When I set up HA Singleton for the whole application (two instances), I was able to verify that the other instance will be selected as new HA singleton provider once the current HA singleton provide is down.
However, if the current HA Singleton provider (instance #1) is processing messages in a cluster queue, and I shut it down. When the other instance (instance #2) becomes the new HA singleton provider, the remaining messages produced by the previous HA singleton (instance #1) becomes unavailable to the new HA singleton provider (instance #2).
However, when I restart instance #1, I can see message processing in instance #2 until all remaining messages are completely processed.
It looks like the messages produced by instance #1 becomes available to instance #2 only when instance #1 is not shut down. So the messages are not replicated to the cluster, but distributed in each instance. Is there anyway that I can make the messages available to instance #2 even after instance#1 was completely shut down?
