We are migrating from Weblogic to JBoss 4.0.1SP1 but we get into a fail-over issue.
We have two Apache servers and two JBoss servers. To setup fail-over between Apache and JBoss, we read
All we need is fail-over, NOT HTTP Session Replication, and NOT HAJMS. In short, just fail-over, not clustering.
But we read the article
"http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3872837#3872837" which says we need the "all" config.
Our concern in using the "all" config is that it puts JMS in the deploy-hasingleton directory which requires all the complicated steps to set up JMS to be shared from one JBoss server. We don't need this.
In Weblogic, we have both servers running their own JMS and the MDB's running on each server consumes messages from their respective JMS queues.
So my question is: Is it possible to use the "default" config to setup mod_jk_1.2 between Apache and JBoss?
If the answer is no and we have to use "all" config, then can we have each JBoss server running their own JMS, by copying C:\jboss-4.0.1SP1\server\default\deploy\jms\*.* into C:\jboss-4.0.1\server\all\deploy\jms?
Please advise as we can getting ready to production soon and this is becoming a major roadblock.
You still need to run "all" and weed out the unnecessary services.