I really love the concept of MAAS underlying an OpenStack implementation, but there are a few questions about MAAS that I am not entirely clear on.
Should all hosts be set to network boot at all times or after they have been registered and allocated as a service, should they boot to disk?
After juju bootstrap is executed, I turn on the machine that has been allocated (note WoL isn't working, I suspect it's being blocked on the network), the machine boot's up and then juju status executes correct, agent running and all that good stuff.
If I 'reboot' the machine (testing power failure/problem whatever), juju status comes back but the agent-state is no longer in running state, and so far I have to destroy the environment and restart.
In all cases I have never been able to deploy any services to any of the other nodes. I deploy the service with juju, note which node it was assigned, and then start the system. The system just boots up into a basic node. If I SSH to it I have to enter password, so it's not setting up the ssh key or anything.
This is on Ubuntu 12.04.1 LTS systems and HP GL360G7 hosts. The MAAS management server is running as a VM but all on the same network.
At this point I am not sure if I am doing something wrong or if there is a problem somewhere else. Is the idea that anytime a host is rebooted it should be rebuilt from the ground up, or is something else going on behind the scene to tell it to boot the local image. If the latter, why doesn't the agent start on a system that has been successfully setup before (juju bootstrapped system)?
