01 Platform
Systems and containers
Linux servers, Docker, service management, backups. Building a working system and keeping it up.
Systems & networks · an intern’s notebook
I build systemsthat work,then write them down.
I'm a systems and networking intern at Netinternet. I set up servers, chase packets that go missing somewhere on the network, and when a service crashes I try to find out why it crashed instead of just restarting it. This is where I write up those problems and what I learned solving them.
TürkiyeScroll down
Understand first, then change.
Infrastructure work isn’t tested on the day it works; it’s tested three months later, on the day it breaks. So when I fix something I try to write down why — for myself, and for whoever maintains it next.
- 01
Read before you touch
Before changing a server I try to understand why it was set up that way. A fix made without knowing the cause brings the same bug back three weeks later.
- 02
Write it, then run it
Before running a command I write down what it should do. It leaves a note for later, and if I was wrong, nothing is lost.
- 03
Start small
Before building something complex I set up the smallest version that works. Learning deployment at the start beats migrating everything at the end.
The right tool, on the right day.
I’m not learning every tool in the box — I’m learning which one to reach for, and when. These four carry most of my day-to-day work.
02 Network
Network and web
DNS, TLS, routing, firewalls. Building the whole path from the browser to the server.
03 Automation
Automation
Turning everything done by hand twice into a script. Backups, monitoring, certificate renewal.
04 Data
Data
Databases and file formats. Keeping content in files — easy to back up, easy to move.
Systems I’ve built.
Everything here runs on my own server and was set up by hand. All of it can be rebuilt from scratch: there are scripts, and there’s a daily backup.
- 01This blog's infrastructureThe site itself: Next.js static export, Caddy, SQLite and automatic rebuilds. Why does a post go live 40 seconds after I write it?
- 02Automated backup systemA single script that backs up eight Docker volumes and their configuration every day, then opens the archive to check it can actually be read.
- 03Server monitoringuptime-kuma and ntfy, set up so that a notification lands on my phone when a service goes down.
- 04A fully local AI botI hooked a Discord bot up to a model running locally with Ollama; not a single request leaves the server.
- 05SSH and password managementTurning off password logins in favour of key-based access, and running a password vault with Vaultwarden.
$
$
This page is being served from this very box. The numbers are live.
Latest pages from the notebook.
- The first ten minutes of a customer outageThe first step in fixing an outage isn't fixing it — it's working out which step you skipped. Because once you reboot, you've wiped out the cause too.Read
- Two operating systems, one job: the real differences between Windows Server and LinuxWhich one is better is the wrong question. The difference is this: in which one does the error message tell you what to do, and in which one are you left to figure it out yourself?Read
- Ten questions the Ubuntu Server installer asksThe installer shows ten screens and it feels like you can click through them all. None of them is trivial — get one wrong and the server is installed, but useless.Read
- First day in the server room: meeting a physical serverThe first time I walked into the server room, I saw how different the gap between physical and virtual is from what the theory says.Read