My homelab is a mess. What’s that phrase again? “The carpenter’s house is never finished” or is it “The shoemaker’s children always go barefoot”? Whatever it is, it applies to me because I’m a terrible sysadmin at home.
Over the years, my homelab has undergone several transformations, progressing from the use of subpar Linksys routers, consumer-grade Dell and HP desktops, unmanaged switches, and old rack-mount servers with high power consumption and heat output.
Operating within the limitations of old equipment was challenging, and I was constrained to using outdated and insecure operating systems, experiencing peculiar stability and performance issues attributable to aging hardware. I needed robust, scalable, and contemporary enterprise-grade hardware to continue working towards my goal of centralizing all my data in-house.
Previous Homelab Issues
Fortunately, last summer, I was able to downscale from three HP 360 G6s running Proxmox and a Supermicro 840 chassis running TrueNAS to a Dell R730 running vSphere 8 and an InforTrend DS3024 running SANWatch, a proprietary Linux-based OS for creating and managing iSCSI LUN targets.
My homelab network is antiquated. I’m currently using HP 1810 and 1910 series switches with Aruba iAP access points. I intend to upgrade my network equipment when I have the necessary bandwidth (heh.).
Going Forward With My Homelab
My present hardware setup is relatively new, runs the latest software, runs cool, and only draws 700 watts of electricity.
So, what am I hosting in my homelab? Not much. I haven’t ran hot and heavy for a while. Two domain controllers, two DFS servers, a docker server with various torrent downloaders + Plex, a web server for my personal resume, Minecraft, VCSA, and Veeam NFR.
There’s a lot wrong with my current setup.
All of my servers, even the ones exposed to the internet, are on the same VLAN. The Plex server allows full read/write access to my DFS multimedia share and is exposed to my WAN.
Unfortunately, Veeam is not running regular backups, and some servers are not being backed up.
Furthermore, I have no means of auditing, monitoring, and updating my network.
There is no automation, documentation, and standardization in my homelab. They have been perpetually on my to-do list. Though I’m not an incompetent sysadmin, I must admit that laziness is a significant factor hindering my efforts.
As such, I am now preparing to restructure my network from the ground up, focusing on security, compliance, and doing things right.
I aim to centralize and self-host most of my data in my homelab, and I believe the best way to achieve this goal is to slowly start fixing things one by one.
Conclusion
In conclusion, I intend to document my journey on Twitter, blog posts, and my wiki. With some luck, I’ll be able to create a capable lab that garners significant nerd cred.
The real work begins tomorrow.
It’s a lot of fun running a homelab. I host pretty much everything in single node K8s on a single 6-core/12-thread machine. Besides common things like a file server and a video streaming service, some less common services I self-host are:
A weather station reading wireless Acurite sensors placed around my property via an off-the-shelf SDR usb module. Pretty graphs are presented in Prometheus/Grafana.
Headless game streaming service. I can play Windows games on my Arm macbook at the airport over my home VPN, out of a debian-xfce container, running in a k8s cluster, running on Ubuntu server. All with minimal lag. Still amazes me.
I’m often thinking about what I want to do next. I think next is an offsite backup for my ZFS mirror. Thinking about setting up a small machine at my brother’s, 300 miles away, and sending daily/weekly/monthly ZFS snapshots to it. Been spending a lot of time looking for dirt cheap hardware deals on Facebook marketplace and offerup.