Skip to content

Start typing to search

Backup and Recovery

How to design backups that actually restore: how many copies, which media, offsite rules, retention, and the restore tests that prove any of it works.

A backup is not a job that finished successfully. A backup is a copy from which you have actually restored. Everything on this page follows from that distinction, because it is where most environments quietly fail: the schedule runs green for two years and the one restore that matters does not work.

Think of a backup system the way you would think of a fire escape in a building. Nobody judges it by how tidy it looks. It is judged once, under load, on the single occasion someone needs it, and the only meaningful test is walking it end to end while nothing is on fire.

What this guide covers

The articles in this cluster take the pieces one at a time: how many copies you need and where they live, what retention actually buys you, how ransomware changed the rules, and how to run a restore test that produces evidence rather than reassurance.

Where to start

If you are building a backup design from nothing, start with the 3-2-1 rule, which governs how many copies exist and where. If you already have backups running and want to know whether they are real, start with restore testing instead.

Backup questions administrators actually ask

Often enough that the first real restore is not the first restore. A quarterly test of a full system, plus a monthly test of a single file or mailbox, catches most of what silently breaks.
No. A snapshot usually lives on the same storage as the original data and depends on it. If the storage fails or is encrypted by ransomware, the snapshot goes with it.
No. Replication copies the current state, including corruption and deletions. It protects against hardware loss, not against mistakes or malware.

Articles in this guide