2

I've been playing with MAAS and juju (using 14.04 LTS) for the first time and have a single node in my cluster which PXE boots fine via WOL when I do a juju bootstrap.

I'm having an issue where after destroying the juju environment, bootstrapping it again errors out because the juju-db service is still running on that node.

The node reverts from Allocated to Ready within MAAS once the environment is destroyed, but needs a manual reboot before a subsequent bootstrap can take place. There are no entries in maas.log following the destroy.

It doesn't feel right to have to do this - surely MAAS should power down the node down once juju has finished with it? Or at the very least juju should properly clean up after itself during the destroy? Or am I missing something here?

1 Answer 1

2

As discussed on IRC this is a limitation of using WOL with MAAS, which can only power on nodes and not power them off (Sleep on LAN anyone? ;) ). Juju will ask MAAS to power off the node but MAAS can't.

1
  • Thanks - feels like something the MAAS guys should perhaps look at, as ought to be possible to issue a remote shutdown command. In the meantime, perhaps Juju should try and handle the situation where juju-db is already running more gracefully? Jul 10, 2014 at 8:28

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .