Consider a single node HQ configuration. If you kill and restart the HQ server before the HQ client has exhausted all its reconnect attempts, the HQ client (publishers and receivers) reconnect just fine. However the HQ clients are seeing their exception listeners being invoked.
This invocation of the exception listeners seems to be incorrect. If the JMS client reconnected successfully, there should be no reason to invoke the exception listener.
The JMS spec, sec 4.3.8 on Exception Listener states:
A JMS provider should attempt to resolve connection problems itself prior to
notifying the client of them.
My interpretation is that when a JMS client reconnects, it has resolved the problem and should just continue on without notifying the client.
Tested with 2.1.2Final.
Thanks
Aspi Engineer
Putnam Investments
