[Day 01] - My Journey from Laptop to Lab: Building a Headless Debian Server for DevOps & AI

Source: Dev.to

Introduction

Hey everyone! As an undergraduate who’s always looking to get my hands dirty, I decided it was time to level‑up my old Dell Inspiron 15‑3567. Instead of letting it gather dust running a resource‑heavy desktop OS, I embarked on a mission: transform it into a lean, mean, headless Debian server for my DevOps and agentic‑AI experiments.

This isn’t just about getting a machine running; it’s about deeply understanding the underlying tech. I hope to learn a lot of fundamental concepts throughout the process and share my experience with you. I also appreciate your supportive comments and ideas, because I’m not an expert at all—I learn by searching, asking AI, and trial‑and‑error.

Below is a breakdown of what I’ve been up to, and, more importantly, what I’ve learned (sometimes the hard way!).

The “Before”: My Trusty (but Thirsty) Linux Mint Laptop

My Dell Inspiron, sporting an Intel Core i3‑7100U and 8 GB of RAM, was running Linux Mint 21. While Mint is fantastic for daily use, my inxi -Fxxxrz output revealed its Achilles’ heel for my new ambitions: ≈80 % RAM usage at idle! That’s not going to cut it when I want to run large language models (LLMs) and Docker containers.

Goal: Strip the OS down to the absolute essentials and free up as much RAM as possible.

I Chose Debian – Because Gemini Told Me To…

When I searched for the best OS to learn, set up AI tools, and practice DevOps, Gemini suggested Debian 12 “Bookworm.” A quick Bing search showed that Debian 13 “Trixie” is also available.

I decided to ignore Gemini’s recommendation, download the Debian 13 “Trixie” ISO, and give it a try—because I don’t trust AI blindly.

Features I Expect from Debian

• Minimalism & Performance – A headless Debian server (no graphical desktop) can idle at an incredible 150–200 MB of RAM, leaving ~7.5 GB of my 8 GB RAM for actual AI tasks.
• Industry Standard – Debian is the bedrock for many enterprise servers and the base for Ubuntu, making it invaluable for real‑world DevOps skills.
• Stability – For a dedicated lab, I need something that won’t randomly break, letting me focus on projects, not OS debugging.

Gemini also suggested Arch Linux, noting that “as an undergraduate, nothing teaches you how Linux actually works better than an Arch installation.” I’ll start with the “Debian way” first, then move to Arch later. My long‑term goal is to understand how to build a Debian‑based server from the ground up, since I already use Debian‑derived OSes for my servers.

Navigating the ISO Minefield: “Free” vs. “Non‑Free” Firmware

This was my first major learning curve!

1. SSH into the laptop over the local network – the first time I used ssh instead of AnyDesk.
2. Download the mini.iso with wget.
3. Inspect block devices with lsblk (List Block Devices).

Once I identified the correct block device, I flashed the downloaded OS onto it using dd (a powerful tool that overwrites everything, so always double‑check your target drive!).

I thought I was golden, but then the installer hit me with this:

The Revelation: My Wi‑Fi and Ethernet Need “Non‑Free” Firmware!

The messages ath10k/QCA9377 and later rtl_nic weren’t random errors—they were the installer politely (but firmly!) telling me that my Qualcomm Atheros Wi‑Fi and Realtek Ethernet chips required proprietary binary blobs to function.

These blobs are “non‑free” not because you have to pay for them, but because their source code isn’t openly available, which conflicts with Debian’s strict “free software” philosophy in its default installers.

The Solution

I scrapped the mini.iso and downloaded the official Debian 13 Netinst ISO that includes non‑free firmware. This version bundles the necessary closed‑source blobs, allowing the hardware to actually turn on.

Next, I created a bootable device with the new ISO using the same dd command. Stay tuned for Day 02 to see what happened next!

End of Day 01 notes.