It may happen we have to re-run the role, for instance in case of scale-out.

We should handle such case in a better way:

• if the VM is already booted, it shouldn't wait for its SSH
• if the VM is already booted, it shouldn't try to configure it
• if the VM is already booted, it should only fetch its MAC addresses

If we're able to do this, we should avoid losing time on those tasks, as well as clean the actions and therefore the log.