How do I set up backups I can rely on? — answered by Engineer Saeed Baatiah
The practical rule is three copies: the original, one on an external drive, and one in the cloud. The reason is that each medium fails differently — an external drive does not survive ransomware if it stays connected, and the cloud does not survive accidental deletion without version history. And most important of all: **actually test a restore**. An untested backup is not a backup, it is an assumption.
The reasons the three media are each necessary differ, and understanding that stops you settling for one: an external drive restores fast but is encrypted along with the machine if it stays connected during a ransomware infection; the cloud survives that but does not survive a cascading accidental deletion without version history; and the original is exposed to theft, fire and failure. No single medium covers all three.
And testing a restore is the difference between a real backup and a comfortable assumption — and it is the most-skipped step. Once every three months: pick a random file, actually restore it, and open it. The number of situations this test alone uncovers is large — backups that stopped months ago with no notification, backups that run but do not include the folder that matters, backups encrypted with a forgotten password, and corrupt files being backed up corrupt.
And a practical point for small businesses: disconnect the external drive after each backup. A permanently attached drive is convenient but is not a backup in any real sense — it is another disk inside the same blast radius. Automate what you can and keep a notification for backup failure, because a backup that fails silently is more dangerous than none: it gives you confidence with nothing behind it.
- Three copies: original, external drive, cloud — each fails differently
- Test a restore quarterly — actually recover a file and open it
- Disconnect the drive after backing up — always-attached is in the same blast radius
- A failure notification — silent failure gives false confidence