Create a new puppet role the new storage service:
- The use of a multi-instance Cassandra configuration, despite only having a single instance configured.
- The use of a JBOD configuration despite having since learned that we cannot recover from a single disk failure. It's likely this would require repartitioning.
- That admin::groups would match those used for RESTBase (this is almost certainly wrong).
- LVS configuration
- Secrets, keys, certs, etc
- Templating variables for deployment