[ClusterLabs] Antw: Re: Xen OCF RA (Xen_Status_with_Retry()): cnt should be configurable

Ulrich Windl Ulrich.Windl at rz.uni-regensburg.de
Mon Jul 6 06:19:46 UTC 2015


>>> Tom Parker <tparker at cbnco.com> schrieb am 03.07.2015 um 16:34 in Nachricht
<55969D77.9080000 at cbnco.com>:
> xm list shows the running vm even when it's in the pygrub countdown.  I
> have just double checked that in my SLES SP3 environment.

OK, I was wrong: Actually the VM is listed with state "--p---" while in pygrub. The RA checks for "status 1" or "status 2" (or "[-r][-b][-p]---"). I'm not sure whether "--p---" corresponds to "status 1" or "status 2".

Regards,
Ulrich

> 
> I also think the status with retry was only called in the case where a
> vm that we expect to be running suddenly is not.  It's to work around an
> issue in Xen where a rebooting vm momentarily disappears from xm list. 
> In all cases we have never seen a vm being gone for more than 1 or 2
> seconds.
> 
> Tom
> 
> On 07/03/2015 05:25 AM, Dejan Muhamedagic wrote:
>> Hi,
>>
>> On Fri, Jul 03, 2015 at 08:50:30AM +0200, Ulrich Windl wrote:
>>> Hi!
>>>
>>> I had a look into the Xen RA of SLES11 (resource-agents-3.9.5-0.37.38.19). In 
> Xen_Status_with_Retry() there is a 5-second re-probe interval that cannot be 
> configured.
>>> If you use pygrub with "timeout 8" to start your Xen VMs, I see a potential 
> problem:
>>> While pygrub is counting down the Xen monitor may be unable to detect the 
> VM. Right?
>> Are you sure about that? Shouldn't the guest be listed anyway?
>>
>> If so, then yes, we may need to have this count exposed as a
>> parameter. The reason it wasn't is that it is there just to cover
>> a deficiency in xen which nobody should really care about.
>>
>> Thanks,
>>
>> Dejan
>>
>>> So should that "cnt=5" in Xen_Status_with_Retry() be configurable?
>>>
>>> Regards,
>>> Ulrich
>>>
>>>
>>>
>>> _______________________________________________
>>> Users mailing list: Users at clusterlabs.org 
>>> http://clusterlabs.org/mailman/listinfo/users 
>>>
>>> Project Home: http://www.clusterlabs.org 
>>> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf 
>>> Bugs: http://bugs.clusterlabs.org 
>> _______________________________________________
>> Users mailing list: Users at clusterlabs.org 
>> http://clusterlabs.org/mailman/listinfo/users 
>>
>> Project Home: http://www.clusterlabs.org 
>> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf 
>> Bugs: http://bugs.clusterlabs.org 
> 
> 
> _______________________________________________
> Users mailing list: Users at clusterlabs.org 
> http://clusterlabs.org/mailman/listinfo/users 
> 
> Project Home: http://www.clusterlabs.org 
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf 
> Bugs: http://bugs.clusterlabs.org 








More information about the Users mailing list