[Pacemaker] Monitor a resource without the cluster reacting to the result...

[Andrew Beekhof]

On Fri, Mar 27, 2009 at 10:30, Joe Bill <foxycode at yahoo.com> wrote:
>
> --- On Thu, 3/26/09, Andrew Beekhof <beekhof at gmail.com> wrote:
>
>>> .... what command should I type to cause the cluster
>>> to perform a "monitor" operation at a specific check
>>> level on that resource, and return the appropriate
>>> OCF status of the operation, *without* the cluster
>>> reacting like triggering a failover ?
>>
>> Basically this is a bad idea.
>
> :-)
>
> Sorry Andrew, I'm afraid I'm going to ask you to follow this up with some explanation and arguements.
>
>>
>> If you don't want the cluster to react to an event, don't
>> have it perform one.
>>
> Andrew, like any ***serious*** system control software I've seen, the cluster should have switches for testing components ON-LINE.
>
> You don't expect engineers to bring down a nuke plant to test it's components and emergency procedures, do you ?
>
>> I saw later that you want to prevent the cluster
>> from doing anything for the resource, simply set
>> is-managed=false for the resource in question.
>
> This is incorrect. I want the cluster to react to everything BUT to this specific invocation of MONITOR.

Then I don't understand what the purpose of the monitor is as there
will be no other actions for the resource until the monitor completes.

>> Then, when you're done with your maintenance, set it back
>> to true.
>
> This is like turning off all the nuke plant's alarms while the maintenance teams are cleaning up the keyboards in the control room.
>
> I'm sure we can do better than that.
>
>>
>> There may be a valid need for this one day, but blocking
>> other actions for the resource is definitely not it.
>>
> Blocking but at most to an extent, no ?

What is an extent in this context?

> Won't the cluster NEVER interrupt an ongoing MONITOR operation beyond a certain amount of time ?

Two actions will absolutely never be concurrently performed for a
single resource.
This is enforced by the lrmd.

> Also what happens if, while a MONITOR operation is going on, I issue a:
> "crm_resource -M" on the same resource to migrate it ?
>
> Will the cluster wait until the monitor operation is finished ?

Clearly it must.
Otherwise the monitor would fail.