[Pacemaker] monitor on disabled nodes

Radoslaw Garbacz radoslaw.garbacz at xtremedatainc.com
Thu Sep 19 23:18:23 UTC 2013


On Wed, Sep 18, 2013 at 8:58 PM, Andrew Beekhof <andrew at beekhof.net> wrote:
>
> On 19/09/2013, at 2:13 AM, Radoslaw Garbacz <radoslaw.garbacz at xtremedatainc.com> wrote:
>
>> Hi,
>>
>> I have a question regarding the "monitor" operation on disabled nodes.
>>
>> I noticed that this operation is called even, when an agent is disabled for a node. Is it an indented behavior
>
> Yes.  We have to check the resource's status on that node to make sure it isn't already running there.
>
>> or is there something wrong with my configuration?
>>
>> Another question is, if it is intended, whether there is any way to distinguish such a call from a normal health check of a running agent (like "ocf_is_probe" for "validate-all" operation)?
>
> Be careful not to unconditionally return 'stopped' for probes.  There are plenty of situations where we could probe for a resource with the expectation that it _is_ running.
>

I check the status of the resource (in this case - nfs - it is
important indeed), just have to avoid certain configuration checks if
not on the head.

Thank you for your help,


-- 
Best Regards,

Radoslaw Garbacz
XtremeData Incorporation




More information about the Pacemaker mailing list