Backup server seems to be handling redelivered messages

[beltane]

It appears to me that the hornetq instance that I've configured to be the backup is handling messages that are scheduled for redelivery.

Hornetq 2.2.14 running in Tomcat 7.0.27, set for live/backup.

When I connect to the two instances via jmx, I only see the queues in the live server, the backup only shows MBeans related to configuration, like the BroadcastGroup, ClusterConnection, and Server.

The backup server has this in the log:

INFO: HornetQ Backup Server version 2.2.14.Final (HQ_2_2_14_FINAL, 122) [2ed31396-6b10-11e2-86fb-fd5350827b40] started, waiting live to fail before it gets active

The backup server works as expected when I shut down the live server; it takes over.

What concers me is I see in the log that the backup server is handling old messages scheduled for redelivery.  This isn't about duplicate messages, because if the message appears in the backup servers log file, it doesn't show up in the live servers log for a particular redelivery attempt.  Though, I have seen the same message being redelivered on the live server and the backup server, though only in separate redelivery attempts.

Why would this be?  Why wouldn't these messages be redistributed to the live server and if I believe the jconsole view of the MBeans, and there are no queues on the backup server, what consumer on the backup server would be handling these messages?

Thank you for the help.

[ataylor]

What concers me is I see in the log that the backup server is handling old messages scheduled for redelivery.  This isn't about duplicate messages, because if the message appears in the backup servers log file, it doesn't show up in the live servers log for a particular redelivery attempt.  Though, I have seen the same message being redelivered on the live server and the backup server, though only in separate redelivery attempts.

Why does it concern you, the backup server will take over responsibilty of all messages when it comes up.  maybe i am misunderstanding what you are saying, could you try to explain.

why would this be?  Why wouldn't these messages be redistributed to the live server and if I believe the jconsole view of the MBeans, and there are no queues on the backup server, what consumer on the backup server would be handling these messages?

messages arent redistributed between live and backup, the backup simply takes over responsibilty when the live server crashes, and the same happens when the live takes over from the backup stopping