They had an eight-month backup. On the day it mattered, nothing was restored. This is the most costly lesson we see small businesses learn: a backup that runs doesn’t mean a backup that works, and the difference between the two only becomes apparent on the day you need your data.
Background: A construction company with a server that stored projects, cost estimates, and contracts. They called us after a colleague reported that “the backup was working”—it showed up as green every night, with no errors.
I asked for just one thing: a restore test. That is, to take a backup and bring it back to life, just like in a real emergency. It didn’t work. The backup was running, but for months it had been saving the wrong folder, and half the projects were completely missing from the backups. “Green” just meant that the process had started, not that the data was recoverable.
What is worth noting about a case like this:
- An untested backup isn't a backup—it's just a guess;
- An “error-free” report only confirms that the process started, not what was actually saved;
- The only way to find out if the backup works is to restore it before you’re forced to do so by a crisis;
- The restore test must have a designated person in charge and a set frequency; otherwise, it will never be done.
That’s why, for our clients, the restore test is included in maintenance—it’s not just a matter of checking that the system is “up and running.” It’s a simple check that takes very little time, and one that no one wants to perform for the first time only on the day of a disaster.
If you want to know whether your company's backup can actually be restored, we can test it together: https://xdn-cs.ro
This article was generated with AI assistance.