[Pacemaker] colocation and ordering for groups/clones/instances

Andrew Beekhof andrew at beekhof.net
Thu Oct 6 03:01:42 UTC 2011


On Thu, Oct 6, 2011 at 2:47 AM, gustavo panizzo <gfa> <gfa at zumbi.com.ar> wrote:
> On Wed, Oct 05, 2011 at 10:45:32AM +1100, Andrew Beekhof wrote:
>> On Wed, Oct 5, 2011 at 10:21 AM, gustavo panizzo <gfa> <gfa at zumbi.com.ar> wrote:
>> > colocation col_name +inf: clone2 clone
>> > colocation col_name +inf: group clone2

also doesn't do what you think. don't create colocation loops.
pretty sure you just want:
   colocation col_name +inf: group clone2

>> > order ord_name +inf: clone2 group symmetrical=false
>> > order ord_name +inf: clone clone2 symmetrical=false
>>
>> drop the symmetrical=false parts
> why? the oposite order won't work (VG before clvm)

it doesn't mean the opposite for starts.  "clone2 group" means start
clone2 before starting group and stop group before stopping clone2

> i've dropped it and the issue still happening
>
>>
>> >
>> >
>> > when i restart a machine (not the one with RG on it), clone starts on it
>> > RG gets stopped where is running until clone is fully up
>>
>> that doesn't sound right. can you include the result of cibadmin -Ql
>> when the cluster is in this state?
> attached, in this moment node2 is entering the cluster, and the
> resource group qsprod (it was running) went down partially

I'll have a look in a bit

>
> thanks
>>
> --
> 1AE0 322E B8F7 4717 BDEA  BF1D 44BB 1BA7 9F6C 6333
>
>
> _______________________________________________
> 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
>
>




More information about the Pacemaker mailing list