I'm trying to migrate/reproduce our actual JBoss infrastructure, based on EAP 4.3, with the new version. In that infrastructure we have an F5 BigIP that load balances several JBoss 4.3 clustered instances which deploy our software. The cluster also exposes JMS in and out queues reachable by external clients with HA-JNDI.
Our software doesn't need of session replication but relies on session stickyness and consists in a couple of EARs, one that contains two WARs and a lot of JARs some of what contain Message Driven EJBs (MDBs), the second a lot of WARs and a persistence based on JPA/Hibernate.
What we are facing, configuring two EAP 6.1 instances in standalone-full-ha mode, is that: the main jboss instance successfully starts and deploys all the resources. The second jboss instance starts to deploy all the servcices but it stops waiting for the "live lock" from the other server, due I think to the JMS/HornetQ configuration, but doesn't make available the webapps in the WARs, also the administration console is not available.
If I tear down the first instance (by kill or CTRL-C) the second one receives the live token, and finishes all the deploys, making available all the web apps.
What I'm doing wrong? The standalone configuration is ok for our purposes or should l use the domain configuration in order to reproduce our actual cluster?