8. Backup and Recovery¶
Backup types¶
| Type | What it saves | Trade-off |
|---|---|---|
| Full | Everything, every time | Simple to restore, slow and large every run |
| Incremental | Only what changed since the last backup (full or incremental) | Fast and small, but restoring means replaying the whole chain in order |
| Differential | Only what changed since the last full backup | A middle ground — faster than full, simpler to restore than incremental |
The 3-2-1 rule¶
A widely used standard: 3 copies of your data, on 2 different types of storage, with 1 copy stored somewhere offsite. A backup that lives on the same drive as the original isn't really a backup — it fails for exactly the same reasons the original would.
This lesson, already in practice¶
Every time this course's own wiki gets updated, the exact same pattern happens: snapshot CT 251 first, then deploy. That snapshot is a point-in-time copy specifically so a bad change can be undone instantly — the same core idea as a backup, just scoped to one container instead of a whole system. Proxmox's own backup tool, vzdump, does the same thing at a larger scale: full backups of entire VMs/containers, schedulable and restorable.
Try it
Explain why the CT 251 snapshot taken before every wiki update counts as a backup in the sense this lesson means, and what it would take to turn that habit into something that satisfies the full 3-2-1 rule.
Tip
A backup you've never actually tried restoring isn't verified to work — the only way to know a backup is good is to have restored from it at least once.
Play it
Open the Operating Systems Lab and use the Backup tab. Once you've been through all eight lessons, try the Quiz tab for a review drawing from everything in this module.