How I Back Up the Home Lab — and What I Haven’t Tested Yet
The backups are real. A full disaster restore test is still not.
All I wanted was…
to make sure I could put Home Assistant back if I broke it.
Then it escalated quickly.
The uncomfortable bit
Backups can look reassuring because the archives exist. That is not the same as proving a recovered system will boot, authenticate and resume its integrations.
What is backed up
The design combines recurring virtual-machine backups with separate application and configuration archives on independent local storage. Application databases are captured with methods intended to produce consistent snapshots rather than relying on blind copies of active files. Home Assistant native backups are treated as manual checkpoints, while recurring platform backups provide the routine protection.
What I checked
The recorded audit checked archive readability, compression and selected database integrity. It also compared important live configuration with archived material. Those are useful checks, but they deliberately stop short of a restore-and-boot claim.
The recovery plan starts by identifying the failed layer, preserving surviving evidence and choosing a successful backup that predates the damage. It documents an order for restoring dependent services and calls for isolated validation before a cutover.
What I have not tested yet
No isolated full restore-and-boot exercise or fresh end-to-end failover test is recorded. Some health checks cover only part of the service estate. There is also no independently proven off-site copy or bootstrap-recovery process.
The point
Backups should be described by the evidence they provide. Archive integrity is valuable; it is not an application recovery test. Writing that limitation down is a design strength, not a failure.