Overhaul existing backup automation

This ensure we centrally collect all logs and applies the same mount/umount and locking technique to all existing backups.

EDIT: I decided that this whole mount/umount/flock business is much too fragile and will reboot this into an OpenStack-based strategy instead: Just rsync from/to the (continuosly) attached volume and use a Python script to keep snapshots of that volume using the OpenStack API. This way, the snapshots are still "out of band", but with hopefully less complexity inside the systems.