Disaster Recovery

Hard drives fail, cloud providers have outages, and ransomware encrypts whatever it can reach. The 3-2-1 rule is just a way to make sure no single one of those can take you down at once.

What this looks like in practice

• Copy 1: Your working files (local machine)
• Copy 2: External drive or NAS (different device)
• Copy 3: Cloud backup or offsite drive (different location)

If it's text, it belongs in Git. Code, configuration, documentation, scripts — all of it. Version control gives you:

• Complete history of every change
• Ability to roll back mistakes
• Built-in offsite backup (GitHub, GitLab, etc.)
• A clear record when more than one person is involved

Don't limit this to code. Infrastructure as code, dotfiles, even notes — if it matters, version it.

If you've never restored from a backup, you don't yet know whether it works.

Schedule regular restore drills — quarterly at a minimum. Actually restore the files, actually spin up from your backups, and time how long it takes. It's a lot easier to find the gaps during a drill than during a real outage.

Restore drill checklist

• Can you actually access your backups?
• Do you have the passwords/keys needed?
• How long does a full restore take?
• Is the restored data complete and usable?
• Who else knows how to do this if you're unavailable?