Implement recovery flow for the case when a slot range is completely lost #131
Description
When a leader and all its followers go out the data on the slot range assigned to them is lost, the operator can not recover from this scenario currently.
The manual procedure would be
1. scale down the operator
2. if the follower of the lost master is still up but empty, delete the pod
3. get the config of a leader, save it, change the name of the pod and the number of redis node
4. apply the config for the new pod
5. on a working leader: redis-cli cluster meet new-pod-ip port
6. repair the missing slots: redis-cli --cluster fix ip-working-leader:port --cluster-fix-with-unreachable-masters
7. rebalance: redis-cli --cluster rebalance ip-working-leader:port
8. scale up the operator, wait until it recreates the remaining follower too
The operation can be done via API or flags on the RDC resource.