[Pacemaker] Dependency Trees

[Andrew Beekhof]

On Fri, Feb 22, 2013 at 11:39 AM, Andrew Beekhof <andrew at beekhof.net> wrote:
> On Fri, Feb 22, 2013 at 1:40 AM, Donald Stahl <don at blacksun.org> wrote:
>>> No.
>>>
>>> [quote]
>>> If you say "colocate A with B" and there is nowhere B is allowed to
>>> run, then A wont be allowed to run either.
>>> But once the cluster has figured out where they go, it doesn't stop
>>> them being started in parallel.
>>> [/quote]
>>>
>>> in this case, A = OraListener1 and B = OraBin1
>>>
>>> You only get the "Start A then start B" part by adding the ordering constraint.
>>
>> I'm not sure I understand where the "sequential" option comes into
>> play with a colocation set then.
>>
>> Or to put it another way- I'm not sure what practical difference there
>> is in these examples:
>>
>> http://clusterlabs.org/wiki/FAQ#Collocation_Sets

I've added some more text to that link, hopefully it provides some
additional clarity.

>>
>> colocation myset inf: app1 app2
>> colocation myset inf: ( app1 app2 )
>
> I'm pretty sure the latter doesn't do anything (bad choice of defaults
> by the crmsh) because the set isn't colocated with anything.
>
> You'd need something as well as the set for it to make sense,  eg.
>
>    colocation myset inf: app0 ( app1 app2 )
>
>> The first seems to say that app1 is dependent on app2 being able to
>> run somewhere.
>>
>> Does the second one simply say that they must not be split up, but if
>> only one can run, that's fine?
>>
>> -Don
>>
>> _______________________________________________
>> 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