[Pacemaker] Why monitor fails in my RA

[Dejan Muhamedagic]

On Mon, May 14, 2012 at 01:01:48PM +0200, Greg wrote:
> W dniu 05/09/12 13:57, Dejan Muhamedagic pisze:
>> Hi,
>>
>> On Wed, Apr 25, 2012 at 10:41:05PM +0200, Greg wrote:
>>>
>>> Hi,
>>>
>>> I try to write redis resources agent working in master-slave. My
>>
>> Are you aware of a pull request for one redis resource agent:
>>
>> https://github.com/ClusterLabs/resource-agents/pull/37
>>
>> It's been there a while, blocked mainly because it uses debian
>> specific daemon start/stop machinery.
>>
>> Thanks,
>>
>> Dejan
>>
> [cut]
>
> Thx for pointing me on this script. My environment is slightly different  
> so I can't use this script without modification.

This premise doesn't sound right. An RA should be generic enough
to fit in any environment. Having said that, I'm not implying
that either of these RAs is wrong, but that possibly one or the
other are lacking in flexibility.

> I've notice that your  
> script in this scenario :
> - master is promoted and goes down (for example OOM, admin mistake etc.)
>
> Your script will reply: OCF_ERR_GENERIC
>
> My script wil reply: OCF_FAILED_MASTER
> (from manual this is best option but it's not working the way I've  
> expected because after recovery to slave, and in case of promotion to  
> master node is not monitored in this state anymore)
>
> Can you tell me why have you chose answer OCF_ERR_GENERIC ?

Not my script. Somebody wants to contribute it to our collection
of resource agents. You can ask questions at github, I suppose
that the author is still around.

Thanks,

Dejan

> Thx
>
> --
> Grego
>
>
> _______________________________________________
> 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