Simple question - will HornetQ clustering still work with the persistenceEnabled option set to false?
Under JBM1, a shared database was necessary for clustering, but since v2 doesn't use a shared database, I'm guessing clustering is now independent of this (i.e. "shared nothing").
We won't require a shared database for clustering.. no.
For failover, we will have the concept of backup servers, the backup will use either a shared disk or a replicated storage.
The concept of fail over and clustering is separated now on HornetQ.