Implement upgrade of Portworx cluster

[harsh-px]

At a high level, use the below sequence

1. Start a Job on all nodes that starts pulling new PX image. This reduces shared volume downtime which starts on step 3.
2. Find all pods using shared volumes. “Drain the controller” managing these shared volume pods
3. This will start terminating all pods
   • If px volumes are found still attached, bail the upgrade and revert 2
4. Start PX rolling upgrade
   • During upgrade of each node, oci-mon pod on each node will take care of draining apps with regular volumes
5. “Undrain the controller” managing shared volume pods
6. List PX pods that are still not ready and feedback to user to monitor them and reboot node if a specific message appears in PX journal logs

[harsh-px]

Converted to a Project: https://github.com/portworx/talisman/projects/2