I have tried backup setups like the ones discribed in the article. It can certainly be made to work.
However, I have since found a better way:
- See http://rdiff-backup.stanford.edu/
It combines several nice features:
- May use already running services (on my servers, SSH is normally already running, while NFS is most often _not_)
- Easy access to most recent backup data.
- Access to previous backups: The incremental history may extend as far as space on the backup host allows. Unfortunately, the restore syntax is rather difficult to express.
- Super-efficient network use. Means that the backup server may actually be placed far from the backed up hosts == better data security in case of fire, water damage, theft, etc.
Currently, rdiff-backup in reality needs to run as root on the backup server, as well as on the backed up server. However, the software is currently being improved in this regard: Soon, the software doesn't need to be run as root on the backup server.
Written in Python 2.2. Good support-climate (well-working mailing list).
Really worth a try.