Building my Proxmox cybersecurity homelab
A segmented lab with OPNsense, Splunk, and an Active Directory environment, provisioned with Packer, Terraform, and Ansible.
My lab runs on a single Proxmox host I call lanServer. I built it following The Homelab Almanac by Michael Taggart.
Network layout
Everything routes through an OPNsense VM, with three bridges keeping traffic separated.
flowchart LR
WAN((WAN)) --> OPN[OPNsense router]
OPN --> INFRA[Infrastructure<br/>192.168.99.0/24]
OPN --> ISO[IsolationLab<br/>10.1.99.0/24]
INFRA --> JB[KDE Neon jumpbox]
INFRA --> SPL[Splunk SIEM]
ISO --> KALI[Kali]
ISO --> AD[Active Directory lab]
| Bridge | Subnet | Purpose |
|---|---|---|
| WAN | upstream | Internet uplink |
| Infrastructure | 192.168.99.0/24 | Management and core services |
| IsolationLab | 10.1.99.0/24 | Attack and target machines |
Infrastructure as code
Once the core was up, I automated builds instead of clicking through installers:
- Packer builds VM templates.
- Terraform provisions VMs on Proxmox from those templates.
- Ansible configures them after boot.
- HashiCorp Vault holds secrets.
Things that broke
KDE login loop after pkcon update. The jumpbox stopped letting me log in after an update. I rolled back to a Proxmox snapshot, which is the best argument I know for taking snapshots before every update.
Login loop after chsh to fish. Changing the default shell locked me out of the desktop session. Lesson: test a new login shell in a terminal before making it the default.