[ClusterLabs] "Connecting" Pacemaker with another cluster manager

[Timo]

On 05/23/2017 09:44 AM, Kristoffer Grönlund wrote:
> Timo <timo at kroenchenstadt.de> writes:
> 
>> Hi,
>>
>> I have a proprietary cluster manager running on a bunch (four) of nodes.
>> It decides to run the daemon for which HA is required on its own set of
>> (undisclosed) requirements and decisions. This is, unfortunately,
>> unavoidable due to business requirements.
>>
>> However, I have to put also Pacemaker onto the nodes in order to provide
>> an additional daemon running in HA mode. (I cannot do this using the
>> existing cluster manager, as this is a closed system.)
>>
>> I have to make sure that the additional daemon (which I plan to
>> coordinate using Pacemaker) only runs on the machine where the daemon
>> (controlled by the existing, closed cluster manager) runs. I could check
>> for local VIPs, for example, to check whether it runs on a node or not.
>>
>> Is there any way to make Pacemaker "check" for existence of a local
>> (V)IP so that I could "connect" both cluster managers?
>>
>> In short: I need Pacemaker to put the single instance of a daemon
>> exactly onto the node the other cluster manager decided to run the
>> (primary) daemon.
> 
> Hi,
> 
> I'm not sure I completely understand the problem description, but if I
> parsed it correctly:
> 
> What you can do is run an external script which sets a node attribute on
> the node that has the external cluster manager daemon, and have a
> constraint which locates the additional daemon based on that node
> attribute.

Exactly what I needed! Thanks!

Timo

> Cheers,
> Kristoffer
> 
>>
>> Best regards,
>>
>> Timo
>>
>> _______________________________________________
>> Users mailing list: Users at clusterlabs.org
>> http://lists.clusterlabs.org/mailman/listinfo/users
>>
>> Project Home: http://www.clusterlabs.org
>> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>> Bugs: http://bugs.clusterlabs.org
>>
>