1. cluster of 2 nodes - A - B
2. send 500 to each (db status - A:500, B:500)
3. close connection between A and B (db unchanged)
4. wait until nodes realize that they're alone (db unchanged)
5. disconnect A from DB and wait for failover to B (db status - A:0, B:1000, only B is in jbm_postoffice)
6. disconnect B from DB and wait for B to shutdown (db status - A:0, B:1000, only B is in jbm_postoffice)
7. reconnect A to DB and wait for node A to start again -> failover won't occur, i've waited more than 10 minutes (db status - A:0, B:1000, A,B is in jbm_postoffice)
8. read from A - 0 messages read - null is returned from consumer.receive(1000) method
9. read from B - exception
After step 7. all messages should be available from A. Can you please look at it and confirm the issue?
Edit: Clustered queue deployed on both nodes were used for storing messages. Clients used "ClusteredConnectionFactory".