[ClusterLabs] Pacemaker

[Ken Gaillot]

On 11/11/2016 03:59 AM, Siwakoti, Ganesh wrote:
> Hi Ken Gaillot,
> 
> Sorry for late reply and thank you for response me.
> 
> i’ve some reason to using -inf rules. 
> 
> Two resources group with three nodes (2 Active 1 Standby)
> 
> When resources group PM1 failed in pm1.local it’ll failover to pm3.local
> which is common node can failover both resources group. If PM1 failed
> again in pm3.local it’ll return back to own node pm1.local. problem is
> that in this situation if PM2 failed in pm 2.local It’ll failover to
> pm3.local even PM1 already failing in this node.that’s why I was using
> -inf rules because if resource failed once on one node I want to prevent
> failover any resources group on  same node.but I don’t know how to
> define it, so my configuration of -inf was long, define every resources.
> Have you any ideas about it. Thanks in advance. 

If you want failure of one resource to prevent any other resources from
running on that same node, set on-fail=standby (or fence) on all operations:

http://clusterlabs.org/doc/en-US/Pacemaker/1.1-pcs/html-single/Pacemaker_Explained/index.html#_resource_operations