Skip to content

Milestone v1.0: Bare-Metal Hypervisor Foundation & LXC Zoning (Legacy)

Platform Standard: Historical Architectural Baseline
Status: Deprecated (Superseded by Sovereign Cloud v3.0)
Original Timeframe: Milestone v1.0 Architecture
Physical Hardware: Intel Core i5 Mini PC (16GB RAM, 256GB SSD, 1TB HDD)

1. Context & Early Objectives

In the initial iteration of the homelab, the primary objective was establishing a robust virtualization layer directly on bare-metal hardware and experimenting with workload isolation, storage partitioning, and hypervisor resource management.

Initial Resource Allocation

Layer / Zone Physical / Virtual Host Resource Allocation Storage Pool Bound Purpose
Hypervisor Base Proxmox VE 8 (Host OS) 4GB RAM, 4 vCPUs 256GB NVMe (pve-root) Core KVM hypervisor daemons, bridge networking, and local storage pools.
Zone A (Production) LXC Containers & VMs 6GB RAM, 2 vCPUs NVMe (local-lvm) Early lightweight services: DNS (Pi-hole), local Nginx reverse proxy, and identity experiments.
Zone B (Academy) Ephemeral Lab VMs 6GB RAM, 2 vCPUs NVMe (local-lvm) Certification practice sandboxes and "Kubernetes The Hard Way" lab clusters.
Secondary Disk Physical 1TB SATA HDD Raw Mount (/mnt/hdd) HDD-Storage (ext4) Dedicated target for ISO installer images and VM snapshots.

2. Legacy Architecture Topology

v1 Architecture

3. Architectural Deficits & Pain Points

Operating Milestone v1.0 in a production homelab environment revealed critical systemic limitations:

  1. Carrier-Grade NAT (CGNAT) Isolation:
  2. The residential ISP assigned private non-routable IPv4 addresses.
  3. Traditional router port forwarding was technically impossible.

  4. External webhooks from GitHub, Stripe, or alert systems could not reach internal containers without opening insecure holes.

  5. Severe RAM & Resource Fragmentation:

  6. Partitioning a 16GB host across multiple discrete VMs and LXC containers introduced substantial OS memory overhead.

  7. Idle Linux kernels and container runtime daemons consumed ~7.5GB of RAM before application workloads even started.

  8. Storage Contention & Snapshot Failure:

  9. Early unconstrained VM snapshots filled the primary NVMe root partition before automated storage policy routing was implemented, causing hypervisor lockups.

  10. Configuration Drift:

  11. Systems were configured using manual shell commands and localized bash scripts without version control or idempotency guarantees.

4. Key Learnings & Catalyst for Modernization

  • Lesson 1 (Consolidation): Rather than running multiple competing VMs on limited memory, consolidating workloads onto a single production Kubernetes node (k3s-prod) eliminates hypervisor fragmentation.
  • Lesson 2 (Infrastructure as Code): Manual configuration is technical debt. All future provisioning was mandated to use Terraform and Ansible.
  • Lesson 3 (Edge Traversal): Overcoming residential CGNAT requires an active edge traversal mechanism rather than local router tampering.

These pain points directly catalyzed the development of Milestone v2.0 (Hybrid Cloud Relay) and the subsequent Sovereign Cloud v3.0 architecture.