Disaster recovery

Now here’s a tale of woe that actually turned out OK, but there were a few tense moments.

My computer (a Mac Pro) has a 1 TB boot disk and was about 75% utilized. I had a backup strategy that had been partially tested, but I’d never had to use it in a recovery mode. I have a 1 TB external disk dedicated to Time Machine, a nifty incremental backup and recovery system that comes with OS X. I excluded most video files because I didn’t want the Time Machine drive getting filled up with large files that I only keep on the boot disk for a few days. Time Machine runs every hour.

I also have a 320 GB drive that gets a bootable system disk copy once a week. Because it is so much smaller than my system disk, I again excluded a number of folders containing video and other temporary stuff. This backup drive was out of date by a few weeks because when I upgraded OS X to Snow Leopard 10.6, I didn’t want to overwrite my 10.5 backup until I was sure the new update was working well. Unfortunately, I forgot to turn the backup process back on.

That described the state of my backups. Why was that important? Being in the cleaning up and throwing out mood (see the previous post), I cleaned up my “office” and got rid of some dust in and on my computer. I have no idea what bad thing I did, but the Mac refused to boot afterwards. After a bit of unhelpful troubleshooting, I booted from the install DVD and ran Disk Utility to check the hard drives. Bad news. The system disk was damaged and unrecoverable (by Disk Utility). Now, I suppose I could have hunted down better disk recovery utilities, but since I had a current Time Machine backup, I just opted for a full restore from Time Machine.

Some 12 or so hours later (it ran overnight), I had a recovered system disk, although a much leaner one. About 500 GB disappeared in the restore. Now, this wasn’t catastrophic—just inconvenient. After a couple of days of checking the recovered system out, I turned Time Machine back on. Somehow, it figured the system was really different and went into a time consuming backup. Twenty or so hours later, it said everything was resynchronized. One nice thing about the Mac Pro is that it has enough CPU power and bandwidth that backups have a negligible impact on normal operation.

I learned a couple of things. First of all, it’s better to have a bootable image backup drive that is the same size as the system drive. I’ve ordered another 1 TB drive for that purpose.

Second, I shouldn’t worry about the Time Machine drive filling up with “temporary” files. It knows how to handle that, and will purge the oldest backup files as necessary to make room for new files that need to be backed up.

Third, and my most serious shortcoming, I have no archive. I have data (like photos, family videos, financial data) that I don’t want to lose. It has value not only to me, but to my heirs. I’ve ordered an external e-sata adapter (and another 1 TB drive) that you can just drop an e-sata drive into. I’ll schedule a monthly image copy to it, and then figure out how to safely and conveniently store that drive outside the house. I’m not exactly sure how that will work—safe and convenient seem opposed to each other.