EJB failover works because when you lookup your EJB in JNDI, you are actually downloading a proxy that contains within it the current cluster topology for the EJB (i.e. what nodes the EJB is deployed on) along with load balancing and failover logic. If you try to use that proxy to invoke on a crashed node, the proxy detects the failure and retries the call on one of the other nodes where the bean is deployed. When the call reaches that other node, the EJB container retrieves the bean' state from the distributed cache (JBoss Cache). View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4190240#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...