[Pacemaker] Cluster Shutdown

[Lars Ellenberg]

On Thu, Nov 04, 2010 at 11:01:29AM +0100, Samuel CUELLA wrote:
>  Hi List,
> 
> I'm currently trying to implement a complete cluster shutdown to
> react to a 'catastrophic' failure like a power outage.
> The problem is that if I try to shut it down node by nodes,
> resources are migrated. Of course, this is the correct behavior.
> 
> I searched the internet and the official clusterlabs documentation
> and the only thing a found is a manual procedure to upgrade the
> cluster engine
> (http://www.clusterlabs.org/wiki/Upgrade#Complete_Cluster_Shutdown).
> However, it's not what I'm looking for.
> I'd like the cluster nodes to be instructed that a shutdown
> procedure is ongoing and therefore that there is no need to start
> migrating resources
> over and over.

# crm configure property stop-all-resources=true
wait for that to happen.
shutdown all nodes.

> I know that Novell as addressed this issue in NCS (Novell Cluster
> Services). A 'cluster shutdown' command will do the trick.
> Is there such a feature in Pacemaker/Heartbeat that I missed ?
> If not, is there plans to implement this on the pacemaker roadmap ?
> 
> Cheers,
> Samuel

-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.