Lack of Openstack resources leaves undead service

[Stuart Bishop]

My test suite spawned too many instances, causing me to end up with 10 machines. The 10th failed to spin up, and I'm left with a service I cannot destroy and a machine I cannot terminate.

$ juju status
machines:
  "0":
    agent-state: started
    agent-version: 1.11.0
    dns-name: 10.55.63.139
    instance-id: e86d2de7-d71f-482d-9f85-6139aeb0275f
    series: precise
  "1":
    agent-state: started
    agent-version: 1.11.0
    dns-name: 10.55.63.126
    instance-id: ae1cdf02-e93e-4da1-b541-f3f7a4b9d564
    series: precise

   [... trim ...]

  "10":
    agent-state-info: '(error: cannot run instance: failed to run a server with nova.RunServerOpts{Name:"juju-lcy2-machine-10",
      FlavorId:"1", ImageId:"8f6560c9-3810-47db-b8ba-b78cf5d4f69d", UserData:

   [... trimmed huge hex dump included in status ...]

, SecurityGroupNames:[]nova.SecurityGroupName{nova.SecurityGroupName{Name:"juju-lcy2"},
      nova.SecurityGroupName{Name:"juju-lcy2-10"}}}

      caused by: Resource limit exeeded at URL https://nova-lcy02.canonistack.canonical.com/v2/b448eaffea0f41588b365395216bdd7b/servers)'
    instance-id: pending
    series: precise
services:
  postgresql:
    charm: local:precise/postgresql-5
    exposed: false
    life: dying
    units:
      postgresql/0:
        agent-state: pending
        machine: "10"