Hi all, Scenario: - two machines, heartbeat 2.1 in V1 config style, all resources defined in just one group - kill the active node (e.g. echo b >/proc/sysrq-trigger) - standby node tries to take over all the resources - OK - standby fails to start any resource for any reason (e.g. drbd device in inconsistent state), heartbeat releases all the resources - OK - formerly active node comes up again, initiates communication with the peer node and is happy without trying to takeover resources - Not OK
I tried it with 2.1.4 and 2.1.3 with patched DRBD peer outdater, the behavior is exactly the same. Am I expecting something not fundamental from heartbeat if I would like to start resources on the formerly active node when it starts up and detects the standby node is not running the resource group? I know that V1 config style is obsolete, but should not proposed scenario work also in this setup type? At least I do not see any attempt to start of resources. Please find attached corresponding log excerpts from both nodes for the simulated situation. Thanks.
ha-debug-xen9.log
Description: Binary data
ha-debug-xen11.log
Description: Binary data
_______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems