Infrastructure

Infrastructure Information

As a hobby, I run a lot of my own infrastructure, including the server serving this website! If you're a nerd, curious, or both, this page has a bunch of information as to how everything works. As a disclaimer, I wouldn't take this as an example of the "best" way to do everything I'm running, I'm sure it can be improved in many ways - It's just how I've enjoyed setting everything up.

Server Specifications

A picture of my server in the rack. The case is black, with an AMD Ryzen sticker visible. It has a vent on the left for airflow and a locked door on the right with front USB ports alongside power on/off and restart buttons behind it.

Jupiter

This is the older of the two servers. Originally, I built it in 2021, however, in December 2024 I upgraded it from a PG6405 to a Ryzen 5 7600, meaning a lot of parts have since been replaced. Then, in 2025, I bought myself a server rack and mounted it to it, where it is today.

CPU: Ryzen 5 7600 (6 cores, 12 threads)

RAM: 64GB DDR5-4800

OS: Void Linux

Services: Bluesky, CDN, Forgejo

A picture of my server, in the rack. It's grey with 32 2.5 inch drive bays visible in the front, as well as an Intel Xeon sticker. The first 10 drive bays are empty and you can see inside the chassis, while the remainder have blank drive caddies.

Saturn

While the hardware is older than Jupiter's, the server is new to me, having acquired it in late 2025. The server was originally going to be thrown out by the previous owners, but instead I was able to take it. Now, it runs a significant amount of the services on my infrastructure.

CPU: 2x Intel Xeon Silver 4114 (20 cores, 40 threads)

RAM: 128GB DDR4-2333

OS: Void Linux

Services: Fediverse, Website


The services running on each server aren't necessarily a hard restriction, from time to time they may be moved across to the other. Most of them are in virtual machines and the data for that is stored on Jupiter because it has fast NVMe drives, which, even over the network are quite fast. This means that while Saturn can fail and services can be moved across to Jupiter, a failure of Jupiter is difficult to recover from.

Both servers are connected across two networks, one by the ISP provided router, and my internal network. Jupiter also acts as a gateway for my devices and the virtual machines to access the internet, as well as providing DNS, DHCP and broadcasts a wireless network. In order to serve these services, each server has an instance of Cloudflared running on it which, using cloudflare tunnels, exposes resources to the internet - Should either server lose connection, the other Cloudflared instance can start serving requests.

Recently, I've been migrating services away from Docker containers on the host, and towards VMs. These run a mixture of operating systems, primarily Debian, Alpine Linux and FreeBSD depending on the needs of the particular service running.

Status Page

Infrastructure

Homepage