Steven Dake wrote:
Likely a defect. We will have to look into it. Please file a bug as
per instructions on the corosync wiki at www.corosync.org.
FWIW:
Cluster with three nodes, pace-1, pace-2 and pace-3.
pace-1:
Sep 26 10:33:40 pace-1 crmd: [2716]: info: do_shutdown_req: Sending
shutdown request to DC: pace-2
pace-2:
Sep 26 10:33:44 pace-2 corosync[531]: [pcmk ] info: update_member:
Node 1984866496/pace-1 is now: lost
Sep 26 10:33:44 pace-2 corosync[531]: [pcmk ] info:
send_member_notification: Sending membership update 280 to 2 children
Sep 26 10:33:44 pace-2 crmd: [601]: info: ais_status_callback: status:
pace-1 is now lost (was member)
Sep 26 10:33:44 pace-2 crmd: [601]: WARN: match_down_event: No match for
shutdown action on pace-1
So, pace-2 doesn't understand 'shutdown action'. Is pace-1 waiting for
an 'ACK' or something from pace-2, before stopping lrmd and/or crmd?
_______________________________________________
Pacemaker mailing list
Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
