[ClusterLabs] corosync[3520]: [CPG ] *** 0x55ff99d211c0 can't mcast to group dlm:ls:lvm_testVG state:1, error:12
Jan Friesse
jfriesse at redhat.com
Tue Jan 5 08:17:17 EST 2021
Ulrich,
> Hi!
>
> In my test cluster using UDPU(!) I saw this syslog message when I shut down a VG:
> Nov 30 13:38:28 h16 pacemaker-execd[3681]: notice: executing - rsc:prm_testVG_activate action:stop call_id:71
> Nov 30 13:38:28 h16 LVM-activate(prm_testVG_activate)[7265]: INFO: Deactivating testVG
> Nov 30 13:38:28 h16 corosync[3520]: [CPG ] *** 0x55ff99d211c0 can't mcast to group dlm:ls:lvm_testVG state:1, error:12
> Nov 30 13:38:28 h16 kernel: dlm: lvm_testVG: leaving the lockspace group...
> Nov 30 13:38:28 h16 kernel: dlm: lvm_global: dlm_recover 7
> Nov 30 13:38:28 h16 kernel: dlm: lvm_global: remove member 118
> Nov 30 13:38:28 h16 kernel: dlm: lvm_testVG: group event done 0 0
> Nov 30 13:38:28 h16 kernel: dlm: lvm_testVG: release_lockspace final free
> Nov 30 13:38:28 h16 kernel: dlm: lvm_global: dlm_recover_members 2 nodes
> Nov 30 13:38:28 h16 kernel: dlm: lvm_global: generation 3 slots 2 1:116 3:119
> Nov 30 13:38:28 h16 kernel: dlm: lvm_global: dlm_recover_directory
> Nov 30 13:38:28 h16 kernel: dlm: lvm_global: dlm_recover_directory 0 in 0 new
> Nov 30 13:38:28 h16 LVM-activate(prm_testVG_activate)[7288]: INFO: testVG: deactivated successfully.
> Nov 30 13:38:28 h16 pacemaker-execd[3681]: notice: prm_testVG_activate stop (call 71, PID 7258) exited with status 0 (execution time 251ms, queue time 0ms)
> Nov 30 13:38:28 h16 pacemaker-controld[3684]: notice: Result of stop operation for prm_testVG_activate on h16: ok
>
> What kind of prtoblem is that, and what may be causing this?
This happens when application calls cpg_mcast_joined to the group which
doesn't exists - so application doesn't called cpg_join yet or (that's
probably the case here) after it called cpg_leave. No matter what, this
is probably the problem in dlm_controld so best is to ask dlm expert.
Regards,
Honza
>
> Regards,
> Ulrich
>
>
>
> _______________________________________________
> Manage your subscription:
> https://lists.clusterlabs.org/mailman/listinfo/users
>
> ClusterLabs home: https://www.clusterlabs.org/
>
More information about the Users
mailing list