In most companies we visit, backups „exist”. Somebody set them up once, a green icon blinks, the software reports success. The trouble is that success means only one thing: the data was written. It says nothing about whether it can be read back.
Three ways backups fail
- The wrong thing is copied. The software backs up a folder that was the main workspace two years ago. Today the documents live elsewhere.
- The copy sits next to the original. An external drive permanently plugged into the same computer gets encrypted along with it.
- Nobody has tried a restore. Archives get corrupted, the password is long forgotten and the software that created them is no longer installed anywhere.
A test that takes half an hour
Once a quarter, pick a random file from a few months ago and restore it — onto a different computer than the one it came from. Not „check that it is listed”, but open it and see that it reads.
If that takes longer than half an hour, or requires calling somebody who happens to be on holiday, it is not a backup. It is a lottery.
The 3-2-1 rule, without the mystique
Three copies of the data, on two different media, one off site. That is all. The last one matters most and is skipped most often: it protects against fire, flood, theft and encryption — events that take everything standing in one room.
How much time can you afford to lose
Before choosing a solution, answer two questions: how many hours of work the company could recreate by hand and how many hours it can be down. The first tells you how often to run backups. The second — how long a restore may take. Without them, every backup quotation is guesswork.