From-Scratch Rebuild Process

[homelab][disasterrecovery]

In the event of a disaster, rebuilding the entire homelab once homelab-as-code is complete should look like this:

  • Install Proxmox on desired hosts
  • Set up desired VM storage (ha-zfs-storage, ceph etc.)
  • Add hosts to cluster & set up HA
  • Add new hostname & IPs to ansible inventory group(s)
  • Add a new host_vars file for each new host (delete old ones as necessary)
  • Verify proper storage is referenced in proxmox setup playbook
  • Generate new API key and add to secrets.yaml
  • Manually run:
    • proxmox setup playbook
    • terraform to provision VMs
    • talos bootstrap script
    • get new kubeconfig using talosctl
    • use kubectl to apply manifests for infrastructure resources:
      • tailscale
      • traefik
      • longhorn
      • authentik
      • gitea
    • restore authentik/gitea from backups
    • install & bootstrap flux against gitea repo
  • The rest should happen automagically