Comment 4 for bug 1504725

I am also hitting the same bug.

I can restart the rabbitmq-server multiple time on one of the controller but if I also restart the second rabbitmq-server in the HA pair, some openstack services will fill their logs files with almost the same error. The impacted services are:

neutron-dhcp-agent
neutron-l3-agent
neutron-metadata-agent
neutron-openvswitch-agent
openstack-nova-compute (exceptions are occurring at a slower rate)

Similar bug report:
https://bugs.launchpad.net/neutron/+bug/1496774

[root@openstack1 ~]# pip freeze | grep oslo.messaging
oslo.messaging==1.8.3

RPMs are from RDO/kilo
[root@openstack1 ~]# rpm -qa | grep oslo-messaging
python-oslo-messaging-1.8.3-1.el7.noarch

# cat /var/log/neutron/openvswitch-agent.log
http://paste.openstack.org/show/478217/

nova-compute will also throws similar stack traces at a slower rate

# cat /var/log/nova/nova-compute.log
http://paste.openstack.org/show/478218/

Relevant parts of neutron.conf
http://paste.openstack.org/show/478219/

Relevant parts of nova.conf
http://paste.openstack.org/show/478220/

rabbitmq.config
http://paste.openstack.org/show/478221/

# rabbitmqctl -p openstack list_queues name policy
http://paste.openstack.org/show/478222/