" socket-binding="adjacent-broker" endpoint="http-acceptor"/>  
      <http-connector name="http-connector" socket-binding="http" endpoint="http-acceptor"/>  



      The following questions arised about cluster connections:

      • What exactly is the attribute "connector-name" on the cluster-connection for? In the documentation I find "The name of connector to use for live connection" which is not absolutely clear to me.
      • Based on which configuration attribute(s) will the cluster topology be created during runtime?
      • How does the propagation of the cluster topology from the broker to the clients works exactly (timing, pull or push)?
      • Is it possible to disable the propagation of the cluster topology to the clients or to override the received topology on the client side?


      Thanks and best regards,