Unable to kill-controller
John Meinel
john at arbash-meinel.com
Sun Apr 3 13:45:01 UTC 2016
I'm not sure what happened, but I can't bootstrap, and I can't
kill-controller to try again. I started out with "failing to get
controller-uuid". I then manually killed the machines, but now I get
"unable to get bootstrap information".
Clearly there is *some* bootstrap information, or I would be able to "juju
bootstrap", but apparently there isn't enough information so I can't "juju
kill-controller".
I feel like we're missing something in our implementation. I realize this
is *probably* just version skew, where I was using one of the earlier dev
releases, and then it broke against a new dev release. However, it also
feels like we're missing a more fundamental "kill should make due with the
information it has, and cleanup". "juju destroy-environment --force" was
our old big hammer to wipe the slate clean. But now our attempts at wiping
the state clean all seem to trip up at every turn. The best I've found is
to "rm -rf ~/.local/share/juju" but that, of course, means I lose *all* of
my controllers and my credentials.
It feels like things like "kill-controller" shouldn't be "allowed" to fail.
Throwing exceptions during cleanup is usually something that isn't helpful.
Thoughts?
John
=:->
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/juju-dev/attachments/20160403/94d12ac6/attachment.html>
More information about the Juju-dev
mailing list