Compute node is deployed or updated successfully; later Open vSwitch agent is down

You successfully deploy, update, or apply a fix pack to a compute node, but later the Open vSwitch agent on the compute node is down.

Symptoms

After you successfully deploy, update, or apply a fix pack to a compute node, you are unable to start a new virtual machine.
You check the status of the Open vSwitch agent by running the following command on your controller node:
neutron agent-list
The Open vSwitch agent is down.

Causes

This problem can occur for various reasons, one of which is that there is an error in the /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini file on the compute node.

Resolving the problem

To resolve the problem, try to restart the Open vSwitch agent by running the following command on the compute node:
service neutron-openvswitch-agent restart
Take one of the following actions:
  • If the agent status on the controller is active, try to start the virtual machine.
  • If the agent status is not active, see the /var/log/neutron/openvswitch-agent.log file on the compute node.
If you applied a fix pack to multiple PowerKVM compute nodes, it is likely that the agent is down on all of them. You can resolve the problem similarly for each compute node.