[Pacemaker] N+1 and equal priority resource groups

Igal Baevsky ibaevsky at marketfactory.com
Mon Apr 28 14:59:55 UTC 2014


> node-2 fails, group-2 moves to node-4
> node-2 returns
> node-1 fails, group-1 moves to node-2
> 
> and no admin intervention in between.
> 
> The more common use of N+1 is where the +1 is a shared failover.  
> So resources are only moving between their primary node and the +1.
> 
> I probably could make your use-case work, but its unlikely to happen for 
1.1.12
> 
> I've filed it as a feature request for future reference: 
> 
>    http://bugs.clusterlabs.org/show_bug.cgi?id=5209
> 

Thanks Andrew!

I'll track http://bugs.clusterlabs.org/show_bug.cgi?id=5209

In the mean time I'll stick to the classic "N+1 shared failover" setup.






More information about the Pacemaker mailing list