Ugarit: initial beta (by )

I'm pleased to announce the release of the first beta release of Ugarit, a backup/archival system based around content-addressed storage, written in Chicken Scheme.

This initial release supports archives stored in the filesystem, including on remote servers via NFS and other such protocols. Future versions plan support for storage of archives in S3 or on remote hosts via SFTP/SSH, and a pluggable storage backend system allows for many other forms of archive to be created.

Ugarit provides efficient snapshots and restores, without requiring intelligence of its storage. Anything that works roughly like a filesystem can be used as a Ugarit backend, and it is designed to minimise the size of data sent to the archive, in order to reduce transfer and storage costs on services like S3, and snapshot time.

I've tested it on various test filesystems, ranging from a contrived example with all sorts of funny things like FIFOs and devices in, up to 500MB of /usr/pkgsrc and >2GB of /usr. I'm going to see if I can borrow some big hardware at work to test it on some many-hundreds-of-gigabytes filesystems as well, to see if I can find any scaling issues, and I'm currently putting it into place as my personal backup system. However, this is still beta software, so please be careful and test your backups!

For details and installation instructions, see the Ugarit project page.

Future developments planned include:

  • File modification time caching, reducing the time taken to identify changed files to snapshot.
  • Encrypted archives.
  • Replicated archives, supporting both fault-tolerance over multiple archives and local caching, where extractions are serviced from a local archive, but if the local archive is lost (even just partially), a remote archive can provide the missing data.
  • More storage backends
  • FUSE support, so you can browse your archive as a read-only filesystem

3 Comments

  • By Faré, Tue 27th Jan 2009 @ 2:51 pm

    Tough question: hardlinks... (and note how inodes numbers may change from run to run).

  • By Improbulus, Sat 31st Jan 2009 @ 5:22 pm

    Congrats on the release! I imagine there will not be a Windows version of Ugarit, am I right or am I right..??

  • By Maurice Snell, Mon 11th Aug 2014 @ 12:48 pm

    Hi Alaric. I like the sound of Ugarit, but there's one thing I don't understand: what happens when two or more different blocks happen to return the same hash value? How do you avoid corrupting the data by thinking that both blocks have the same content when their hashes match?

    Maurice

Other Links to this Post

RSS feed for comments on this post.

Leave a comment

WordPress Themes

Creative Commons Attribution-NonCommercial-ShareAlike 2.0 UK: England & Wales
Creative Commons Attribution-NonCommercial-ShareAlike 2.0 UK: England & Wales