So I recently acquired three raspberry pi 3B’s. I want to get into hosting my own services as I’ve been looking at home labing for quite some time out of curiosity and a piece of entertainment, but I never had a moment to jump in. Now having three sad pi’s doing no work in my drawer all day I felt like this was the time to start. I want to deploy services onto them like pi-hole, pi-VPN, home assistant, et cetera, … Just light weight services to get my toes wet.
I looked into my options and I thought that clustering them together made the most sense. This way, there is redundancy if something ever were to happen to one, scalability if I were to acquire more pi’s, and distributing of resources, and ease of management since there is only one node to talk to. The only problem is that when reading into kubernetes and watching video’s on how other people do it I feel kinda overwhelmed, and it looks so complicated.
It’s not that I’m new to linux, the command line, or deploying containers. Not to say I’ve got all the required skills. Subnetting, and all else network wise is still very new and abstract to me (if anyone has some great guids for noobs lmk, it’d be appricated, none of what I watched seems to have clicked). But for some reason it just looks way to hard to do. And I can barely understand what the guide is doing.
Am I forgetting an easier way, am I just overcomplicating things, because that is what it feels like. Or am I just affraid to jump in for real?
A combination of glusterfs, keepalived, and docker swarm might be for you.
Glusterfs to create a persistent storage pool for access from all nodes. Keepalived to create one dynamic and highly available network address to point everything external to. Docker swarm for the benefits of a cluster without all the k8s.