[Pacemaker] Three questions...

[Romi Verma]

On Fri, Mar 13, 2009 at 7:23 PM, Andrew Beekhof <beekhof at gmail.com> wrote:

> On Thu, Mar 12, 2009 at 14:11, Dejan Muhamedagic <dejanmm at fastmail.fm>
> wrote:
> > Hi,
> >
> > On Sat, Mar 07, 2009 at 11:40:52AM +0530, Romi Verma wrote:
> >> Thanks Dejan,
> >> it is really good info for me . one more question, suppose we are
> running a
> >> stonith as clone on four nodes cluster. if one nodes fails and it needs
> to
> >> be fenced then who will be elected among rest three nodes to fence this
> >> errant node. how this election happens. is there any priority for nodes
> also
> >
> > No, there is no coordination between nodes. All of them will try
> > to reset the node.
>
> Surely not.
>
> Thats contrary to how I understood stonithd worked.
> I thought, at the very least, the instance on the DC tries to shoot
> the node itself before escalating to other nodes.
> But even then, it was my understanding that it picks one node at a
> time to perform the fencing.


This really make sense.  but  what if the DC itself fails and become errant
node then  which node will be choosen to come forward and fence .

Regards,

_______________________________________________
Pacemaker mailing list
Pacemaker at oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20090314/d40209c4/attachment-0001.htm>