[Pacemaker] Upgrading a master salve resource agent, Maintenance best practices

David Vossel dvossel at redhat.com
Mon Nov 19 19:42:39 UTC 2012



----- Original Message -----
> From: "Jesse Hathaway" <jesse.hathaway at getbraintree.com>
> To: "The Pacemaker cluster resource manager" <pacemaker at oss.clusterlabs.org>
> Sent: Monday, November 19, 2012 1:06:34 PM
> Subject: Re: [Pacemaker] Upgrading a master salve resource agent, Maintenance best practices
> 
> 
> 
> On Mon, Nov 19, 2012 at 12:55 PM, David Vossel < dvossel at redhat.com >
> wrote:
> 
> 
> 
> 
> I do not have much experience managing pacemaker deployments, but
> this is what I would try.
> 
> Set the resource option, is-managed, to false.

Ha,  I got this backwards.  Should be 'true'. Default is 'false'.  I'm not sure where my mind was, hopefully this was an obvious mistake :)

> http://www.clusterlabs.org/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/s-resource-options.html
> 
> Go in and make whatever changes you need to make to both the cib and
> the active resource. After you are done, set the resource's
> is-managed option back to true, or remove it from the config
> entirely since the default is true.
> 
> Alternatively you could set the entire cluster in maintenance mode by
> setting the cluster option, maintenance-mode, to true. This will
> automatically make all the resources un-managed until maintenance
> mode is turned off again.
> 
> 
> Unfortunately, placing the resource or cluster into maintenance mode
> does not help, since when the resource is managed again pacemaker
> notices that the resource has changed and the policy engine makes
> any necessary actions, e.g. in this case restarting the resource.

What version of pacemaker are you using?

I just tested this using the Stateful resource with the master branch at commit 
e4f9338b6e33906117159657a27ff1c03746e0ad.

I made a multistate resource and added some fake instance attributes to it.

After that I put the resource in un-managed mode, changed some instance attributes around, and then set the resource back into managed mode.  The master resource did not restart for me. It picked back up at monitoring the Master resource.

-- Vossel 


> --
> Jesse Hathaway, Systems Engineer
> Braintree
> 917-418-8423
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started:
> http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
> 




More information about the Pacemaker mailing list