Is it possible to implement failover without a shared file system, if I don't need persistence? I see that there is a shared-store element that can be set to "false." I don't have access to a shared file system, and I plan to use only non-durable queues. I don't see the use of a shared file system if I'm using non-durable queues. The only thing I would need is a queue on each node in the cluster. Any way to configure this set up?
The shared store is used to configure live/backup groups. If you set to false, it will use replication mode (supported in 2.3.0.Alpha). If you don't use backups, you can ignore it.
Howard