this post was submitted on 28 Apr 2025
14 points (100.0% liked)

graybeard

251 readers
23 users here now

Stories, links, experiences from calculator manipulators with a few grays in their beard

founded 2 years ago
MODERATORS
 

All on my own!

I've heard horror stories about it taking months upon months in the past and they've been keeping me from trying. Silly me :)

It took under two days; and most of the time was trying to figure out all the problems caused by being behind a an http proxy. Looking back, I can see it can be done in a minute or so now that I've got a few bits of config nailed down.

Granted this is a very bare bones cluster atm, but applying that test nginx deployment and seeing all three replicas running, one on each node, has sent me buzzing :)

Next step - generating an internal intermediate and shoving it into the cluster.

Happy Monday!

you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 2 points 1 day ago (2 children)

I have an interest in this as well, what resources did you use?

[–] [email protected] 1 points 1 day ago

Official docs, that's it. kubeadm init with a pod cidr and dns name for the control plane.

For network and network policies I went with calico. I think flannel and ovn are better, but our prod and staging on eks run calico, hence the choice. Metrics server afterwards and now I'm at the before mentioned spot of trying to figure out certs. I think cert-manager is my choice, but at the same time I got sidetracked with reworking our internal ca. Current is rsa and I want ed25519 plus a bit more automation.

[–] [email protected] 1 points 1 day ago* (last edited 1 day ago)

I started a few weeks ago. My life is minikube helm and skaffolds right now.

They're is a really good ad free like 4 hour or something intro a lady put on YouTube. Otherwise it's just me the docs and the LLM.