Add additional data pools to cephfs

To (properly) support ceph file systems with erasure coding or with partial ssd backing, it would be nice if one could specify multiple data pools for `ceph.automation.ceph_fs` module. Or, a little cleaner maybe, if one could specify an `additional_data_pools` key such that `data` designates the main data pool without relying on yaml sorting.

Start by locating the ceph.automation.ceph_fs module and any related tests or configuration examples. Determine how the current data pool is selected, then verify support for additional data pools while keeping data as the main pool; done means erasure-coded and partially SSD-backed Ceph filesystems can be specified without relying on YAML sorting.

Written by the indexing model from the issue text.