Troubleshooting installation

Troubleshooting an installation of Nova is similar to the ones described in the previous chapters. While we have still not spun up an instance, let's take a look at some failure scenarios when we do spin an instance up and it refuses to come up.

Note

Please remember that the Nova log files are situated in the /var/log/nova/ directory of the nodes where any of the Nova services are installed.

Here are some of the preliminary checks:

  1. Check the Keystone log to see an authentication or token failures.
  2. Check whether RabbitMQ is functional and reachable from all the different compute nodes.
  3. Check whether the installation and configuration have been done properly.

If the authentication was not a problem, we follow the following procedure:

  1. Check nova-api.log on the controller node to check whether the request was properly received. If the request was not properly handled, you will see the reason here.
  2. Execute the nova list command and note down the instance ID of the system.
  3. Execute the nova console-log <uuid of the VM> command to see the logs.

These logs should point you to the right direction. If the system is stuck at building or scheduling, and the VM has not begun to be built yet, then there can be two possible reasons:

  • Not enough resources
  • Issues with the hypervisor

The first log file to be checked will be of the nova scheduler; then check nova-compute.log on the compute node. These are general guidelines and further troubleshooting may be needed on the hypervisor side.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset