Project Phoenyx – Part 1: My Journey to the Perfect 2-Server Infrastructure

Category: Homelab / Servers / Tutorials | Read Time: approx. 8-10 minutes

Introduction: Why Are We Doing This?

Let's be honest: We all started small. Maybe with a Raspberry Pi gathering dust in a corner, or a cheap VPS where we wildly installed everything the internet had to offer. And we all know the result: Eventually, you lose track. Was port 8080 for the web server or for that test project from last week? And – far more importantly – what actually happens if this one machine dies tomorrow?

I was at that exact point. I wanted to move away from "tinkering chaos" towards a serious, stable environment. I wanted to host my own cloud, be independent of big providers, but still sleep soundly at night.

Welcome to Project Phoenyx.

In this new blog series, I’m taking you on a journey. We aren't just setting up a server. We are building an infrastructure. We will set up two servers to function as a team: One stands at the front, the other holds the fort in the back. Does that sound like Enterprise IT? Maybe. But the beauty is: We’re doing it in a way that beginners can understand and replicate.

The Strategy: Why Two Servers Are Better Than One

You might be asking yourself: "Why the effort? Wouldn't one big server be enough?"
Technically: Yes. But in terms of security and organization, the "all eggs in one basket" method is risky.

My concept for this guide is based on a strict separation of duties. We divide our tasks into two areas: "Production" and "Management."

Think of it like a restaurant.

No one would print the safe combination directly on the menu. That is exactly what we are implementing with our two servers.

Our Candidates

We are christening our two servers Server01 and Server02. Both will run on Ubuntu 24.04 LTS – a system known for its stability that will support us for years to come.

The Worker: Server02 (The Frontline)

Let's start with the server in the spotlight. Server02 is our workhorse. This server is directly connected to the internet and provides the services we use daily. It needs to be fast, it needs to be accessible, and it needs to be configured to handle requests efficiently.

Here is what we will install on this server – and why:

1. Traefik (The Bouncer)

When requests come from the internet (someone visits your website), we need someone to direct traffic. Traefik is our "Reverse Proxy." It’s modern, extremely fast, and loves Docker. The best part about Traefik: It automatically handles HTTPS certificates (the green lock in the browser). We never have to worry about expiring certificates manually again. It automatically routes requests like cloud.my-domain.com or mail.my-domain.com to the correct container.

2. Mailcow (The Post Office)

Running your own mail server is considered the supreme discipline of hosting. Many advise against it. However, with Mailcow, it becomes manageable. It is a complete suite that brings everything: virus protection, webmail, calendar, and contacts. Why host it yourself? Because your emails are nobody's business but yours. On Server02, Mailcow will be our communication hub.

3. Vaultwarden (The Safe)

Saving passwords in the browser is convenient but risky. Using services like LastPass in the cloud requires trust. Vaultwarden is a lightweight version of the well-known "Bitwarden" server. With it, you host your own password manager. Your passwords never leave your control. Since this service is critical, it runs on the frontline but will be (as we’ll see later) massively secured.

4. Your Website & Web Projects

Whether WordPress, a static blog, or a small app: This is their home. Thanks to Traefik, we can run as many websites in parallel as we want without them getting in each other's way.

5. Monitoring Agent

So that we know how Server02 is doing, it gets a little "spy." This agent monitors CPU, RAM, and disk space and reports this data live to our second server.

The Manager: Server01 (The Fortress)

Now we come to the heart of my security strategy. Server01 is the server operating in the background. It is less "public" than its brother. Its job is protection, monitoring, and administration. If Server02 gets hacked or burns down tomorrow, Server01 is our life insurance.

1. Monitoring HQ (The Cockpit)

All data converges here. We will install a dashboard (e.g., Grafana or Checkmk) that shows us colorful graphs and warning alerts. Is Server02 overloaded? Is the backup drive filling up? Server01 will tell us before it’s too late.

2. Backup (The Safety Net)

This is the most important point of the entire series. A server without a backup is just data loss waiting to happen.

We will set up an automated process that pulls all data from Server02 (the mails, the passwords, the websites) to Server01 regularly.

Should Server02 fail, we have a 1:1 copy of the data ready on Server01. We are physically separating "live data" from "backup data."

3. VPN (The Secret Passage)

I don't want to expose every administrative port to the internet. Databases or the monitoring dashboard have no business being on the public net. That’s why we are setting up a VPN server on Server01 (e.g., WireGuard). If I want to administer something, I connect via VPN. To the internet, these doors are invisible.

4. Gitlab (The Brain)

Here I store my code and my configuration files ("Infrastructure as Code"). If we make changes to our server configuration later, we save these changes in Gitlab. This way, we can undo any mistake and know exactly what we changed and when.

Outlook: What’s Next?

As you can see, we have big plans. But don't worry: We’ll eat this elephant one bite at a time. This series is explicitly written for beginners who are ready to learn. We will use the console (terminal) a lot, but I will explain every command.

In the coming weeks, we will build this architecture brick by brick.

What awaits you in Part 2:

We will pour the foundation. We will install both servers and prepare them identically.
This means:

I look forward to implementing this project together with you. Get two fresh servers (or virtual machines) ready – in the next part, we get down to business!

Do you have questions about the architecture or requests for specific details? Feel free to write me an email. See you next time!