Comment 9 for bug 1302205

Workaround seems to have just been temporary. I added a second node, then ran juju status, and both non-bootstrap nodes now appear to be down:

ubuntu@ms01a:~$ juju status
environment: manual
machines:
  "0":
    agent-state: started
    agent-version: 1.17.7
    dns-name: 10.0.128.4
    instance-id: 'manual:'
    series: trusty
    hardware: arch=arm64 cpu-cores=1 mem=1895M
  "1":
    agent-state: down
    agent-state-info: (started)
    agent-version: 1.17.7
    dns-name: 10.0.128.7
    instance-id: manual:10.0.128.7
    series: trusty
    hardware: arch=arm64 cpu-cores=1 mem=1895M
  "2":
    agent-state: down
    agent-state-info: (started)
    agent-version: 1.17.7
    dns-name: 10.0.128.8
    instance-id: manual:10.0.128.8
    series: trusty
    hardware: arch=arm64 cpu-cores=1 mem=1895M
services: {}