default casing build real quick

Playing again with #DNS settings on my #archlinux. This time I decided to use #systemd-resolved to benefit of #quad9 with #dot

#linux #networking #homelab
https://www.schwitzd.me/posts/secure-dns-networkd-resolved/

New blog post: https://blog.mei-home.net/posts/cnpg-barman-plugin-migration/

A short (for me, stop snickering!) post on migrating from the integrating barman cloud backup in CNPG to the separate plugin.

I've been spending the past few months migrating my #homelab to more open source options. It's been a fun journey away from Hyper-V and Windows. Most of my virtual machines were #Linux anyway. Moving to a more efficient AMD based mini pc for hosting my various workloads.

I'm kind of wondering if this new #AI era is going to usher in a new (old) segment of business focused on "Cloud Last" and bringing things back on premises.

Da bin ich ja mal gespannt. Auf gut Glück mal drucken. Direkt all in. #unifi #homelab #3ddruck

What have been peep's experiences with building RAID arrays with 16TB, 20TB, etc. hard drives?

Too big and slow to pivot when a drive breaks? Lower reliability per spindle compared to smaller units? No differences of note?

Context: I have multiple 6 disk draid2 configs with 8TB spindles. The drives are aging out and could use an excuse to to reduce the number of spindles (less heat, noise, power, weight). Thinking about 3-4 way raidz[1|2] — or maybe something different.

Opinions and comments appreciated.

Stealthbox overview: self-host media, chat, files, and local AI offline on your hardware. Deploy in minutes. FOSS and modular; no accounts, no tracking, no subscriptions. Your data, your rules.

Watch: https://odysee.com/@Stealthbox:9/output:285?r=GKp916NtmepiSg5AGTUWyGXtrQcrwf57

Making the most out of a small LLM

Yesterday i finally built my own #AI #server. I had a spare #Nvidia RTX 2070 with 8GB of #VRAM laying around and wanted to do this for a long time.

The problem is that most #LLMs need a lot of VRAM and i don't want to buy another #GPU just to host my own AI. Then i came across #gemma3 and #qwen3. Both of these are amazing #quantized models with stunning reasoning given that they need so less resources.

I chose huihui_ai/qwen3-abliterated:14b since it supports #deepthinking, #toolcalling and is pretty unrestricted. After some testing i noticed that the 8b model performs even better than the 14b variant with drastically better performance. I can't make out any quality loss there to be honest. The 14b model sneaked in chinese characters into the response very often. The 8b model on the other hand doesn't.

Now i've got a very fast model with amazing reasoning (even in German) and tool calling support. The only thing left to improve is knowledge. #Firecrawl is a great tool for #webscraping and as soon as i implemented websearching, the setup was complete. At least i thought it was.

I want to make the most out of this LLM and therefore my next step is to implement a basic #webserver that exposes the same #API #endpoints as #ollama so that everywhere ollama is supported, i can point it to my python script instead. This way it feels like the model is way more capable than it actually is. I can use these advanced features everywhere without being bound to it's actual knowledge.

To improve this setup even more i will likely switch to a #mixture_of_experts architecture soon. This project is a lot of fun and i can't wait to integrate it into my homelab.

Hoping someone can help debug this

Playing with a WaveShare e-paper display but it’s acting weird. I’m doing a partial refresh of the time from an ESP8266 every 30 seconds. Screen is being driven by an #Arduino Uno.

Even though the only call is to update the time, every other time it fills the screen except for the partial with static. Then it refreshes and is back to normal.

Anyone seen this?

OMG, Netbox is making my ADHD brain so happy right now!

https://github.com/netbox-community/netbox

#PatchTuesday 2025-09
Patching experience in my test lab (#homelab).

Windows 11 - No issues on any test machines / VMs.
Windows 10 - No issues on any test machines / VMs. I finally see the prompt to enroll in extended updates as a local user.

Windows Server 2016 - Failed the updates 2 months in a row on the test VM. Will be rebuilding the VM.
Windows Server 2019 - Normal, uneventful.
Windows Server 2022 - Normal, uneventful.
Windows Server 2025 - Normal, uneventful.

Sorry for skipping August.

junction box build with XT60 & banana safety sockets

I got passkey login working in Authentik​​

I'm.... Not exactly sure how?

I also setup an email to notify me when a user is created (for when a friend uses the invite link)

Next up: oops synapse is still kinda broken? Then recipe app, then I should work more on this file server for my parents

#Kubernetes #Homelab #Authentik

Camping season is over, so I'm now using the Ecoflow lifepo4 powerstation to replace my older APC UPS w/ lead acid batteries.

While the APC went crazy with beeping and cooling when power was lost, the Ecoflow just doesn't blink an eye. Transfer time (spec says <10ms) seems right, all my gear just keeps running.

The Ecoflow has a lot more knobs and settings that cut power without warning, if they want to be a serious UPS they have to make it harder to accidentally shut down.

Wer viele Server per SSH verwaltet, kennt das Chaos mit Keys, Configs und Port-Weiterleitungen. SSHPilot bringt Ordnung ins Terminal und bindet Serverdaten direkt im Dateimanager ein – schlank, übersichtlich, open source

https://linuxundich.de/gnu-linux/sshpilot-linux-alternative-putty-securecrt/

Ah yes, late night maintenance at my ISP. I may have been sleeping but my monitoring wasn't. #Grafana #Prometheus #homelab

@tailscale My fresh Linux Mint 22.2 installation is an “illegal number” for your install script, eh?

We’ll just see about that!

Seriously, though, if y’all haven’t had a look at #Tailscale yet, you’re missing out! I was never able to successfully get my #homelab self-hosted until Tailscale made it DEAD EASY. Now I can achieve the dream of having my own #PersonalCloud and put an end to Apple reaching into my wallet every month.

Shout out to #TsDProxy (Tailscale Docker Proxy). Such a neat tool and works wonderfully.

Please do yourself a favour and go check it out if you #selfhost services in #docker and proxy through #tailscale.

Ii was looking for this exact solution and couldn't really believe it when I found it.

https://almeidapaulopt.github.io/tsdproxy/

Finally organized my thoughts into a blog post around the state of hypervisors in my home lab, with some new hardware. I also called out a few nice features of XCP-NG and Proxmox that I discovered while I was evaluating both of them.

https://medium.com/@a.j.longchamps/home-lab-hardware-refresh-september-2025-edition-3e50767b63e1

Been a minute since I have stepped up the #homelab but I have decided to shift my desktop to a new server. Maybe throw Fedora Server on it but I will keep it on Bluefin.

Got Paperless-NGX, NextCloud, and OnlyOffice setup to move off of cloud storage and away from MSFT/Google soon. We have been drowning in paper lately so having it stored will allow us to get the house under control a bit.

The Framework 12 has kinda been my machine when not on a work Mac and I like it a lot. Need to keep playing with it because outside of work I have been avoiding computers