Comment 2 for bug 1886429

One way to reproduce during B&R of AIO-DX:
deploy AIO-DX, remove armada label on node controller-0. delete armada pod, armada pod moves to controller-1.
take the backup. do the restore up to including controller-0 unlock. Then easiest way to reproduce is to constantly apply cert-manager which will constantly fail with `Failed to copy /etc/build.info`

> /usr/lib64/python2.7/site-packages/sysinv/conductor/kube_app.py(3072)_start_armada_service()
-> if pod and pod.status.phase == 'Running':

print pods[0]
            'phase': 'Running',
            'pod_ip': '172.16.166.134',

armada armada-api-6d7cdc4555-nk4xh 2/2 Terminating 0 165m 172.16.166.134 controller-1 <none> <none>
armada armada-api-74cffccf8c-d48jd 2/2 Running 2 115m 172.16.192.102 controller-0 <none> <none>

line 3063 pod = pods[0] well..the logic only takes one pod into account.
But why isn't kubernetes returning the phase `Terminating`? seems to me a k8s bug