You have a lot of responses here, but I’ll tell what k8s actually is, since a lot of people seem to get this wrong.

Just like k8s, docker has many tools. Although docker is packaged in a way, that it looks like it’s just 1 tool. This is docker desktop. Under the hood there is docker engine that is really a runtime and image management service and API. You can look at this more if you wanted. There is containerd, runc, cri-o. These were all created so that different implementations can all talk to this API in a standard way and work.

Moving on to k8s. K8s is a way to scale these containers to run in different ways and scale horizontally. There are ways to even scale nodes vertically and horizontally to allow for more or less resources to place these containers on. This means k8s is very event driven and utilizes a lot of APIs to communicate and take action.

You said that you are doing kubectl apply constantly and you say feels wrong. In reality, this is correct. Under the hood you are talking with the k8s control plane and it’s taking that manifest and storing it. Other services are communicating with the control plane to understand what they have to do. In fact you can apply a directory of manifests, so you don’t have to specify each file individually.

Again there are many tools you can use to manage k8s. It is an orchestration system to manage pods and run them. You get to pick what tool you want to use. If you want something you can do from a git repo, you can use something like argocd or flux. This is considered to be gitops and more declarative. If you need a templating implementation, there are many, like helm, json net, and kustomize (although not a full templating language). These can help you define your manifests in a more repeatable and meaningful way, but you can always apply these using the same tools (kubectl, argocd, flux, etc…)

There are many services that can run in k8s that will solve one problem or another and these tools scale themselves, since they mostly all use the same designs that keep scalability in mind. I kept things very simple, but try out vanilla k8s first to understand what is going on. It’s great that you are questioning these things as it shows you understand there is probably something better that you can do. Now you just need to find the tools that are right for you. Ask what you hate or dislike about what you are doing and find a way to solve that and if there are any tools that can help. https://landscape.cncf.io/ is a good place to start to see what tools exist.

Anyway, good luck on your adventure. K8s is an enterprise tool after all and it’s not really meant for something like a home lab. It’s an orchestration system and NOT a platform that you can just start running stuff on without some effort. Getting it up and running is day 1 operations. Managing it and keeping it running is day 2 operations.

You’re right to be reluctant to apply everything by hand. K3s has a built-in feature that watches a directory and applies the manifests automatically: https://docs.k3s.io/installation/packaged-components

This can be used to install Helm charts in a declarative way as well: https://docs.k3s.io/helm

If you want to keep your solution agnostic to the kubernetes environment, I would recommend that you try ArgoCD (or FluxCD, but I never tried it so YMMV).

Ive actually been personally moving away from kubernetes for this kind of deployment and I am a big fan of using ansible to deploy containers using podman systemd units, you have a series of systemd .container files like the one below

[Unit]
Description=Loki

[Container]
Image=docker.io/grafana/loki:3.4.1

# Use volume and network defined below
Volume=/mnt/loki-config:/mnt/config
Volume=loki-tmp:/tmp/loki
PublishPort=3100:3100
AutoUpdate=registry

[Service]
Restart=always
TimeoutStartSec=900

[Install]
# Start by default on boot
WantedBy=multi-user.target default.target

You use ansible to write these into your /etc/containers/systemd/ folder. Example the file above gets written as /etc/containers/systemd/loki.container.

Your ansible script will then call systemctl daemon-reload and then you can systemctl start loki to finish the example

The answer to the first two questions are helm charts. They are collections of parametrized yaml files and the most popular way to install things into k8s. You just need one config file for each helm release (values.yaml).

If you want to go declarative with gitops rather than imperative, check ArgoCD or flux.

I use Kube everyday for work but I would recomend you to not use it. It’s complicated to answer problems you don’t care about. How about docker swarm, or podman services ?

I’ll post more later (reply here to remind me), but I have your exact setup. It’s a great way to learn k8s and yes, it’s going to be an uphill battle for learning - but the payoff is worth it. Both for your professional career and your homelab. It’s the big leagues.

For your questions, no to all of them. Once you learn some of it the rest kinda falls together.

