[Pacemaker] Pacemaker cluster with different operating systems

Lars Marowsky-Bree lmb at suse.com
Thu Mar 7 09:32:24 EST 2013


On 2013-03-07T09:22:27, Osman Findik <Osman.Findik at telenity.com> wrote:

> Hi Lars,
> We think that the third node will be in standby, so only corosync and pacemaker processes will execute and no resource will be executed on this node by the cluster.

That is true, but it will still participate in the corosync protocol
(meaning an instability of the quorum node could translate to
instability of your cluster); and if you really run pacemaker there,
you'll have to keep it updated as well so that it doesn't become the
eternal DC.

A third node makes perfect sense if you need a third node in your
cluster for redundancy and load reasons. It's not good to add one at
this layer if all you need is better quorum.

> For other approaches, I only know boothd. And what I know is it is suitable for geo-redundancy purposes. I will look for fence_sanlock.

boothd can also be used as a 1:1 fail-over scenario. Admittedly, it's
overkill and makes the configuration more complex. I'd probably go with
external/sbd, but then, I wrote that ;-)


Regards,
    Lars

-- 
Architect Storage/HA
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde





More information about the Pacemaker mailing list