I've got about 1-2 years of experience with Kubernetes as a platform engineer, and I'm ready to dive deeper this year. I'm aiming to pursue certifications while also building a tiny Kubernetes cluster on my own desktop. I'm thinking about using three Raspberry Pis—one with 8GB for the control node, one with 8GB as a worker node, and another smaller one with 4GB as a secondary worker.
Is this setup viable for what I'm trying to learn? Am I on the right track with my questions? Any advice would really help. My goal is to learn everything from the basics, overcome challenges, and this seems like the first big step: getting started!
4 Answers
Honestly, you might get more bang for your buck with a capable ultra-small form factor PC. You could run a hypervisor like Proxmox to virtualize everything instead.
Starting with what you proposed could work, but you might want to consider expanding later on. Multi-purpose nodes could help with high availability if you have that in mind down the line.
I usually go for all-in-one virtual machines. My lab doesn't really need high availability.
It seems a bit tricky to have a functional control plane without at least three nodes. Are you sure you want to start with this setup?
That's what caught my attention too. With the rising prices of Raspberry Pis, is it even worth it trying to set this up in a diluted form?
I'm not too worried about high availability right now. I'll just be deploying a simple hello world app first, and then scale later. Am I thinking about this right?
Yeah, this is just to kick things off. I definitely plan to grow from this initial setup!