I’m going into a meeting, but I’ll post here with how I do it later. In the mean time, pick one and only one container you want to get started with. Stateless is easier to start with compared to something that needs volumes. Piece by piece brick by brick you will add more to your knowledge and understanding. Don’t try to take it all on day one. First just get a container running. Then access via a port and http. Then proxy. Then certs. Piece by piece, brick by brick. Take small victories, if you try to say “tomorrow everything will be on k8s” you’re setting yourself up for anger and frustration.

@sunoc@sh.itjust.works Edit: To help out I would do these things in these steps, note that steps are not equal in length, and they are not complete - but rather to help you get started without burning out on your journey. I recommend just taking each one, and when you get it working rather than jumping to the next one, instead taking a break, having a drink, and celebrating that you got it up and running.

0. Start documenting everything you do. The great thing about kubernetes is that you can restart from scratch if you have written everything down. I would start a new git repository with a README that contains every command you ran, what it did, and why you did it. Assume that you will be tearing down your cluster and rebuilding it - in fact I would even recommend that. Treat this first cluster as your testing grounds, and then you won’t feel crappy spinning up temporary resources. Then, you can rebuild it and know that you did a great job - and you’ll feel confident in rebuilding in case of hardware failure.

1. Get the sample nginx pod up and running with a service and deployment. Simply so you can curl the IP of your main node and port, and see the response. This I assume you have played with already.

2. Point DNS to your main node, get the nginx pod with http://your.dns.tld:PORT. This should be the same as anything you’ve done with docker before.

3. Convert the yaml to a helm chart as other have said, but don’t worry about “templating” yet, get comfortable with helm install, helm upgrade -i, and helm uninstall. Understand what each one does and how they operate. Then go back and template, upgrade-ing after each change to understand how it works. It’s pretty standard to template the image and tag for example so it’s easy to upgrade them. There’s a million examples online, but don’t go overboard, just do the basics. My (template values.yaml) usually looks like:

<<servicename>>
  name: <<servicename>>
  image:
    repository: path/to/image
    tag: v1.1.1
    network:
     port: 8888

Just keep it simple for now.

4. Decide on your proxy service. Traefik as you see comes out of the box. I personally use istio. I can go into more details why later, but I like that I can create a “VirtualService” for "$appname.my.custom.tld` and it will point to it.
5. Implement your proxy service, and get the (http only still) app set up. Set up something like nginx.your.tld and be able to curl http://nginx.your.tld and see that it routes properly to your sample nginx service. Congrats, this is a huge one.
6. Add the CertManager chart. This will set it up so you can create Certificate types in k8s. You’ll need to use the proxy in the previous step to route the /.well-known endpoints on the http port from the open web to cert-manager, for Istio this was another virtual service on the gateway - I assume Traefic would have something similar to “route all traffic on port 80 that starts with /.well-known to this service”. Then, in your nginx helm chart, add in a Certificate type for your nginx endpoint, nginx.your.tld, and wait for it to be successfully granted. With Istio, this is all I need now to finally curl https://nginx.your.tld!

At this point you have routing, ports, and https set up. Have 2 drinks after this one. You can officially deploy any stateless service at this point.

Now, the big one, stateful. Longhorn is a bear, there are a thousand caveats to it.

Step one is where are your backups going to go. This can be a simple NFS/SMB share on a local server, it can be an s3 endpoint, but seriously this is step 1. Backups are critical with longhorn. You will fuck up Longhorn - multiple times. Losing these backups means losing all configs to all of your pods, so step one is to decide on your stable backup location.

Now, read the Longhorn install guide: https://longhorn.io/docs/1.9.0/deploy/install/. Do not skip reading the install guide. There are incredibly important things in there that I regretted glossing over that would have saved me. (Like setting up backups first).

The way I use longhorn is to create a PV in longhorn, and then the PVC (you can look up what both of these are later). Then I use Helm to set what the PVC name is to attach it to my pod. Try and do this with another sample pod. You are still not ready to move production things over yet, so just attach it to nginx. exec into it, write some data into the pvc. Helm uninstall. See what happens in longhorn. Helm install. Does your PVC reattach? Exec in, is your data still there? Learn how it works. I fully expect you to ping me with questions at this point, don’t worry, I’ll be here.

