Thank you Christine and Andrei, I took a look at the corosync quorum policy configuration options, and actually I would need a more conservative approach, i.e. to consider quorum only if all the nodes are present - any node loss is a quorum loss event for me. At present I check it in an agent, but would be helpful if pacemaker took care of this for me.
I know that it is not the requirement pacemaker was designed for (i.e. it does not use the full power of this cluster environment), but for now the application we use cannot handle any node loss. On Tue, Nov 10, 2015 at 2:13 AM, Christine Caulfield <[email protected]> wrote: > On 09/11/15 22:20, Radoslaw Garbacz wrote: > > Hi, > > > > I have a question regarding the policy to check for cluster quorum for > > corosync+pacemaker. > > > > As far as I know at present it is always (excpected_votes)/2 + 1. Seems > > like "qdiskd" has an option to change it, but it is not clear to me if > > corosync 2.x supports different quorum device. > > corosync 2 does not currently support any other quorum devices. But > watch this space .... > > > What are my options if I wanted to configure cluster with a different > > quorum policy (compilation options are acceptable)? > > > > Have a read of the votequorum(5) man page, there are options for > auto_tie_breaker, and maybe others, that might be useful to you. > > > Chrissie > > _______________________________________________ > Users mailing list: [email protected] > http://clusterlabs.org/mailman/listinfo/users > > Project Home: http://www.clusterlabs.org > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf > Bugs: http://bugs.clusterlabs.org > -- Best Regards, Radoslaw Garbacz XtremeData Incorporation
_______________________________________________ Users mailing list: [email protected] http://clusterlabs.org/mailman/listinfo/users Project Home: http://www.clusterlabs.org Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf Bugs: http://bugs.clusterlabs.org
