Here I go with more about backups again! I recently posted about using JungleDisk to backup my documents to Amazon's S3 storage. When I posted that I had still been running the original big old backup to the cloud, plus an annoying update where I moved around a few gigabytes of stuff and so my system hadn't been sleeping for a week or so. I assumed that was because it was grinding away at the backup. Since then I've discovered that TinyGod won't sleep automatically when JungleDisk is running. I can put it to sleep manually (with the sleep option on the System menu) but that's not what I want. I have a support ticket open with the JungleDisk people, we'll see what happens. In all fairness the current version of the JungleDisk software is listed as beta, so maybe this problem will go away soon. In the meantime, after I reboot TinyGod I run this command to unload JungleDisk: sudo launchctl unload /Library/LaunchDaemons/com.jungledisk.service.plist Obviously the usual caveats apply. If you don't know what that means, probably don't run it. If you're not comfortable mucking around at the command line, don't run this. It may not fix the problem. Etc. If you followed my Twitter feed on Friday you know it was a little more complex than this. There are two other issues that complicated my diagnosis. First off is that running this shutdown doesn't seem to always fix the sleep issue. I think the problem is that if I run the JungleDisk Monitor application then sleep is just screwed up until I reboot, even if I later unload the daemon. The second thing is that I had installed an application called FuzzyClock while the big backup was running. FuzzyClock also prevents sleep. I was somewhat amused by getting the "fuzzy times" in my menu bar, but not enough to put up with this side effect so FuzzyClock got the boot. I probably would have normally noticed the sleep-blocking the first day I ran the application, but it snuck under the radar while the big backup ran. At any rate, if you care about having a system sleep automatically when idle I'd recommend staying away from FuzzyClock and JungleDisk at the moment. Personally I'm willing to manually shut-down JungleDisk like this and make it a process I run every week or so as opposed to having no offsite backup, but your mileage could certainly vary.

Back in 2007 I did a whole set of posts about backups, laying out my "clone your drives" strategy and listing what software I used to make the backups, and going through several different choices for Windows. About a year and a half later that's still what I do for the most part - every machine has an external drive and I weekly clone the entire boot drive to the external backup. I've added Time Machine more for the "oh shit, I'd like to get yesterday's version of a particular file" coverage, and since a lot of things are heavily sync'ed between my laptop and my desktop that means I really have something like six copies of my main files. Windows has been mainly ghetto-ized into a virtual machine so it's a single file that gets cloned and backed up that way and to hell with trying to run a Windows backup. (I still have Boot Camp and the Bart PE/DriveImage XML solution for that install, but I don't think I've used Boot Camp in 2009 yet. Which means the first day I want it I'll have to let Microsoft play patch-and-reboot for quite a while. Be nice if they could get their shit together about updates.) There was one remaining glaring hole in the entire system: if a tragedy destroys all of the hard drives in my office and living room then all six copies of whatever can be destroyed at once. D'oh! I finally got around to solving that last week, and my suggestion is Jungle Disk. Jungle Disk is a cross-platform piece of software that backs up data to Amazon S3 storage. When I installed the OS X version it walked me through adding S3 to my existing Amazon account and then configuring the backup process. Jungle Disk can run backups automatically, throttle bandwidth usage, provide encryption, will archive multiple versions of files, and costs $20 for a license on as many computers as you want. It took around a week being throttled during the day to get my Documents folder and my Aperture photo library up to the cloud - that's roughly 26 Gigabytes of data. Now that it's all up there it looks like it will be able to fairly easily handle updates and changes as we go. I just finished installing it on Horton to backup my Subversion archive, MySQL databases, and all of the web pages. S3 charges $0.15 per Gb of storage per month and $0.10 Gb of transfer in (plus some other twiddly charges for overhead requests). My projected bill for S3 as of March 1st is $4.98. So for $25 ($20 for JungleDisk and $5 for Amazon S3) I've backed up every document and photo I have to the cloud. If every hard drive I own is destroyed at the same time, once the insurance claim clears and I buy a new Mac I can pull down everything I care about. Sure it would take a while, but none of it's GONE. That's worth $5/month no doubt. If you care about your data, the Amazon S3/Jungle Disk combo seems pretty solid to me. I'm aware of some free services that do this, but there's no business model there that makes sense to me, and if you're using a backup service you don't really trust then what's the point? I don't think Amazon is going anywhere and since they are charging money (even if it is a piddly amount of money) then if I have a problem I can be aggressive about complaining and yelling. Anyway, so far I'm pretty happy with how it all works and I can sleep easy knowing that even if a meteor crushes my house and demagnetizes the rubble I'm still able to restore my critical data.