Cluster Bridge is reconnecting to its own node, what will cause eventual message loss on certain races

We have hit regression in LODH #2 test case (see HornetQ test plan). We have tried to execute same test with EAP 5.1.1 and technical preview included in this build and test is ok.

HornetQ version included in EAP 5.1.2 ER1 is broken and cannot be used for testing. We have prepared own distribution from branch for EAP and we are using this version for testing. Same code base will be used for EAP 5.1.2 ER2.

We are working on the investigation.