Leave a comment via GitHub (Giscus). Sign in with GitHub to post.
Introduction
In the world of software engineering, the classic question is “Build vs. Buy.”
When it comes to managing our digital lives—storing terabytes of photos, documents, and videos—we face this exact choice. Should we keep paying monthly fees for services like Google Drive, OneDrive, or iCloud? Or should we build our own solution?
As a Computer Science Master’s student, facing rising subscription costs on a limited budget, relying on SaaS wasn’t sustainable. Naturally, the alternative seemed to be buying an off-the-shelf NAS like Synology or QNAP. However, I chose a different path: repurposing my existing workstation hardware.
By building instead of buying, I avoided the “black box” nature of proprietary OS interfaces. This project isn’t just about saving money; it is a HomeLab designed to translate academic theories—virtualization, container orchestration, and storage management—into practical, production-grade experience. This post breaks down how I engineered a TrueNAS Scale based private cloud with High Reliability, Tiered Storage, and DevOps principles using consumer-grade hardware.
This migration wasn’t a spur-of-the-moment decision. For over a year, I had architected and actively maintained a Proxmox VE environment as my daily driver.
However, my setup was fragmented, consisting of multiple Virtual Machines (VMs) connected via a standard Linux Bridge.
(▲ Legacy Architecture: Direct IP exposure, manual configs, and high VM overhead.)
As illustrated above, despite serving me for a year, this architecture had significant drawbacks:
nginx.conf and patching 4 separate Guest OS kernels.I realized I needed a system that prioritized Storage Stability and Efficiency, leading to the migration to TrueNAS Scale.
To handle mixed workloads—database I/O, large file storage, and concurrent container execution—I moved away from the “just enough” mentality and adopted a Workstation-grade configuration.
Crucially, I implemented a strict Tiered Storage Strategy:
(▲ ZFS Storage Architecture: Logical isolation of Boot, Apps, and Data pools.)
⚡ Power Efficiency Note: Despite utilizing a workstation-class Intel Core i9, the system maintains a remarkably low baseline power consumption of 28–31W. This demonstrates that with proper power management, high-performance hardware can still be highly energy-efficient for 24/7 HomeLab operation.
The core value of this private cloud lies in its ability to securely deliver multiple services to the public internet, effectively breaking free from expensive Big Tech ecosystems.
To secure these services, I transformed the network topology to a “Single Entry Point, Reverse Proxy” architecture.
(▲ Modern Architecture: Secured flow via Cloudflare WAF -> NPM SSL -> Docker Containers.)
I deployed a comprehensive stack to regain full digital sovereignty:
(▲ Production Dashboard: The container fleet running on TrueNAS Scale. Note the minimal resource footprint—CPU idles near 0% even with 4 active services.)
This TrueNAS Scale infrastructure is more than just a storage server; it is a platform demonstrating Full-Stack Infrastructure capabilities.
From hardware assembly and ZFS tuning to container orchestration and network security, every layer required deep system understanding. Through this HomeLab, I have been able to validate DevOps theories in a production-like environment, continuously refining my skills in System Administration and Architecture Design.
Future Roadmap:
| Component | Specification | Architectural Reasoning |
|---|
| CPU | i9-10900 | 10 Cores / 20 Threads. High concurrency headroom for containers & DBs. |
| RAM | 64GB DDR4 | ZFS Optimization. Maximizes RAM caching to minimize HDD read latency. |
| Boot Drive | NVMe SSD | OS Isolation. Separates system logs/updates from App I/O. |
| App Drive | SATA SSD | High IOPS Tier. Dedicated pool for random R/W (Docker/PostgreSQL). |
| Data Drive | 4TB HDD x 2 | Throughput Tier. RAID 1 Mirror prioritization for cold storage safety. |