[ClusterLabs] Simple clarification regarding pacemaker
kwenning at redhat.com
Wed Apr 27 02:29:59 EDT 2016
On 04/26/2016 09:09 PM, K Aravind wrote:
> Thank you for the quick responses :)
> I starting to understand :)
> One more quick question.
> Let's say I have a 2 node cluster with stonith-enabled=false and no
> quorum policy = ignore
> And master-max=1
> Now connection between nodes went down.
> Now a split brain happened
> So now both nodes are active
> Now connetion came back again.
> Now does pacemaker see that 2 nodes are active and since master-max is
> 1 would it demote one of the nodes ?
Pacemaker will. But this is probably not gonna solve your problem as the
fact who is primary and who secondary
(master/slave role in pacemaker-terms) doesn't say anything about from
which of the disks which part of the
data is being taken. Primary just means that the drbd-device is being
accessed from that side (btw. can be on both
sides provided you are using a filesystem that can cope with that).
Although it is not recommended drbd can be configured to automatically
recover - by deletion of the changes on
one side done during split-brain-phase.
> Users mailing list: Users at clusterlabs.org
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Users