[ClusterLabs] Colocation by Node

[Andrei Borzenkov]

02.10.2018 23:49, Ken Gaillot пишет:
...
>>>>
>>>> Is a configuration like this possible? Without creating two
>>>> primitives for 'ocf:esos:scst' and ditching the clone rule? Or is
>>>> the
>>>
>>> No, there's no way to constrain against a particular clone
>>> instance.
>>
>> Hmm ...
>>
>> commit 6ebb25aaa749903b77ad810c8b100745bb3f2a5f
>> Author: Andrew Beekhof <andrew at beekhof.net>
>> Date:   Fri Apr 30 15:01:25 2010 +0200
>>
>>     Medium: PE: Bug lf#2169 - Allow constraints to apply to specific
>> instances of clones (regression tests)
>>
>> It does not appear to be documented anywhere and I am not even sure
>> whether it is enabled by default, but code appears to be there.
> 
> Ah, I remember now running across that when pulling out old code from
> 2.0.
> 
> It's in the testing-ground "pacemaker-next" schema, so someone would
> have to mess with the validate-with property explicitly to try it. I
> doubt it would work reliably with the current clone code. The
> originally proposed use case was colocating different IPs with
> different clone instances.
> 

I guess it can be used for every resource agent that supports
globally-unique option as long as we have some way to associate specific
function with specific instance. But unless this can be done in
semi-automated way like IPaddr2 it is probably easier to simply create
individual primitives instead.