Cassandra: data directory hardcoded into the globs

[dmitry-potepalov]

Presently the patterns we use to pick the files that go into the backup for Cassandra do not start at the keyspace level; they include the data directory data:

SNAPSHOT_GLOB = f"data/*/*/snapshots/{SNAPSHOT_NAME}"
BACKUP_GLOB = "data/*/*/backups/"

Therefore if the data dir is configured to be anything else than literally data in cassandra.yaml, we don't back up anything. E.g. setting up a demo with ccm required me to patch the thing, because ccm puts the data into data0...dataN dirs depending on the count in configuration.

If we want anyone else besides us to use Astacus, the data dir needs to become configurable. Unfortunately, it's non-trivial, because we need to keep compatibility with the existing backups. Shouldn't be very complicated, though.