[Pacemaker] Colocation of multistate resource and group

[Evgeniy Ivanov]

On Tue, Jan 18, 2011 at 6:37 PM, Dejan Muhamedagic <dejanmm at fastmail.fm> wrote:
> Hi,
>
> On Tue, Jan 18, 2011 at 02:52:07PM +0300, Evgeniy Ivanov wrote:
>> Hello,
>>
>> I have a group Gr and multistate resource MasterRsc. I need MasterRsc
>> to be master on the node where Gr works.
>> In MasterRsc::start() I use crm_master to set higher value if Gr works
>> on this node and it works fine.
>> But I have a problem with failovers and colocation:
>>
>> <rsc_colocation id="master_dummy-with-loggroup" rsc="MasterRsc"
>> rsc-role="Master" with-rsc="Gr" score="INFINITY"/>
>
> Add: with-rsc-role="Started"

With this option after group (now I use single resource for test)
makes failover, but master resource doesn't.

>
>> I did failover of Gr and then added this colocation. What happened is
>> that Gr started failover to the node with MasterRsc:master, but not
>> MasterRsc:master to another group.
>
> You lost me here.

Ops, I meant another node, not group :-)

>
>> Is it expected (PM 1.0.3)? What's correct way to achieve this?
>
> Consider upgrading.

Yeah, I will. But for now I need to make it work on 1.0.3.

>
> Thanks,
>
> Dejan
>
>
>>
>>
>> --
>> Evgeniy Ivanov
>>
>> _______________________________________________
>> 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
>



-- 
Evgeniy Ivanov