I have been doing more testing of different scenarios regarding transacted
sessions and different acknowledgement nodes and here is what I've found
regarding this topic/behavior:
My initial tests (and what I provided logs for) didn't use transactions and
used auto acknowledge:
Session session = connection.createSession(false, Session.AUTO_ACKNOWLEDGE);
I also tested with client acknowledgement:
Session session = connection.createSession(false,
Session.CLIENT_ACKNOWLEDGE);
and had the same behavior occur - message consumption stopped after the
second failover.
I then tested with using transactions (which is our normal use case but not
100% of the time):
Session session = connection.createSession(transacted,
Session.SESSION_TRANSACTED);
and this test was successful. I failed over multiple times (at least 3) and
each time message consumption continued. Messages that could not be
committed because a failover was in progress were redelivered correctly once
the client connected to the new master node.
--
Sent from: http://qpid.2158936.n2.nabble.com/Apache-Qpid-users-f2158936.html
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscr...@qpid.apache.org
For additional commands, e-mail: users-h...@qpid.apache.org
