lifehacker pointed to this post on how to store an iTunes library on Amazon S3. Basically, this tells you how to use Jungle Disk and Amazon S3 as a network file store for a music library. I really have no need to do this, but I thought it could be a fairly effortless and inexpensive way to backup my laptop.

Both Jungle Disk and Amazon S3 work as advertised. It took me less than 10 minutes to activate S3 and get Jungle Disk working. Once it is set up, Jungle Disk looks like a network server. It handles network file transfers in the background, so writing files to the disk is quite snappy. Very nice, plus the storage and bandwidth costs at Amazon make this a pretty low cost solution ($0.15 per GB per month and $0.20 per GB in bandwidth).

Unfortunately (for me), Jungle Disk does not work very well with rsync, my backup utility of choice. Since Jungle Disk transfers files asynchronously, rsync transfers files even if they already exist at S3. To make matters worse, the files actually get transfered twice. This is due to the way rsync names the files during transfer.

Amazon S3, Jungle Disk, and rsync are all great. But, taken together, mmmm, not so great. There are other backup utilities out there that would probably work just dandy with Jungle Disk. In fact, Jungle Disk includes its own simple backup function. There’s also s3sync, written in Ruby, that mimics rsync functionality.

So, Jungle Disk is worth taking a look at if you need some simple, cheap network storage. But, don’t bother trying rsync. Did I mention it runs on Mac, Windows and linux?

Update 2007-02-19: According to This thread in Jungle Disk’s forum, Jungle Disk and rsync can play nicely if the right options are used with rsync.