Emergency plans

What to do when bad stuff happens to the homeLab

DoS Attack

Ban the offending IP via Fail2Ban on the VPS (already running on the public-facing proxy - see VPS Proxy).

DDoS Attack

Unplug the router, research current mitigation options, implement whatever’s needed, then plug it back in. Since the VPS absorbs all public traffic by design (see Double NAT), the home network itself shouldn’t be directly reachable regardless.

VPS Compromised

The VPS is treated as disposable and stateless on purpose (see the “Security” section of Double NAT) - it holds no lab data, so the fix is to destroy and rebuild it rather than try to clean it. Restore from the DigitalOcean snapshot (see VPS Proxy), then rotate the WireGuard keypair and any secrets it held (Authentik secret key, Caddy config) before reconnecting it to the tunnel.

VM or Host Rooted/Hacked

Unplug the affected host or VM’s network cable first. VLANs are set up to block cross-talk between segments (see Switch), so this should contain the problem to that one host/VM. If it’s unclear which host or VM is affected, unplug the router as a last resort to cut everything off at once.

Drive Fails

Look up how to fix it and fix it. There are currently no backups for Proxmox VMs/hosts (see Host Nodes) or for the Cisco switch config (see Switch), so a failure on any of those means rebuilding from scratch, not restoring. The router’s boot drive is a known, especially old single point of failure (see Router).

WireGuard Tunnel Down

A cron job on the VPS already pages my phone when this happens (see VPS Proxy). The tunnel is known to sometimes fail to recover on its own after a power outage, for a still-unknown reason (see Router) - if it doesn’t come back within a few minutes, log into the router and restart WireGuard manually.

Power or Internet Is Down

Nothing - no UPS or generator. Everything should come back up on its own once power/internet is restored, aside from possibly needing the manual WireGuard restart above.