Hi, The latest Pacemaker 1.0 can help our problem which I posted to the following entry. http://developerbugs.linux-foundation.org/show_bug.cgi?id=1990
A split brain under 4 nodes circumstances can be recovered successfully! It seems that these patches have the effect for this behavior. http://hg.clusterlabs.org/pacemaker/stable-1.0/rev/fd4066e9888d http://hg.clusterlabs.org/pacemaker/stable-1.0/rev/7508f211e536 Could it be possible to reflect these patches to Heartbeat 2.1.4 tree? Thanks, Junko > Hi, > > This issue might be related to Bug #1546. > http://developerbugs.linux-foundation.org//show_bug.cgi?id=1546 > > When Heartbeat recovers a split brain, > It tries to handle its instance id. > Some nodes receive an old id which has been gotten before, > Almost all nodes can receive the newest instance id successfully. > But, sometimes, some nodes can not. > > I created a new bugzilla and filed the logs. > http://developerbugs.linux-foundation.org//show_bug.cgi?id=1991 > > 9 nodes received "instance=17" during a split brain. > hac01, hac02, hac03, hac04, hac06 hac08 and hac09 receieved it again after > recovering a split brain, > but they can receive the newset id (like instance=18, 20, 23...) and join > the cluster member. > hac02, hac06 received "instance=17" again, and can notice the DC election, > but they freeze... the newest id doesn't come. > Other nodes would take hac02 and hac06 as OFFLINE node. > > This situation is very rare, so is this some timing bug? > > Best Regards, > Junko Ikeda > > NTT DATA INTELLILINK CORPORATION > > > _______________________________________________ > Linux-HA mailing list > [email protected] > http://lists.linux-ha.org/mailman/listinfo/linux-ha > See also: http://linux-ha.org/ReportingProblems _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
