Recovery when all pods are down

Covers 2 types of cluster down scenarios

1. Cluster is running and sudden disaster leads to all pods down, but operator is still running.
2. Pods and operator is down.  Operator is started and sees pods are down and restarts them.

• only pods with persistent volumes will be recovered *

Behavior is undefined when all pods and operator goes down and all start at same time.  Or at least its the same as rebooting couchbase on several physical machines/vms.