Hacker News Reader

Comment by carlmcqueen | Hacker News Reader

I've been trying to reduce my data hoarding. I had a larger back-up hard drive go down and take a lot of data with it. I felt sick knowing I potentially lost years of funny images, videos, music, writings, etc.

What I found is I had back ups of back ups of the things I actually cared about and the things I lost I hadn't truly looked at in a long time. A lot of the music I hadn't listened to in years.

I have found that I can fit on small SSD drives and as a result of the things I do wish I hadn't lost, I'm just keeping up with the practice of keeping multiple back ups.

TL:DR: I have made an effort to keep a smaller data footprint instead of hoarding.

View on HN

If the only instance of a thing is on your "back-up" drive then it isn't a back-up.

Agreed.

I currently have two external drives: an archive drive and a backup drive: both 2TB. (Most things on the backup drive are compressed and deduped, which is why I can fit everything.)

Data that isn't backed up (as in a second copy) doesn't really exist.

Likewise: data that isn't backed up off site isn't really backed up.

Untested backups don't exist either.

There are several horror stories of people going to their offsite tape backups to restore some lost data only to find that their backups didn't work due to a malfunctioning tape drive.

Other stories about encrypted backups where the decryption key is lost (backup data is encrypted with a randomly chosen "session" key, session key is encrypted using a public key, no-one could find the appropriate private key to decrypt the session key.

etc...

I heard about a private school that backed up its fee database (MS Access) to a USB stick every night - in fact what the accountant was doing was backing up the desktop shortcut to the database.

In the event of the failure that revealed the problem, they lost about 7 months worth of data (which was the time since the computer had last been connected to the network, and had synced via Active Directory, or whatever they call it.)

> Data that isn't backed up (as in a second copy) doesn't really exist.

An IMO more realistic statement: Data that is not backed up is implicitly classified as "nice to have" - in many cases a perfectly rational choice, if done consciously.

> Likewise: data that isn't backed up off site isn't really backed up.

You may as well argue that no data is ever "really" backed up because there is always the possibility of simultaneous failure. It's always a choice of how much risk you're willing to accept vs. how much money and effort it would cost to reduce it further.

Yep. 1 copy of data is not enough, because it can be lost. Now, assuming that N copies of data is not enough, it's easy to see that N+1 copies is also not enough, because that extra copy can be lost, thus reducing to N copies, which is not enough.

This proves by induction that no matter how many backups you have, it's not enough.

Reply on news.ycombinator.com