CSI StorageClass for management cluster

[BobyMCbobs]

There are several things that need to persist across Pod restarts

• Coder Postgres database
• PowerDNS MariaDB database

There are a few options for the provider, such as Rook+Ceph, Mayastor, local-path-provisioner, topolvm, Longhorn.

In my opinion to run Rook+Ceph, you need a dedicated team (I can't solely maintain it)

Needs:

• ReadWriteOnce (nothing fancy)
• nothing tied to the node
• something supported

[BobyMCbobs]

Longhorn appears to work well.
It supports replication, volume snapshotting, backups and disaster recovery.

[BobyMCbobs]

Longhorn is successful.
Added in 040b277 and 07eb773