If a broker goes down and comes back up, either due to node failure or loss of connection, we need to resume the connection and restore any management objects necessary.
Resolved in revision 2046.
cumin-0.1.2058-1.el5, qpidc-0.2.656926-1.el5, qpidd-0.2.656926-1.el5, and python-qpid-0.2.656025-12.el5 have been pushed to the staging repo for testing