Looks like neutron agents didn't become available after RabbitMQ cluster failover ... or failover just didn't complete in time before tests were run: I see multiple `rabbit on node 'rabbit@node-1' down` in rabbitmq logs during the time instances were launched.
Step "7. Wait 5 minutes" should really be "Wait until RabbitMQ cluster is reassembled properly.", which can possibly take more than 5 minutes.
Nova could not spawn instances due to problem with binding VIFs by Neutron: http:// paste.openstack .org/show/ 240272/
All events for this particular instance uuid: http:// paste.openstack .org/show/ 240286/
Looks like neutron agents didn't become available after RabbitMQ cluster failover ... or failover just didn't complete in time before tests were run: I see multiple `rabbit on node 'rabbit@node-1' down` in rabbitmq logs during the time instances were launched.
Step "7. Wait 5 minutes" should really be "Wait until RabbitMQ cluster is reassembled properly.", which can possibly take more than 5 minutes.