[ http://jira.jboss.com/jira/browse/JBAS-35?page=comments#action_12310827 ] Ovidiu Feodorov commented on JBAS-35: -------------------------------------
Correction to the previous comment: the problem only shows in JBoss versions older than 3.2.6. Since 3.2.6, JBoss uses JGroups 2.2.7 which fixes the problem. A detailed explanation follows. At step 7 (see previous comment), with JGroups < 2.2.7, the node A initiates a partition merge and remains the coordinator of the healed partition. Based on this configuration, the DRM layer decides that A has been elected new master, and because is still the current master, S1 continues to run. At step 8, the node B runs into a race condition when as the result of the merge, it receives a view update where A is coordinator while B's replicant map hasn't been updated yet and it only contains B. Due to this race condition, B decides that has been elected new master, and because is still the current master, S2 continues to run. JGroups 2.2.7 solves the problem by having A leave the partition and re-join, which is the correct thing to do anyway. The steps 7 and 8 in this case are: At step 7, A leaves and re-joins the partition, its DRM layer decides that A cannot be master anymore, so S1 stops. At step 8, B stays master, so S2 continues to run. > InvalidDestinationException after Cluster Master failover > --------------------------------------------------------- > > Key: JBAS-35 > URL: http://jira.jboss.com/jira/browse/JBAS-35 > Project: JBoss Application Server > Type: Bug > Components: Clustering > Versions: JBossAS-3.2.5 Final > Reporter: Scott M Stark > Assignee: Ovidiu Feodorov > Fix For: JBossAS-3.2.7 Final > Attachments: JBossFailOverError.txt > > > See the attachment JBossFailOverError.txt for more details. > Abstract: I have a 2 node cluster running JbossMQHA. If the Master node is > disconnect from the Network, the slave assumes the master role. But when the > original Master comes back on-line, it reassumes the Master role, but my > subscriber is never notified of the 2nd fail-over and I am unable to publish > on the slave for about 5 minutes. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - If you want more information on JIRA, or have a bug to report see: http://www.atlassian.com/software/jira ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://productguide.itmanagersjournal.com/ _______________________________________________ JBoss-Development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development