Troubleshooting Contrail Nova and Cinder Services Installation
When installing Contrail directly from the provided ISO images for a single node installation, the following server connection error messages may display in the Nova/scheduler.log or Cinder/scheduler.log:
Nova/scheduler.log
ERROR nova.openstack.common.rpc.common [-] AMQP server on localhost:5672 is unreachable: [Errno 111] ECONNREFUSED. Trying again in 1 seconds.
Cinder/scheduler.log
ERROR [cinder.openstack.common.rpc.impl_qpid] Unable to connect to AMQP server: client: 0-10, server: 9-1. Sleeping 60 seconds
To correct the server connection errors, you must reset the rpc_backend server value in both the Nova and Cinder configuration files:
- Review the /etc/nova/nova.conf file.
- Comment out the rpc_backend value.
- Restart the nova-scheduler.
- Review the /etc/cinder/cinder.conf file.
- Comment out the rpc_backend value so that it uses the default RabbitMQ value, or set to the following:
rpc_backend = cinder.openstack.common.rpc.impl_kombu
- Restart both the cinder-api and cinder-scheduler.
NOTE: Both Nova and Cinder services do not report a connected state and do not generate a "Connected to..." log entery in their logs.
For a quick-start video of troubleshooting Contrail services, click Troubleshooting Contrail Services for a Contrail Learning Byte YouTube video.