I have what I thought was a fairly big (230GB) extra hard drive fitted to my
PC which I use for backup purposes. The drive is slowly filling up with all
these Incremental backups (58GB free space) left! Can I get rid of any of these
old backups? If so, how do I choose?
No matter what tool you use, a properly configured backup is likely to
simply collect more and more data in the form of backups as time goes on. If
you’re backing up to an external drive eventually it will fill up, no matter
how big it is.
What to do, meaning what to remove, depends on the type of backup you’re
doing, what you expect to be able to use your backups for and what other
storage options you might want to use.
In an earlier article, I discussed the differences between “full” and “incremental” backups. In short, a full backup is a complete backup of absolutely everything on your machine at the time the backup was taken. An incremental backup includes only those things that have changed since the immediately preceding backup was performed.
If you take full backups each time, then while your disk will fill up much quicker, your options are actually fairly simple: just pick which ones you want to delete. Each backup stands on its own, so there’s nothing to worry about. Perhaps keep yesterday’s, last week’s and last month’s, and delete everything else. (More on the logic behind that seemingly random example below.)
If you’re taking incremental backups, as I suspect most do, things get a bit trickier.
An incremental backup is typically useless on its own. It relies on the incremental backups that came before it, all the way to the initial full backup, to create a picture of the data being backed up. Thus, when you want to clean up, you need to make sure to retain all those.
The best way to think of it is that the initial full backup, and all the incrementals that come after it, comprise one backup set. You can delete it, or not, but only as a complete set.
Now if all you’ve been doing is incremental backups since day one, then you can’t really, safely, delete anything. Removing the initial full backup, or any of the incrementals along the way would cause all the backups in the set taken thereafter to be invalid.
That’s why I recommend a blended approach. Take periodic full backups, and daily incrementals. That way each time you take a full backup you “reset the clock”. That full backup stands on its own, and any of the preceding backups, incremental or otherwise, can be safely deleted depending on your own needs and plans.
Many backup programs, including my recommended Acronis Trueimage, can now actually automate this periodic full and daily incremental approach.
And, indeed, that’s my approach: once a month I take a full backup of my machine, and every night an incremental. Over the course of a month that pretty much fills up the external drive I happen to use, so each month I start over.
Once we know what’s safe to delete from a technical point of view – anything prior to the most recent full backup would be a good rule of thumb – we need to think about what it is we need our backups for.
The most common need for a backup is to recover from a system crash. That is, you’re going along, your system dies, and you end up using the most recent backup to restore your system to its prior state. For me, that means if my machine dies, I can restore it to whatever state it was in as of the last incremental backup taken in the middle of the preceding night.
If that’s all you expect to need, or the only case you need to care about, then that rule of thumb not only defines what’s safe to delete – anything prior to the most recent full backup – but the minimum of what you need to keep – the most recent full backup and all subsequent incrementals. That will always allow you to restore as needed to the most recent backup in case of a catastrophe.
But backups can be useful for more than that.
It’s not terribly uncommon to want something older than the most recent. Perhaps you installed a virus or some spyware and didn’t realize it for a few days. Perhaps you want to recover a document that you deleted last month. Perhaps you’d like to restore your machine to the relatively pristine state it was in shortly after you received it.
All of these scenarios, and more, can be accomodated by keeping the appropriate backups. You don’t have to keep them all, just a select few.
Using myself as an example, again, I:
Keep all the monthly full backups for roughly three months.
Keep the quarterly backups (the monthly full backup from January 1st, April 1st, July 1st and October 1st) for at least six months, if not a full year.
Keep the yearly backup (that January 1 full backup) for as long as I can.
If the machine did not come with installation media (or often, even if it did) I keep the first full backup for as long as I have the machine.
If my math is right that’s a storage requirement of about 6-10 times the size of a full backup throughout the year, growing by one each year as I keep a relatively permanent archive.
Do I need to keep that many? Probably not, but disk space to store it on is relatively cheap.
Note that I keep only full backups. As I said above, they stand on their own; I don’t need to keep track of associated incremental backups at all past the current month’s daily backups.
So what you actually need to keep really depends on your own needs and plans. At a minimum, the most recent of course. But if you can envison using backups for more than just restoring to yesterday’s machine or grabing a file you just deleted by accident – you might consider setting up a system that allows you to keep a few of those snapshots as you move forward.