This week I lost the startup disk of my main server. Although most of my data was in an attached storage, I had several virtual machines in that disk. Fortunately, this Mac Mini was still under warranty, so Apple is replacing the disk for free. But even better, my latest backup in Time Machine is from 6 hours before the crash. And the latest backup in the cloud is from 1 hour before the crash. So I have lost nothing.
In my “How I work” series, I have talked about backups back in 2010. Since then, things have improved a lot. Technology evolved, prices for tech dropped, and I ended up with a multi-site backup scheme that works like this:
- On-site backups: all systems have native OS backup (Time Machine on Macs, File History on Windows, rsync on Linux) that automatically perform backups every 12 hours to a 4TB network storage.
- Off-site backups: all systems are automatically backed up to a remote server every 24 hours, in case something goes wrong at my place (burglary, fire, tornado, alien invasion, whatever). I recommend CrashPlan for off-site backups.
- Cloud sync: all files that are required on-the-fly regardless of the device I am working on are synchronized to the cloud via Dropbox and Google Drive. Please remember to encrypt your most sensitive files before synchronizing them to any of these online services. You may create encrypted sparse bundles on the Mac (via Disk Utility), use True Crypt on Windows, or use the loopback filesystem interface on Linux (you will need cryptoloop and aes loaded in the kernel).
It is very important to have at least two levels of redundancy when it comes to protecting your files from disaster. And one of these levels should always be off-site, for the reasons cited above. This way you will have a lot less headaches when something goes very wrong.