Backing up

The best backup software I’ve found for Windows is Acronis True Image.

Unfortunately, the software really sucks.

Let me illustrate the problem. I told Acronis (on their Facebook wall) that I want a simple “set and forget” backup that doesn’t require a lot of configuration and maintenance, and I asked what the latest version of their product can do for me. Their reply:

An incremental backup stores changes to the data against the latest backup. So you need access to previous backups from the same chain to recover data from an incremental backup. To clean up an older backup chain (Full backup + dependent incremental backups) you’ll need to create a new full backup as a base for next incremental backups. Here you can find different clean up schemes available in the custom backup scheme:…/document…/ATI2017/index.html… For your scenario I’d suggest using an option “Keep size of the backup no more than [defined size]” – to limit maximum size of the backup. After creating a new backup version, the program checks whether the total backup size exceeds the specified value. If it’s true, the oldest backup version will be deleted.

This is the year 2016. I shouldn’t have to deal with “full” versus “incremental” backups. I shouldn’t have to worry about what will happen when my backup disk fills up. I should just be able to plug in a backup disk, hit a button, and have the backup software take care of the rest.

I first started using Acronis True Image a few years ago. I ran into problems almost immediately:

  • First, I decided that my backup scheme would be to start with a complete backup of my computer, and then every subsequent backup would be an incremental – backing up only the files that changed. This seemed to make the most sense to me, and it worked fine … until the backup drive filled up. Then Acronis told me it couldn’t proceed, and asked me what to do. What were my choices? Well, only one, really: delete all of the backups and start over! I couldn’t delete any of the incremental backups, or else a file that changed in that backup would be lost to any subsequent backup. And I definitely couldn’t delete the original full backup, because then all the files that never changed would be gone!
  • So I had to start over. This time, I used “backup chains”: it would make a full backup and then five incremental backups after that, and then another full backup and five more incremental backups, and so on. So if it had to free up disk space, it would only have to delete the oldest “chain” of a full backup with all its incrementals, leaving all the other “chains” intact. The problem here is that I had to store multiple full backups on the backup drive – and this wasted a lot of time and space. I place a high value on time and space, but there really wasn’t any other solution.
  • Still, the software kept filling up the drive and being unable to proceed. So (as in the suggestion I received, above) I set “keep size of the backup no more than”‘ the size of the backup hard drive. This worked until I decided to back up two computers to the same backup drive. How could I predict how large I wanted each computer’s backup to be? If I simply set them both to “half the size of the backup drive,” I could be wasting a lot of space if one backup was a lot smaller than the other. Why can’t the backup software simply see when the drive is close to being full, and clean up after itself?

For my Mac computers, macOS comes with backup software called “Time Machine,” and this gets it right. Time Machine takes advantage of a neat little feature called “hard linking”: a single file can exist in multiple places without taking up more disk space. A “hard linked” file keeps track of how many places it exists, and only when that count goes down to zero does it get erased.

So, with Time Machine, the first time you back up your hard drive, it’ll copy everything to the backup. The second time you back up, only your changed files get copied – all of the unchanged files are hard-linked from the first backup, so that they now exist in two places. If you then delete the first backup, those unchanged files now exist in only one place: your second backup.

The benefit of this approach is that you never need to make any “full backups” after the first, and you can delete any backup at any time (for example, if the backup disk is full) without affecting the integrity of other backups. Also, you can restore your entire hard drive from any backup.

I recommended this approach to Acronis a year ago. Sadly, I don’t think they understood my explanation of it.

All the other Windows backup solutions I’ve found have still been tied to this idea of “full” versus “incremental” backups; I haven’t been able to find any that rethink the problem. And, unbelievably, Acronis is selling their backup solution for Mac … for all those people who want to have to micromanage their backups, I guess.

Leave a Reply