On Mon, Feb 1, 2016 at 7:12 AM, Matthew Williams <matthew.willi...@canonical.com> wrote: > Hey folks, > > the lxd provider has been working fine for me - until this morning. I'm > running on master (1cb8f0) > > I've attached output from juju debug-log juju status and lxd list, can > someone help me work out what I should do to resolve this. > > The history leading up the attached logs is: > > 1) Everything was working > 2) A deployed a charm that had a hook install error (I was debugging said > charm, so that was expected). I couldn't ssh to the unit. I got the > following error: > $ juju ssh foobar/0 > ERROR error fetching address for unit "foobar/0": private no address. > > juju debug log showed: > > machine-4: 2016-02-01 11:53:39 ERROR juju.worker runner.go:229 exited > "machiner": setting machine addresses: cannot set machine addresses of > machine 4: state changing too quickly; try again soon > > After doing what debug I can I destroyed the controller. Now whenever I > bootstrap a lxd environment I get the state server and everything else get's > stuck in allocation. Attached is more detailed stuff from the logs. What > should I do to work out what's going on here?
After "juju destroy-model" (or destroy-environment) you should not have any LXC containers left, so "lxc list" should come up empty. If the model was not properly destroyed then you can run "lxc delete juju-..." to manually destroy the container. I'd also check to see if there is an open issue that matches. If not, be sure to open one. Also, from what Casey reported in IRC, it sounds like there might be some relationship with the LXD config file. We haven't really touched the provider in a while so possibly something changed recently which the provider depends on. -eric -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
