I've configured a two-nodes cluster following chapter 16 of jboss4guide:
* Server "All"
* node1\deploy\jbossweb-tomcat55.sar\META-INF\jboss-service.xml: SnapshotMode, SnapshotInterval, true, false
* Web.xml, with
* Jboss-web.xml, with <replication-config>
* JBoss.xml, with true into each ejb definition.
Clustering services under HTTP (via Apache+modjk) works well, balancing the requests and replicating states, but in my app i've a RMI communication. For that, I followed the steps about "Dynamic Proxy".
First question: As I can understand, "Dynamic Proxy" with Stateless Session Beans doesn't require code change on the client side, am i wrong?
Second: Dynamic proxy is based by defining ip and HA-JNDI port? java.naming.provier.url=server1:1100,server2:1100,server3:1100,server4:1100, for example... I understood that defining that on the client, it will try with each node until it gets a response... and if not, it will try multicast communication.
My situation is that I can deploy the app in the farm correctly, and when i'm using it, if I decide to shutdown the node that's processing the client's requests, the client still trying to communicate with it...:
org.jboss.invocation.ServiceUnavailableException: Service unavailable.
So I think I'm missing something on the client side... but i really don't know what.
Could you help me?
Thanks a lot in advance!!