At the geek dinner on Thursday we met Mike Culver from Amazon Web Services. He was talking about Amazon’s Simple Storage Service (S3) and Elastic Compute Cloud (EC2). S3 provides a very cheap, reliable storage solution. EC2 provides a way to buy computing time and scale up the amount of power within minutes by using virtual servers on a huge grid.
I’ve looked at S3 before, but not found the time to really play with it. A number of places have recommended it for backups. After Mike’s talk I figured I should get my arse in gear and give it a whirl.
Another Y! blogger, one Mr Zawondy, has A List of Amazon S3 Backup Tools, which seems really complete. I did a bit more searching and found Screencasts online tutorial which uses Jungle disk and S3 Browser. Unfortunately this is one of their paid podcasts, and I didn’t feel like hefting over $35 for a 6 month subscription.
Setting up Jungle Disk was really easy though. My only niggle is the need to keep JungleDiskMonitor.app running. It really needs to be a preference panel, with an optional menu item. I really am getting annoyed with Mac apps which are services not being pref panes. Anyway, Jungle Disk installed perfectly, took my keys and allowed me to store and retrieve data with no problems at all.
Jungle disk does has a backup option. However I really like the excellent iBackup. It’s actually a wrapper for the UNIX tools ditto and rsync. However unlike some apps (Xslimmer comes to mind) it adds significant value. Aside from adding a convenient way to manage the crontab, it also maintains profiles for the applications you have. These self-update from the internet and contain information on where all the Application preferences, etc are stored. This means unlike a lot of back-up programs you don’t have to backup all of your home directory to save information like preferences.
While a lot of people might find doing their entire home directory is ok, I prefer not to. I have a mix of work and personal information on my laptop, and I try really hard to keep them separate. I do not want Yahoo! data under a personal account of mine on an Amazon server. Hell no.
So where is all this leading me? Well Jungle Disk does a great job of wrapping communicating with S3. However, ditto or rysnc don’t like it. They don’t think that it has enough space to receive files.
There is a work-around though. You can backup to some local place and then have Jungle Disk copy that across through its backup. This has an obvious downside, you keep your data twice. If it isn’t big data it may not be an issue. The other down side is that you are running two back-up routines. If Jungle Disk had back up support you could call it from iBackup, but it doesn’t. I have mine set to run it a half hour after iBackup. I’m going to see how that goes, while I look for a better solution.