[Pacemaker] Clone resource dependency issue - undesired restart of dependent resources

[Andrew Beekhof]

On Thu, Apr 7, 2011 at 10:37 AM, Christoph Bartoschek
<ponto at pontohonk.de> wrote:
> Christoph Bartoschek wrote:
>
>> Andrew Beekhof wrote:
>>
>>> 2011/4/4 Christoph Bartoschek <bartoschek at gmx.de>:
>>>> Andrew Beekhof wrote:
>>>>
>>>>> 2011/4/1 Christoph Bartoschek
>>>>> <bartoschek at gmx.de>:
>>>>>> Andrew Beekhof wrote:
>>>>>>
>>>>>>> You didn't mention a version number... I think you'll be happier with
>>>>>>> 1.1.5 (I recall fixing a similar issue).
>>>>>>
>>>>>> I see a similar problem with 1.1.5.  I have a two node NFS server
>>>>>> setup.
>>>>>>
>>>>>> The following happens:
>>>>>>
>>>>>> 1. The resources run on node A.
>>>>>> 2. I put node A into standby.
>>>>>> 3. The resources are migrated to node B.
>>>>>> 4. I make node A online again.
>>>>>> 5. The services are stopped on node B.
>>>>>> 6. The services are started on node B.
>>>>>>
>>>>>> In my opinion starting node A again should not cause the services to
>>>>>> stop and start again.
>>>>>
>>>>> No argument there. Can you include a crm_report archive covering the
>>>>> time between 1 and 6 please?
>>>>
>>>> Could you please show me how the call to crm_report should look to give
>>>> you the information you are interessted in?
>>>
>>> crm_report -h should contain sufficient information
>>
>> I hope the resulting file is ok. I called crm_report this way:
>>
>> crm_report  -f "2011-04-04 12:15:00" \
>>             -t "2011-04-04 12:18:00" \
>>             -n "laplace ries"
>>
>
> Is the report enough or is more information needed?

Should be enough.
I just checked the latest 1.1 code and the clones are no longer restarted.

For the rest, you can get the result you want by using 0 for the score
in your ordering constraints.
The current behavior is a bug, the services should only restart if the
clone instance on the same node restarts.

Question though, do the nfs clients need to restart if the server
component does?