Yes I understand this. How could there possibly be data to replicate if the entire cluster is shutdown.
I specified the scenario in which a single instance failure occurs, the failover takes place but yet I'm required to login again. I have also been posting this on the JBoss Seam forums hoping to get some insight. Is there a way to examine/monitor the session data replication? As I said the logs appeared to be clean in terms of no exceptions.
> If you are using memory-replication and you stop the
> entire cluster, then all your data is lost. Were you
> expecting a different behavior? Memory replication
> relies on the heap memory of running replica partners
> in the cluster. Again, if all the instances of a
> cluster
> are stopped, then all your conversational data is
> gone.
>
> For help with documentation please try this:
> https://glassfish.dev.java.net/nonav/javaee5/docs/SJSA
> SEEHAAG.pdf
[Message sent by forum member 'samdoyle' (samdoyle)]
http://forums.java.net/jive/thread.jspa?messageID=248129