RabbitMQ Openstack services using the wrong queues

Issue

After an update, or an outage, or an involuntary service restart you may get intermittent issues when creating or updating resources (ex: instances, neutron ports, volumes, stacks and so on)

You may get the following ERROR messages while trying to update or create a new resource (instance in this case):

2017-06-23 20:25:55.414 25527 ERROR oslo.messaging._drivers.impl_rabbit [req-cef7a1d7-a635-42d6-8c4e-b5d47e2bfcee - - - - -] Failed to publish message to topic 'reply_d32bd2f977374abcb2acdb762423ac05': The exchange reply_d32bd2f977374abcb2acdb762423ac05 to send to reply_d32bd2f977374abcb2acdb762423ac05 still doesn't exist after 60 sec abandonning...
2017-06-23 20:25:55.415 25527 WARNING oslo_messaging._drivers.amqpdriver [req-cef7a1d7-a635-42d6-8c4e-b5d47e2bfcee - - - - -] reply_d32bd2f977374abcb2acdb762423ac05 doesn't exists, drop reply to 92d4cbced29f491d8ab0b513272d32ea

These resources failures could be anything from a stack update, a volume attach or an instance create.