Most backup plans assume you still have a network and a machine that boots.
Cold restores assume the opposite.
When the hardware dies and there’s no internet, can you rebuild from scratch?
If you haven’t practised, the answer is probably no.
Backups Are Only as Good as Your Restore
An image you’ve never tested isn’t a backup. It’s wishful thinking.
Cold restore means starting with nothing but:
- Hardware or replacement hardware
- Offline copies of your data and configs
- A way to get the OS back up without an active net connection
If any one of those is missing, your recovery plan is fiction.
Keep the Essentials Offline
You can’t wget your way back to life if the network is down.
You need:
- OS install media burned to USB or stored on a local drive
- Checksums and keys offline for validation
- Backups on drives that aren’t always plugged in (to avoid ransomware collateral)
Cloud is great when it’s there. It’s useless when it isn’t.
Document the Build
Future you won’t remember the install flags.
Cold restore docs should include:
- Disk partitioning and mount points
- Users and permissions
- Services and where their configs live
- How to restore from offline backup to live data
Make it idiot-proof. Because disaster days make you an idiot.
Rehearse While It’s Calm
The worst time to learn cold restore is when you need it.
Spin up a spare drive or VM and run the playbook now:
- Install OS
- Restore services from offline backup
- Verify everything runs
It’s dull. It’s worth it.
Uptime Vanity vs Survival
You don’t need five-nines uptime. You need a way back from zero.
Cold restores are the difference between inconvenience and total data loss.
Plan for the day you wake up to a dead box and no internet.
Do it now, while the lights are still on.