Instances can get into a 'highlander' state - no obvious way to kill them

[James Troup]

With vanilla Ubuntu 12.04 (including precise-proposed) we often see
instances get into a 'highlander' state. They're in an 'error' state,
like so:

RESERVATION r-n1d0t747 c519923c921a404c96ebc8210a4ec67a juju-canonistack2, juju-canonistack2-10
INSTANCE i-000000bb ami-000000bf server-187 server-187 error None (c519923c921a404c96ebc8210a4ec67a, alce) 0 m1.small 2012-07-02T02:12:56.000Z nova monitoring-disabled instance-store

This is particularly problematic when combined with things like LP
#1092108. There doesn't seem to be any reasonable way to kill the
instances (ec2-terminate-instances as either the owner or an admin
user has no effect).

(I realise there's not a lot of detail here; I'm happy to any more
 that's asked for, but the focus of the bug (from my POV) is the
 unkillable nature of these instances, not how/that they're in an
 error state.)

[Chuck Short (zulcss)]

Elmo,

So something like nova delete --force as an admin?

chuck

[Launchpad Janitor (janitor)]

[Expired for nova (Ubuntu) because there has been no activity for 60 days.]