Sometimes happens that jms client is not able to failover to backup - attached thread dump from consumer + logs from servers.

Test scenario:

1. Start server B and server A in collocated topology - servers contain backup for each other
2. Start producer and consumer connected to server A - clients are active for the whole duration of the test
3. Kill server A -> clients should failover to backup server A on server B but sometimes it won't happen

Reproducer attached - how to use it:
1. Download and unzip reproducer.zip
2. Prepare EAP server with profiles for server A and B in collocated topology by using - "sh prepare-servers.sh"
3. Open a new console and start server B - "sh start-server2.sh" (bind to 127.0.0.2)
4. Open a new console and start server A - "sh start-server1.sh" (bind to 127.0.0.1)
5. Open a new console and start producer connected to server A - "sh start_producer.sh" (connected 127.0.0.1 - queue/Inqueue)
6. Open a new console and start consumer connected to server A - "sh start_consumer.sh" (connected 127.0.0.1 - queue/Inqueue)
7. Kill server A using - "sh killA.sh" or use "kill -9 ..." (note "jps -m" prints info about processes)
8. Now clients should hang but if not stop all servers and clients and continue by step 3. (I hit it at first try in messaging lab)

Scripts can be easily modified. In the end i decided to set highest priority. Ping me if there are any problems.