(Apologies in advance for the wonkiness of this post. Feel free to skip it if backup plans aren’t your cup of tea.)
Backing up your computer: You know you should do it, but you don’t. I only got religion a year or two ago, but now I’m pretty anal about keeping extra copies of everything. (The thought of my digital life — including every email I’ve sent or received since 1993 — disappearing is shattering.) Here’s my system, presented in the hopes that it might be useful to someone.
My computer: I’ve got two internal hard drives, a 160GB boot drive named Huey and a 250GB drive named Earl. Earl holds my MP3s; Huey holds everything else. It’s a Power Mac G5 running 10.4.7.
The most important content on my computer is my email. It’s also the part of my computer most likely to change from day to day, so if I lost a few days worth of material in a disk crash, email would be the most likely to be lost. So:
First Line of Defense: A nightly backup of my email on Huey to Earl. I do it with Email Backup, a dead-cinch-easy way to automate the process. (It looks like it’s just a front end to a cron job, but hey, it’s free and one fewer step.) This way, if Huey goes belly-up one day, I’d lose at most one day’s worth of mail.
But what about the rest of Huey? There’s tons of important stuff on there. Or what if Earl goes? So:
Second Line of Defense: A weekly backup of both Huey and Earl to external Firewire drives. I use the great SuperDuper for this, since it’s perfect for incremental backups. (In other words, you don’t have to recopy the entire hard drive each time you back up — SuperDuper can tell which files have been updated and copy only those.) Huey and Earl each get their own backup drives (250GB Porsche drives named Speedy and Dorian). And the backup drives are bootable — meaning that if Huey suddenly dissolves, Speedy can take its place in less than 30 seconds.
A lot of folks would have SuperDuper do its backup automatically. But that would require me to keep Speedy and Dorian powered on and mounted all the time. I don’t like that idea for a couple reasons: It would put a lot of wear on the drives and it would leave the drives to something external and bad (hackers, lightning, etc.) that could screw up my computer.
Instead, I leave the external drives turned off. But I set up my Yahoo! Calendar to send me an email at midnight every Sunday morning reminding me to backup. All it takes is turning on the drives and a couple clicks in SuperDuper. On average, it takes about two minutes to backup Earl and about 20 to backup Huey. But you can go on merrily working while it works.
So, I’ve got a complete backup of my entire computer that’s at most a week old. But…that backup is sitting in the same room as my computer. What if my apartment catches fire? What if burglars break in and steal my computer — without thoughtfully leaving my backup drives behind?
I hope to eventually get a completely off-site backup in here. Even broadband speed isn’t fast enough for reasonable online remote storage of 500GB of stuff. I’ve thought about backing up my MP3s to offsite DVDs, but damn does that sound daunting. (At 216GB, it’d take 47 DVDs at the moment.) I could buy another set of backup hard drives and keep them at my office. But I’m not there yet.
So, again, I fall back on the backup that’s most important to me:
Third Line of Defense: A monthly backup of my email to Amazon S3. For those unfamiliar, S3 is basically online disk space that Amazon rents out to anyone who wants it. (“It gives any developer access to the same highly scalable, reliable, fast, inexpensive data storage infrastructure that Amazon uses to run its own global network of web sites.”) There are a million potential uses for it, but backup is one of the most obvious.
Accessing AWS directly takes a lot more programming mojo than I have. But a program called Jungle Disk provides a usable front-end (for Mac, Windows, or Linux). On a Mac, at least, you log on to the web space as you would a normal local server, and you copy and paste in the Finder as you would anything else. Jungle Disk handles the actual uploading. (It’s a goofy Mac app — normal interface cues don’t work — but it’s wholly functional.) I get another email from Yahoo! Calendar on the first of every month reminding me to backup.
(A hint for Apple Mail users: Zip your mail directory (~/Library/Mail/) before you upload it. Apple Mail stores each message as a separate file for Spotlight purposes, and all those tiny files really add to the size of your upload and the time of your backup. And if you really want to get everything email-related, you should also backup ~/Library/Mail Downloads/, ~/Library/Application Support/AddressBook, and ~/LIbrary/Preferences/com.apple.mail.plist.)
Now, Jungle Mail is dog slow. It takes almost an hour to upload my complete email archive, which compresses to about 130MB. You can’t do incremental backups, so it wouldn’t work well for huge data stores. But it all happens in the background, so you’re free to do other things. (Like write this post, for instance.) And it works. And if a dirty bomb explodes in my closet tomorrow, my email will be safe on some server in Seattle.
One incredibly small caveat: Amazon S3 isn’t free. They charge for both storage space and upload bandwidth. But the prices are laughably cheap. My bill last month: $0.07. The month before: $0.05. At this rate, my total outlay will reach one American dollar sometime in early 2008.