Nova has good logging and will most likely have a pretty good error message to indicate what is going wrong if you have trouble getting instances to launch. You would want to check logs for nova-api, nova-conductor, and nova-scheduler on the control tier, and on each compute node where there is a nova-compute service running. Let's look at what each of these agents does so that you know where to look when something is going wrong.

When a Nova command is executed, it talks to the nova-api service. If an error is received directly when a Nova command is executed, this is generally where you can find more detail beyond the immediate error message returned.

Once a command to act on an instance is accepted into the Nova infrastructure, there are two services that handle passing actions to the compute nodes—nova-scheduler and nova-conductor. Nova-scheduler is just what its name suggests. It handles the decision making as to where to schedule resources across...