[Pacemaker] Monitoring on master node not running after standby is connected

[Andrew Beekhof]

On 16 Oct 2013, at 12:21 am, Juraj Fabo <juraj.fabo at gmail.com> wrote:

> Juraj Fabo <juraj.fabo at ...> writes:
>> 
>> Hello Andrew
>> 
>> 
>> thank you for the response.
>> 
>> I've patched crmd, cleaned the cluster, done the scenario steps and 
> created crm_report which is attached.
>> 
>> After loading the cluster configuration both nodes were running the IFDS-
> Stateful resource monitoring properly, however after restarting pacemaker on 
> the slave node monitoring on the master node was cancelled
>> 
>> 
>> ifds1 was slave, monitoring of IFDS-Stateful running 11seconds interval
>> ifds2 was master,  monitoring of IFDS-Stateful running 10 seconds 
> intervallog messages in ifds_ha.log
>> 
>> 
>> with best regards 
>> 
>> 
>> 
>> Juraj
>> 
> 
> Hello
> 
> I've downgraded pacemaker in my cluster to version 1.1.8 just to see whether 
> the monitoring of master slave resource will be running on both nodes in 
> parallel.
> It is actually running, after entering HotStandby node the monitoring on 
> hotstandby is monitoring postgresql resource in defined interval and on 
> master in interval defined for role master.
> This time I used my target config with postgresql and other resource, I do 
> not have logs for test cluster with dummy "Stateful" resource.
> 
> Now I'm confused due to fact, that using 1.1.8 pacemaker (in which this bug 
> was fixed) is ok, while 1.1.10 not - at least in my config and according to 
> my understanding. I did not see other reports with similar symptoms, but 
> OTOH with 1.1.8 monitoring is running on both nodes.
> 
> Should I provide some more logs/symptoms from 1.1.10 case to be able to 
> proceed?

I am almost certain you're being affected by:

   https://github.com/beekhof/pacemaker/commit/58962338

Apply that patch and you should be fine.

Apologies for not making the connection sooner :-(

> 
> Thank you in advance
> 
> Juraj
> 
> 
> 
> _______________________________________________
> 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