[Pacemaker] Temporarely suspending monitoring

[Bart Coninckx]

On Thursday 12 August 2010 00:48:32 Dejan Muhamedagic wrote:
> Hi,
> 
> On Thu, Aug 12, 2010 at 12:14:59AM +0200, Bart Coninckx wrote:
> > On Wednesday 11 August 2010 23:55:42 Bart Coninckx wrote:
> > > On Wednesday 11 August 2010 23:01:22 Vince Gabriel wrote:
> > > > > -----Original Message-----
> > > > > From: Bart Coninckx [mailto:bart.coninckx at telenet.be]
> > > > > Sent: Wednesday, August 11, 2010 10:49 AM
> > > > > To: pacemaker at oss.clusterlabs.org
> > > > > Subject: [Pacemaker] Temporarely suspending monitoring
> > > > >
> > > > > Hi,
> > > > >
> > > > > We're using the Xen resource agents with an operation "monitor"
> > > > > that repeats
> > > > > every 60 seconds.
> > > > > For backing up the Xen machines, we use "xm save" and "xm restore"
> > > >
> > > > which
> > > >
> > > > > takes
> > > > > them offline for a short amount of time (and copies the memory
> > > >
> > > > contents to
> > > >
> > > > > a
> > > > > file). Of course, when the monitor check happens right after the
> > > > > Xen
> > > >
> > > > guest
> > > >
> > > > > has
> > > > > been taken offline, it will try and restart it, possibly before the
> > > >
> > > > Xen
> > > >
> > > > > guest
> > > > > has been restored in the proper way.
> > > > > Is there a way to avoid this? For instance by temporarely
> > > > > suspending
> > > >
> > > > the
> > > >
> > > > > monitoring?
> > > >
> > > > To stop monitoring:
> > > >
> > > > #crm resource unmanage <rsc>
> > > >
> > > > To start monitoring:
> > > >
> > > > #crm resource manage <rsc>
> > > >
> > > > > Thx!
> > > > >
> > > > >
> > > > > B.
> > > > >
> > > > > _______________________________________________
> > > > > 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://developerbugs.linux-
> > > > > foundation.org/enter_bug.cgi?product=Pacemaker
> > > >
> > > > _______________________________________________
> > > > 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://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacem
> > > >aker
> > >
> > > Hi, thx for the reply - I'm using Heartbeat with CRM, not really
> > > Pacemaker. I figured things wouldn't be too different between the two,
> > > but it seems they are. I've looked at the crm_resource command and
> > > "crm_resource -p is_managed" seems to do the same thing. Am I correct?
> > >
> > > thx again,
> > >
> > >
> > > B.
> > >
> > > _______________________________________________
> > > 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://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemak
> > >er
> >
> > Hmpf, in this thread Andrew Beekhof explains that stopping to manage a
> > resource does not stop monitoring:
> >
> > http://readlist.com/lists/lists.linux-ha.org/linux-ha/0/2861.html
> >
> > maybe I need to try the linux-ha list
> 
> There's another way to disable monitoring, by setting the
> "enabled" attribute to false for the operation. It can be done
> with crm_resource, but I can't recall the magic incantation.
> 
> Thanks,
> 
> Dejan
> 
> > B.
> >
> > _______________________________________________
> > 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://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
> 
> _______________________________________________
> 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://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
> 

As usual it is surprisingly easy:

xen1:~ # cibadmin -M -X '<op id="op_mon_lx05" name="monitor" interval="60" 
timeout="30" start_delay="0" disabled="false" role="Started"/>'
xen1:~ # cibadmin -M -X '<op id="op_mon_lx05" name="monitor" interval="60" 
timeout="30" start_delay="0" disabled="false" role="Stopped"/>'

B.