validate raid levels

It sometimes happens that due to a misconfiguration a partition lands on a single drive instead of a (non-0) RAID. hw-health has access to all the tools necessary to validate the health of both drives and arrays, so it could also check for the proper redundancy of block devices.

As a first step, it should alert on the root partition not being backed by a redundant array (hardware or software).

This check should be enabled by default, return an unknown state if it can't determine the raid level, and allow operators to disable it if needed.

There are then potential additional checks that would be useful, but I'm not sure if it would be appropriate for hw-health to handle them or if it should be up to other charms to do so:

* nova ephemeral images should be on redundant arrays
* ceph OSDs should *not* be on redundant arrays