The backup plan

When we got our first computer, 20 years ago, we managed our backups with floppy disks. We had maybe two or three of them. As for the “backup software” we used Norton Commander. It wasn’t really meant for the task, but we had only a bunch of files.

Those disks were enough for us. When we got a new computer it was a matter of minutes before every important file was accessible to the new machine.

Ten years later my father bought a scanner, for his old photos. The TIFF that were coming out of it were huge. We had to buy a CD recorder to fit them somewhere. And that changed everything. The two or three floppies became the three or four CDs, 0.6GB of data each (it seemed a lot back then). They were lighter than the floppies but there was so much data inside that moving to a new computer was almost a matter of hours.

But we felt safe. When I was looking at the digitalized pictures of my grandfather hunting game, I thought about him a bit like Han Solo in Jabba de Hutt’s mansion: he was preserved in the polycarbonate plastic forever. Safe, ready to be seen by a click of the mouse.

Except that now the original photos are still doing fine, while their digital counterpart has been trashed, because the CDs got unreadable. All CDs are equal, but some CDs are more equal than others. The ones from the music shop were more equal than the ones from the supermarket.

But we kept using them, because we’re resilient folks (or should I say stubborn?). And so at the end of the bachelor, with a tiny iBook G4, I was moving to the Netherlands. In my bag some of my carefully labeled backup CDs. Again? Well, I was also carrying another device, an external hard drive, that I used with iBackup. External? It was mind blowing, at the time. Oh boy, spinning plates, moving parts, you could write and rewrite on the same hard disk forever, and at the same time carry it in your pocked and attach it everywhere.

But I wasn’t relying only on (less equal) CDs and the external HD. People say that data needs to be redundantly stored, which is a way to say that, if something bad happens, the consequences will not redound upon your data. And in the case of an hard drive, your odds can be worse than a cd, because they fail a lot. That means that I started sftp’ing the most important files to my university server. Mathematica files, C++ code, LaTeX, you name it. For me, it was the beginning of cloud storage. Now, I’ve already said what redundant storage means: by all means you can imagine that cloud storage means something with throwing your data out the window.

So what I have now, 2011? Well, I kept the be-redundant-with-your-data and upload-your-data-to-the-cloud styles. Plus I grew the need of having the same data in one or more devices. Therefore my home directory is synced between my Macbook Pro and my iMac SpiderOak. This means that all the data is first encrypted, locally, and then backed up on the web. Then I have a Dropbox folder where I keep the data accessed by mobile app (mostly 1Password and Nebulous Notes). Is that it? No, besides that I have two external ZFS HDs which hold my home folder in two different physical location. I use an rsync script to do that.

My backup rig seems still silly compared to someone like John Gruber. He, and many more, uses SuperDuper, however I haven’t decided on that yet. I can’t boot from a ZFS HD, and all my external HDs are ZFS. But it’s fine for me, and I think I’ll never lose my data, because:

  • if something happens to my HDs, I have my Macs + SpiderOak
  • if something happens to my Macs, I have the HDs + SpiderOak
  • if a cataclism ruins both my house and my offsite HD, I’ll have other things to worry about, but just in case I hope it spared SpiderOak, on the other side of the world.

To wrap it up with the words of Merlin Mann

  • If it’s not automated, it’s not a real backup.
  • If it’s not redundant, it’s not a real backup.
  • If it’s not regularly rotated off-site, it’s not a real backup.