Proxmox VE

Homelab Build – Learning Networking, Self-Hosting, and Linux

Introduction: Why I Built a Homelab

My journey into the world of homelabbing started with a desire to expand my technical skills beyond theoretical knowledge. I wanted hands-on experience in networking (VLANS, subnets, network segmentation), self-hosting services (Pi-hole, Nextcloud, Jellyfin, Home Assistant, Docker Swarm), and virtual machine management for development environments. More importantly, I wanted to understand how containerisation and orchestration tools like Docker, Podman, and Kubernetes are utilised in enterprise environments.

After successfully completing my CompTIA Linux+ certification, I was motivated to explore Linux further, particularly in the areas of container orchestration. Building a homelab was not just a means of experimentation but a strategic decision to simulate real-world scenarios and build practical skills.

In this article, I’ll detail the hardware and software selections I made, explain my rationale behind each decision, and provide insights into the learning experiences along the way.

Objectives of My Homelab:

I aimed to:

• Learn Networking: VLANs, subnets, firewalling, and secure network segmentation.
• Practice Virtualisation: Running VMs and containers efficiently using Proxmox VE.
• Explore Container Orchestration: Docker, Podman, Kubernetes (enterprise-focused solutions).
• Host Personal Services: Self-host solutions such as Pi-hole, Nextcloud, Jellyfin, Home Assistant.
• Test and Deploy Small AI/ML Workloads: GPU acceleration and virtualisation for experimentation with AI.

Networking and Firewall Setup

A crucial component of my homelab was the network infrastructure. I needed to create a segmented, secure network to simulate enterprise-level setups. For this, I used pfSense Community Edition as my firewall/router. Setting up pfSense and segmenting my home network into VLANs was an invaluable learning experience that provided hands-on knowledge in:

• VLAN configuration
• Subnetting
• Network segmentation for security
• DHCP and DNS management

To host pfSense, I purchased a used Dell Optiplex 9020 SFF Desktop with the following specs:

• Intel Core i5-4590
• 250GB SSD
• 16GB RAM

I chose a dedicated machine instead of virtualising pfSense on Proxmox to avoid losing internet access during Proxmox reboots. This separation provided flexibility and reliability.

Server Build: Hardware Overview

My goal was to build a reliable, server-grade machine without overspending. While many homelabbers use spare desktops or laptops, I built a compact but powerful server from scratch.

Core Components:

• Motherboard: ASRock Rack X570D4U (ECC support, AM4 socket)
• CPU: AMD Ryzen 7 5700X (8 cores, 65W TDP)
• RAM: 64GB Micron DDR4 3200MHz ECC UDIMM (2x32GB)
• PSU: Corsair RM Series 650w 80+ Gold Fully Modular
• Case: Fractal Pop Mini Silent (Black)
• CPU Cooler: Noctua NH-D15 chromax.black (high-performance, fits perfectly in the case)

Storage Setup

OS Storage:

• Drives: 2x 250GB SSDs (Samsung 870 EVO + Crucial MX500)
• Configuration: ZFS RAID 1 (Mirror) for redundancy

VM and Container Storage:

• Drives: 2x 1TB NVMe SSDs (Kingston FURY Renegade + Corsair MP600 PRO XT)
• Configuration: ZFS Mirror for redundancy and performance

Note: RAID is not a backup solution. I plan to implement a dedicated Proxmox Backup Server or NAS in the future for actual data backups.

Networking Infrastructure

I wanted 10 Gb networking to ensure high-speed data transfer between my Proxmox server and workstation. I balanced performance with cost since enterprise 10 Gb switches can be expensive.

Networking Hardware:

• NIC: Dual LAN SFP+ PCI-e Card (Intel 82599/X520-DA2 Controller)
• Switch: QNAP QSW-M408-4C 10GbE Layer 2 Managed Switch (supports VLANs, LAGG, user-friendly UI)

This setup allowed me to create VLANs and Link Aggregation Groups (LAGG) while maintaining high-speed connectivity. I’ll detail the switch configuration and VLAN setup in future write-ups.

GPU for AI and Media Workloads

Interested in experimenting with small-scale AI workloads and GPU virtualisation (IOMMU passthrough), I added:

Intel ARC PRO A40 GPU (6GB GDDR6) – A cost-effective, entry-level workstation GPU suitable for initial AI exploration, virtualisation testing, and media transcoding. If needed, it can be upgraded easily in the future for heavier workloads.

GPU:

• Intel ARC PRO A40 6GB GDDR6
• Supports Ray Tracing and AI workloads
• Can be passed to VMs or Containers via IOMMU passthrough in Proxmox

Virtualisation Platform: Proxmox VE

For the hypervisor, I chose Proxmox VE, a powerful, open-source virtualisation platform based on Debian 12 Bookworm.

Key Features Used:

• Support for VMs and LXC Containers
• ZFS for robust storage management
• IOMMU passthrough for GPU and PCIe devices
• Snapshots and backup tools

Proxmox has been excellent for managing multiple environments and learning enterprise-grade virtualisation.

Lessons Learned:

Building and configuring this homelab taught me significantly more than I initially expected. Some highlights include:

• Proxmox VE Management: Setting up ZFS storage pools, optimising VM and container performance, and GPU passthrough using IOMMU.
• Advanced Networking: VLAN segmentation, subnetting, firewall rules, DHCP/DNS management, and routing via pfSense.
• Container Orchestration & Linux: Gained practical experience using Docker and Podman, including container networking and storage management.
• Hardware Selection & Assembly: Learned to select compatible, reliable components resembling enterprise-grade solutions, balancing cost and performance.

Future Plans and Learning Goals

• Set up a Proxmox Backup Server or Truenas for proper backups
• Configure reverse proxy with SSL (e.g., Nginx Proxy Manager or Traefik)
• Explore Kubernetes using K3s or MicroK8s for lightweight deployments
• Deploy Grafana + Prometheus for monitoring and visualisation
• Experiment with larger AI workloads and possibly upgrade GPU

Documentation & Further Learning:

I plan to publish additional articles detailing the configuration steps, best practices, and lessons learned from each aspect of my homelab:

• VLAN and pfSense configuration guides.
• Proxmox setup and optimisation tutorials.
• Kubernetes deployment and container orchestration strategies.
• GPU virtualisation and AI workload walkthroughs.

Conclusion

Building this homelab has been one of the most rewarding learning experiences. From networking to containerisation and virtualisation, I now have a deeper understanding of enterprise technologies and practical skills that I use daily. I plan to write detailed articles about each component of this homelab to help others who want to embark on a similar journey.