[Openstack] Nova connects to rabbitmq on localhost during migrate (ocata)

Кирилл Беспалов k.besplv at gmail.com
Thu Apr 6 14:32:08 UTC 2017


You enable notification driver (via driver = messagingv2), but you are
missing to set transport_url  in notifications config section:

[oslo_messaging_notifications]
transport_url = rabbit://openstack:XXXXXXXXXX@openstack2.example.com

2017-04-06 16:54 GMT+03:00 Jens Rosenboom <j.rosenboom at x-ion.de>:
> Sorry for forgetting to CC the ML on my first reply.
>
> 2017-04-06 13:19 GMT+00:00 Kim Gert Nielsen <kgn at one.com>:
>> On Thu, Apr 06, 2017 at 12:49:23PM +0000, Jens Rosenboom wrote:
>>> 2017-04-06 11:37 GMT+00:00 Kim Gert Nielsen <kgn at one.com>:
>>> > Hi,
>>> >
>>> > I have just test my ocata installation and most of it seems to be fine.
>>> > But when resizing an instance I get the error:
>>> >
>>> > 7447a2-9602-4d15-8196-9151af6a0e86] AMQP server on 127.0.0.1:5672 is
>>> > unreachable: [Errno 111] ECONNREFUSED. Trying again in 32 seconds.
>>> > Client port: None
>>> >
>>> > on the compute node. The weird thing is that I have configured a
>>> > transport_url:
>>> >
>>> > <nova.conf from the compute node>
>>> > [DEFAULT]
>>> ...
>>> > transport_url = rabbit://openstack:XXXXXXXXXX@openstack2.example.com
>>>
>>
>> Hi,
>>
>> Thanks for responding so fast :)
>>
>>> This is taken from the nova.conf during creation of your cell, see
>>> https://docs.openstack.org/developer/nova/cells.html#fresh-install
>>> Do you have the same config file on your controller-node and was it in
>>> place when you created the cell. As you see in the docs you can also
>>> override it in the command.
>>
>> I use chef so I have the same config file (With the exception of $myip)
>> on the controllers and I have made the steps for the cells only I
>> haven't done any configuration using cells.
>
> Oh, a chef user, nice. Do you use your own cookbooks or the ones from upstream?
>
> For upstream, I only recently added cells creation here:
> https://review.openstack.org/#/c/443711/12/Rakefile
> But that is only a first step in order to get the integration tests
> working, for a proper deployment one would need to make this
> idempotent and such.
> Still I think running Ocata without cells set up will be difficult.
>
>> How do you see that this 'localhost' value comes from the cells
>> misconfiguration ?
>
> In fact that was more guessing than seeing. You might want to check
> your nova_api database and look at the cell_mappings table there to
> confirm.
>
> _______________________________________________
> Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
> Post to     : openstack at lists.openstack.org
> Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack




More information about the Openstack mailing list