[neutron][rabbitmq][oslo] Neutron-server service shows deprecated "AMQPDeprecationWarning"

Ken Giusti kgiusti at gmail.com
Fri Jan 10 15:24:51 UTC 2020


On Thu, Jan 9, 2020 at 2:27 PM Ben Nemec <openstack at nemebean.com> wrote:

>
>
> On 1/9/20 5:57 AM, Amjad Kotobi wrote:
> > Hi Ben,
> >
> >> On 7. Jan 2020, at 22:59, Ben Nemec <openstack at nemebean.com> wrote:
> >>
> >>
> >>
> >> On 1/7/20 9:14 AM, Amjad Kotobi wrote:
> >>> Hi,
> >>> Today we are facing losing connection of neutron especially during
> instance creation or so as “systemctl status neutron-server” shows below
> message
> >>> be deprecated in amqp 2.2.0.
> >>> Since amqp 2.0 you have to explicitly call Connection.connect()
> >>> before using the connection.
> >>> W_FORCE_CONNECT.format(attr=attr)))
> >>> /usr/lib/python2.7/site-packages/amqp/connection.py:304:
> AMQPDeprecationWarning: The .transport attribute on the connection was
> accessed before
> >>> the connection was established.  This is supported for now, but will
> >>> be deprecated in amqp 2.2.0.
> >>> Since amqp 2.0 you have to explicitly call Connection.connect()
> >>> before using the connection.
> >>> W_FORCE_CONNECT.format(attr=attr)))
> >>
> >> It looks like this is a red herring, but it should be fixed in the
> current oslo.messaging pike release. See [0] and the related bug.
> >>
> >> 0: https://review.opendev.org/#/c/605324/
> >>
> >>> OpenStack release which we are running is “Pike”.
> >>> Is there any way to remedy this?
> >>
> >> I don't think this should be a fatal problem in and of itself so I
> suspect it's masking something else. However, I would recommend updating to
> the latest pike release of oslo.messaging where the deprecated feature is
> not used. If that doesn't fix the problem, please send us whatever errors
> remain after this one is eliminated.
> >
> > I checked it out, we are having the latest pike Oslo.messaging and it
> still showing the same upper messages. Any ideas?
>
> Hmm, not sure then. Are there any other log messages around that one
> which might provide more context on where this is happening?
>
> I've also copied a couple of our messaging folks in case they have a
> better idea what might be going on.
>
>
This deprecation warning should not result in a connection failure.  After
the amqp connection code issues that warning it immediately calls connect()
for you, which establishes the connection.

What concerns me is that you're hitting that warning in latest Pike.  That
version of oslo.messaging should no longer trigger that warning.  If at all
possible can you get a traceback when the warning is issued?

When did these failures start to occur?  Did something change -
upgrade/downgrade, etc?

Otherwise if you can reproduce the problem can you get a debug-level log
trace?

thanks,


> >>
> >>> Thanks
> >>> Amjad
> >
> >
>
>

-- 
Ken Giusti  (kgiusti at gmail.com)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20200110/a4c25b51/attachment.html>


More information about the openstack-discuss mailing list