Many of the posts I read here are about Docker. Is anybody using Kubernetes to manage their self hosted stuff? For those who’ve tried it and went back to Docker, why?
I’m doing my 3rd rebuild of a K8s cluster after learning things that I’ve done wrong and wanted to start fresh, but when enhancing my Docker setup and deciding between K8s and Docker Swarm, I decided on K8s for the learning opportunities and how it could help me at work.
What’s your story?
Kubernetes is useful if you have gone full cattle over pets. And that is very uncommon in home setups. If you only own one or two small machines you cannot destroy infra easily in a “cattle” way, and the bloatware that comes with Kubernetes doesn’t help you neither.
In homelabs and home servers the pros of Kubernetes are not very useful: high availability, auto-scaling, gitops integrations, etc: Why would you need autoscaling and HA for a SFTP used only by you? Instead you write a docker-compose.yml and call it a day.
The one exception to this is if you’re using your homelab to learn kubernetes.
That was the only time I used K8s and k3s on my homelab.
And for anything that I do want to set up in a HA/cattle kind of way, I use Docker Swarm, as it feels like a more comfortable extension of docker compose.
This mostly, I haven’t seen a compelling reason to leave my docker setup.
I think the biggest reasons for me have been growth and professional development. I started my home cluster 8 years ago as a single node of basically just running the hack/ scripts on my Linux desktop. I’ve been able to grow that same cluster to 6 hosts as I’ve replaced desktops and as I got a bit into the used enterprise server scene. I’ve replaced multiple routers and moved behind cloudflare, added a private CA a few times, added solid persistence with rook+ceph, and built my ideal telemetry stack, added velero backups into Backblaze b2, and probably a lot more I’m not thinking of.
That whole time, I’ve had to do almost zero maintenance or upgrades on the side projects I’ve built over the years, or on the self hosted services I’ve run. If you ignore the day or so a year I’ve spent cursing my propensity to upgrade a tad too early and hit snags, though I’ve just about always been able to resolve them pretty quickly and have learned even more from those times.
And on top of that, I get to take a lot of that expertise to work where it happens to pay quite well. And I’ve spent some time working towards building the knowledge into a side gig. Maybe someday that’ll pay the bills too.
One line from your comment struck a chord. The part about maintenance and upgrades. I feel like I get stuff set up and working and go about my life and then a failure happens at the most inopportune moment. Mostly, the failures are when I have a few hours free and decide to upgrade the OS and everything breaks and all the dependencies fall apart and some feature is no longer supported. That’s where I started looking to K8s to just roll back until I have time to manage it.
While you’re probably right overall, there are many good reasons to use k8s. The api provides all sorts of benefits. Kubectl, k9s, and other operational UIs . Good deployment models and tools like argo. Loads of helm charts that are (theoretically) ready to use.
No, those things aren’t free. There’s a lot of overhead to running k8s.
I run k3s and all my stuff runs in it no need to deal with docker anymore.
I’m not very familiar with kubernetes or k3s but I thought it was a way to manage docker containers. Is that not the case? I’m considering deploying a k3s cluster in my proxmox environment to test it out.
Kubernetes is abbreviated K8s (because there’s 8 letters between the “k” and the “s”. K3s is a “lite” version. Generally speaking, kubernetes manages your containers. You basicaly tell K8s what the state should be and it does what it needs to do to get the environment as you’ve declared. It’ll check and start or restart services, start containers on a node that can run them (like ensuring enough RAM is available). There’s a lot more, but that’s the general idea.
Could you list some of your “stuffs” that you run on your k3s? I’m curious.
Oh it is not that much, I run adguard DNS with adblocking, searxng as my search engine, vaultwarden as my password manager. All combined with Argo CD as GitOps engine, nginx ingress with cert-manager for lets encrypt certificates, longhorn as storage layer and metallb as loadbalancer solution. I am planning to completely replace my current setup (which is an old sandy bridge powered HP microserver) with a turing pi 2 clusterboard with 4 RPi4 CMs as soon as they get cheaper.
I manage like 200 servers in Google cloud k8s but I don’t think I’d do that for home use. The core purpose is to manage multiple servers and assign processes between them, auto scaling, cluster internal network - running docker containers for single instance apps for personal use doesn’t require this kind of complexity
My NAS software has a docker thing just built into it. I can upload or specify a package and it just runs it on the local hardware. If you have a Linux shell, I guess all you really have to do is run dockerd
to start the daemon, make sure your network config allows connections, and upload your docker containers to it for running
Kubernetes is great if you run lots of services and/or already use kubernetes at work. I use it all the time and I’ve learned a lot on my personal cluster that I’ve taken to work to improve their systems. If you’re used to managing infra already then it’s not that much more work, and it’s great to be able to shutdown a server for maintenance and not have to worry about more than a brief blip on your home services.
I use k8s at work and have built a k8s cluster in my homelab… but I did not like it. I tore it down, and currently using podman, and don’t think I would go back to k8s (though I would definitely use docker as an alternative to podman and would probably even recommend it over podman for beginners even though I’ve settled on podman for myself).
- K8s itself is quite resource-consuming, especially on ram. My homelab is built on old/junk hardware from retired workstations. I don’t want the kubelet itself sucking up half my ram. Things like k3s help with this considerably, but that’s not quite precisely k8s either. If I’m going to start trimming off the parts of k8s I don’t need, I end up going all the way to single-node podman/docker… not the halfway point that is k3s.
- If you don’t use hostNetworking, the k8s model of traffic routes only with the cluster except for egress is all pure overhead. It’s totally necessary with you have a thousand engineers slinging services around your cluster, but there’s no benefit to this level fo rigor in service management in a homelab. Here again, the networking in podman/docker is more straightforward and maps better to the stuff I want to do in my homelab.
- Podman accepts a subset of k8s resource-yaml as a docker-compose-like config interface. This lets me use my familiarity with k8s configs iny podman setup.
Overall, the simplicity and lightweight resource consumption of podman/docker are are what I value at home. The extra layers of abstraction and constraints k8s employs are valuable at work, where we have a lot of machines and alot of people that must coordinate effectively… but I don’t have those problems at home and the overhead (compute overhead, conceptual overhead, and config-overhesd) of k8s’ solutions to them is annoying there.
I’d suggest Podman over docker if someone is starting fresh. I like Podman running as rootless, but moving an existing docker to Podman was a pain. Since the initial docker setup was also a pain, I’d rather have only done it once :/
For me the use case of K8s only makes sense with large use cases (in terms of volume of traffic and users). Docker / Podman is sufficient to self-host something small.