Diese Perle unseres internen Schriftverkehrs wollte ich euch nicht vorenthalten.

Gerade wenn man denkt #datev läuft und man sich mit #sfirm rumärgern muss kommt sowas

Okay, I've been asked this question to which I was not prepared: where do you go if you want to learn Kubernetes from the ground up? Consider a person who doesn't even know how to run "docker" in their own laptop, but they want to rise to a point where they can actually manage a Kubernetes cluster. Thoughts?
#HomeLab #Kubernetes #SysAdminLife #k8s #SRE

I have two wolves inside me: one wants to hit the "Merge pull request" button, the other is slapping my hand.
#SysAdminLife #SRE #DevOps #HomeLab #Friday

NPU can be a "Neural Processing Unit" or a "Network Processing Unit". They're two absurdly different things though.

Why? Why can't people just choose different acronyms? Hype?

Yep. I deployed to PROD on a Friday afternoon, and I'm not ashamed to admit (after all, I'm the one who needs to deal with the fall down anyway).
#SysAdminLife #HomeLab

You know what's not normal at all?
I'm drinking a simple cup of coffee with this Chobani Dairy Coffee Creamer Caramel Macchiato, and the smell is simply overwhelming all around my house, like if I lighted up a scented candle. What in the world is in that thing?
#SysAdminLife #Coffee
https://www.chobani.com/products/creamers/dairy/caramel-macchiato

Just read this from the status panel from one of my VPS providers:

“8th March 2025: 100.0000% uptime, 129 minutes maintenance”

The service was certainly down within that maintenance period.

Have I been doing my percentages wrong? If I brought something down for half an hour, that's ~2% of the day, so I would say uptime that day was 98%. Isn't it rather shady to say uptime was 100.0000% (six nines!), because all the downtime was during planned maintenance?

This morning, once again, I come across an interesting project. I head over to GitHub to take a look.

I read that the only supported way to use the project is via Docker because "providing installation instructions, given the underlying complexity, would be impossible." Not a fan of that, but okay, I start to understand… So, as I often do, I decide to open the Dockerfile (there’s no Compose). First line:

FROM alpine:3.14 AS base

No, your project isn’t "just" complex. It’s running on an outdated and potentially insecure base. And that’s a much bigger problem. It probably means it’s such a mess that updating it is impossible.

I close the GitHub page and move on with my life.

This is a beautiful time lapse. Kind of makes me miss doing work inside a data center.
#SysAdminLife #Infrastructure #DataCenter #Hetzner
https://youtu.be/bzs2Gisptm0

Suddenly I realized: there was a time when a lot of open source projects were using Slack to chat, but recently I have seen a spike in new projects using Discord instead. Considering that before Slack it was IRC, I wonder if it's a generational thing, rather than a tooling problem.
#OpenSource #Linux #SysAdminLife #HomeLab

Devs sometimes battle legacy code. But they workaround deploying a newer version, or writing new code.

Infra engineers (like me) sometimes have to battle *legacy hardware*. And it's not like you can just run into a closet and pick up a new one. I wish it would be that easy.

That feeling when your blog posts are being used as a reference to prove a point.
#HomeLab #SelfHosted #SysAdminLife

I'm having a hard time remembering some tool as badly documented as the Phoronix Test Suite...
#HomeLab #SysAdminLife
https://github.com/phoronix-test-suite/phoronix-test-suite/blob/master/documentation/phoronix-test-suite.md

Ooh boy. I completely forgot that there is a new major version of Pi-hole. This is definitely not going to be a straightforward upgrade for me, considering I have 3 instances and so much automation around it. Here it goes, another item to the backlog.
#HomeLab #SelfHosted #PiHole #SysAdminLife

PSA: Don't try this at home

sudo rm -rf /etc/{ceph,cni,kubernetes,rancher,}

My app server was a little busy this morning. Never had load averages that high before! All due to a bug in my processing code. Oops #sysAdminLife #Linux #PHP

Okay, that (kind of) settles it! 40% of the voters said I should keep writing about running things with Docker Compose in my blog, so that's what I'm gonna do.

However a close 36% said they also wanted me to add Podman + SystemD as an alternative. I will take feedback that into consideration, and think about how I can incorporate them, even if just a little bit (no promises though).

And what did I learn through the course of pondering/researching about this subject? Some people just want to use their home lab to self-host stuff, and learn about the stuff they host, not necessarily spend time or learn about the *infrastructure* behind the home lab, or the multiple ways to run it as similar as possible to a production environment (either with less/savaged hardware, or not caring about redundancy/HA).

Those were some very enlightening couple of days. Let's see how I can better make use of this research to tailor the way I share my experiences.

Thanks all who voted/commented! (and all the random posts I read to complement my data)

#HomeLab #SelfHosted #SysAdminLife #Forgejo

https://hachyderm.io/@badnetmask/113997118763827937

Poll: I've been reading some posts about people who converted their Docker Compose deployments to some sort of Kubernetes deployment (k3s, minikube or microk8s).

Since I've been writing a lot of posts about self-hosting apps, I've been wondering about what people around here think would be better.