Open Stack

I believe the issue is that there should be an additional:

grep glance

in the command line that checks for glance queues...

-jay

On 12/03/2012 08:21 PM, Mark McClain wrote:
> I've been digging some more and this bug is an unintended side-effect of this patch:
> 
> https://review.openstack.org/#/c/15735/
> 
> The check for durable exchanges will always evaluate to True because there are certain exchanges that are always durable:
> 
> On a fresh rabbitmq server:
> 
> vagrant at vagrant-precise:/etc/glance$ sudo rabbitmqctl list_exchanges name durable | grep true | wc -l
> 8
> vagrant at vagrant-precise:/etc/glance$ sudo rabbitmqctl list_exchanges name durable | grep true
> amq.rabbitmq.trace	true
> amq.rabbitmq.log	true
> amq.match	true
> amq.headers	true
> amq.topic	true
> amq.direct	true
> amq.fanout	true
> 	true
> 
> 
> Now that Glance has been fixed, shouldn't we remove this work around from devstack?
> 
> mark
> 
> 
> On Nov 30, 2012, at 4:18 PM, Russell Bryant <rbryant at redhat.com> wrote:
> 
>> On 11/30/2012 04:07 PM, Russell Bryant wrote:
>>> On 11/30/2012 01:53 PM, Gary Kotton wrote:
>>>> On 11/29/2012 08:02 PM, Russell Bryant wrote:
>>>>> On 11/29/2012 12:51 PM, Gary Kotton wrote:
>>>>>> The issue is how Quantum was listening on the notifications.
>>>>>> https://github.com/openstack/quantum/blob/master/quantum/agent/rpc.py#L89
>>>>>>
>>>>>> This was the same name as that used by glance - "notifications"
>>>>> but what I'm saying is that what you're talking about is the *queue*.
>>>>> The *exchange* is something different.  Queues on different *exchanges*
>>>>> are completely independent of each other.  Quantum and Glance should be
>>>>> using different exchanges and thus completely independent.  That is how
>>>>> this is all supposed to work.
>>>>>
>>>> I agree with you, but in reality it is not the case.
>>>>
>>>
>>> I think I see the problem.  See this change:
>>>
>>>   https://review.openstack.org/#/c/12518/
>>>
>>> This change is in both master and stable/folsom of oslo-incubator.  It
>>> has made its way into projects without making the change that the big
>>> warning in the commit message describes.  :-)
>>>
>>
>> so were you getting this error when testing with stable/folsom?  quantum
>> master seems like it should be ok and using the 'quantum' exchange.
>> Here's a patch for quantum's stable/folsom:
>>
>>    https://review.openstack.org/#/c/17236/
>>
>> and nova master and stable/folsom:
>>
>>    https://review.openstack.org/#/c/17233/
>>    https://review.openstack.org/#/c/17234/
>>
>> -- 
>> Russell Bryant
>>
>> _______________________________________________
>> OpenStack-dev mailing list
>> OpenStack-dev at lists.openstack.org
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
> 
> 
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>