Frank
Here’s my unpopular opinion upfront: Proxmox is a fantastic product that I have no interest in running as the foundation of my homelab VMs, and TrueNAS is excellent storage software that I also don’t use for that purpose. I run Ubuntu. Plain, unglamorous, been-around-forever Ubuntu Server, and I would make the same call again today.
The homelab community treats Proxmox like a rite of passage and TrueNAS like a moral obligation. I respect both. But the assumption that they’re the correct answer for everyone, and that running Ubuntu means you just haven’t evolved yet, is one I’m going to push back on.
This isn’t a comparison post where I tell you which is “better.” Better for what? Better for who? The honest answer is that the right stack depends entirely on what you’re actually trying to do, and for what I’m doing, Ubuntu wins every time. Let me explain the reasoning.
The Skills Problem Nobody Talks About
My day job is Microsoft Exchange Hybrid and Active Directory. Twenty-eight years of IT, most of it in Windows Server environments. PowerShell is my native language. When I come home and sit down at Megatron, I’m not looking to maintain a second full-time expertise. I’m looking to build things and learn things that compound on each other.
Ubuntu Server is Linux without any training wheels, but also without the specialized vocabulary that comes with Proxmox’s web UI or TrueNAS’s particular way of doing things. That matters more than people admit. Every hour I spend learning how Proxmox organizes its storage pools or how TrueNAS wants me to think about datasets is an hour I’m not spending learning Docker, Python, or actually building the applications I care about. The overhead is real.
I’ve gone down enough rabbit holes to know how this works. I can open one tab about ZFS replication and wake up six hours later with forty-seven tabs open, a half-read white paper about btrfs, and a very strong opinion about something I didn’t care about that morning. That’s fun. It’s also a way to never finish anything.
Ubuntu keeps me in a lane where my existing Linux knowledge transfers directly. apt, systemd, cron, ufw, standard SSH config. Nothing exotic. No proprietary management layer between me and the OS. When something breaks, I’m debugging actual Linux, not debugging a platform that happens to sit on top of Linux.
What I Actually Run and Why It Matters
Look at my homelab network and the picture gets clearer. Scooby is my dev server. Passbolt runs a Linux server instance. Several other machines handle specific workloads. HomeBase, my React/TypeScript asset tracker, runs in Docker on Ubuntu. Crumble, my PHP/MySQL recipe manager, same thing. Paperclip, the multi-agent orchestration platform I’m currently configuring, runs on Ubuntu with Docker containers.
The pattern is: Docker on Ubuntu. Every time.
That’s a deliberate architectural choice, not a default. Docker on Ubuntu gives me a surface area I understand completely, with isolation I trust, and a deployment model that’s consistent across every service I run. When I want to spin up a new container for a new project, I’m not navigating a UI or translating concepts between platforms. I’m writing a compose file and running docker compose up -d. Done.
Proxmox is genuinely excellent for running multiple full virtual machines with different operating systems, managing them through a polished web interface, doing live migration, handling snapshots at the hypervisor level. If I had a lab where I was constantly spinning up Windows Server instances, testing different OS configurations, or running a dozen VMs that all needed to stay cleanly isolated from each other, Proxmox would be an obvious choice. But I’m not doing that. I’m running containerized Linux services. The Proxmox VM overhead doesn’t buy me anything in that scenario.
TrueNAS is a different argument entirely. It’s purpose-built for storage. I have four NAS units: Rollo running QNAP, Lamont, Grady, and FredG on Synology. My storage problem is already solved, and it’s solved with dedicated hardware that’s good at exactly that job. Adding TrueNAS to the picture would mean managing another storage paradigm on top of what’s already working. That’s complexity without a corresponding benefit. The QNAP and Synology units handle the media, the backups, the documents. Ubuntu VMs handle the compute. Clean separation.
The Reproducibility Argument
Here’s something that rarely comes up in homelab discussions: what happens when you rebuild?
Drives fail. VMs get corrupted. I’ve had to rebuild services from scratch more than once, and I’ll do it again. On Ubuntu with Docker, rebuilding is mostly just restoring a compose file and a volume backup. The OS configuration is minimal and consistent. I know exactly what packages are installed because I installed them. There’s no platform state to reconstruct, no web UI configuration to remember and replicate.
Proxmox configurations can get complex fast. ZFS pool configurations, VM templates, network bridges, storage backends, all of that lives in Proxmox’s own config layer. If you’ve documented everything carefully, great. Most people haven’t. Rebuilding from scratch means reconstructing a lot of decisions you made eighteen months ago and only half-remember. I’ve been in IT long enough to know that “I’ll document that later” is one of the most expensive lies we tell ourselves.
On Ubuntu, the documentation requirement is lower because the complexity is lower. A shell history file and a handful of compose files gets me most of the way back to operational.
What This Isn’t Saying
I’m not arguing Proxmox is overengineered junk. It’s not. I follow the homelab communities, I’ve read the documentation, I’ve spun it up in a test environment. It’s genuinely impressive software and it makes a lot of sense for a lot of use cases.
I’m also not arguing Ubuntu is superior in some absolute technical sense. For pure storage workloads, TrueNAS handles things Ubuntu simply isn’t designed to do, and it handles them very well.
What I’m arguing is that the homelab community has developed a status hierarchy where Proxmox is the “serious” choice and Ubuntu is what beginners run before they know better. That framing is backwards. Running Ubuntu on your homelab VMs isn’t a beginner move. It can be a fully considered architectural decision made by someone who’s weighed the options and decided that minimizing platform surface area, maximizing transferable Linux skills, and keeping Docker as the primary deployment model is the right answer for their specific situation.
My situation: a 28-year IT veteran building containerized applications in his spare time, on a network of machines he maintains alone, with no interest in managing complexity for its own sake.
For that situation, Ubuntu is the right call. Not because it’s fashionable. Because it fits.
The best homelab stack is the one that gets out of your way and lets you build the thing you actually sat down to build. For me, that’s Ubuntu. Your mileage will vary, and that’s completely fine, but don’t let anyone tell you that running Ubuntu means you haven’t graduated yet.
Sometimes the experienced move is choosing the simpler tool on purpose.