Longhorn will take time in learning, give yourself grace. Also after you feel comfortable with it, you’ll need to start moving data from your old docker setup to Longhorn, and that too will be a process. You’ll get there though. Just start with some of your lower priority projects, and migrate them one by one.

After all of this, there is still more. You can automount smb/nfs shares directly into pods for media or anything. You can pass in GPUs - or I even pass in some USB devices. You can encrypt your longhorn things, you can manage secrets with your favorite secret manager. There’s thousands of things you’ll be able to do. I wish you luck, and feel free to ping me here or on Matrix (@scrubbles@halflings.chat) if you ever need an ear. Good luck!

I’ve thought about k8s, but there is so much about Docker that I still don’t fully know.

And this is why I do not like K8s at all. The only reason to use it is to have something on your CV. Besides that, Docker Swarm and Hashicorp Nomad feel a lot better and are a lot easier to manage.

I’m the creator of ansible k3s playbooks, event if I not more active maintener. But there’s a big community : give a try and contrib https://github.com/k3s-io/k3s-ansible

K3s (and k8s for that matter) expect you to build a hierarchy of yaml configs, mostly because spinning up docker instances will be done in groups with certains traits applying to whole organization, certain ones applying only to most groups, but not all, and certain configs being special for certain services (http nodes added when demand is higher than x threshold).

But I wonder why you want to cluster navidrome or pihole? Navidrome would require a significant load before service load balancing is required (and non-trivial to implement), and pihole can be put behind a round-robin DNS forwarder, and also be weird to implement behind load balancing.

Firstly, I want to say that I started with podman (alternative to docker) and ansible, but I quickly ran into issues. The last issue I encountered, and the last straw, was that creating a container, I was frustrated because Ansible would not actually change the container unless I used ansible to destroy and recreate it.

Without quadlets, podman manages it’s own state, which has issues, and was the entire reason I was looking into alternatives to podman for managing state.

More research: https://github.com/linux-system-roles/podman: I found an ansible role to generate podman quadlets, but I don’t really want to include ansible roles in my existing ansible roles. Also, it intakes kubernetes yaml, which is very complex for what I am trying to do. At that point, why not just use a single node kubernetes cluster and let kubernetes manage state?

So I switched to Kubernetes.

To answer some of your questions:

Am I really supposed to have a collection of small yaml files for everything, that I use with kubectl apply -f ?? It feels wrong and way too “by hand”! Is there a more scripted way to do it? Should I stay with everything in Ansible ??

So what I (and the industry) uses is called “GitOps”. It’s essentially you have a git repo, and the software automatically pulls the git repo and applies the configs.

Here is my gitops repo: https://github.com/moonpiedumplings/flux-config. I use FluxCD for GitOps, but there are other options like Rancher’s Fleet or the most popular ArgoCD.

As a tip, you can search github for pieces of code to reuse. I usually do path:*.y*ml keywords keywords to search for appropriate pieces of yaml.

I see little to no example on how to deploy the service containers I want (pihole, navidrome, etc.) to a cluster, unlike docker-compose examples that can be found everywhere. Am I looking for the wrong thing?

So the first issue is that Kubernetes doesn’t really have “containers”. Instead, the smallest controllable unit in Kubernetes is a “pod”, which is a collection of containers that share a network device. Of course, pods for selfhosted services like the type this community is interested in will rarely have more than one container in them.

There are ways to convert a docker-compose to a kubernetes pod.

But in general, Kubernetes doesn’t use compose files for premade services, but instead helm charts. If you are having issues installing specific helm charts, you should ask for help here so we can iron them out. Helm charts are pretty reliable in my experience, but they do seem to be more involved to set up than docker-compose.

Even official doc seems broken. Am I really supposed to run many helm commands (some of them how just fails) and try and get ssl certs just to have Rancher and its dashboard

So what you’re supposed to do is deploy an “ingress”, (k3s comes with traefik by default), and then use cert-manager to automatically apply get letsencrypt certs for ingress “objects”.

Actually, traefik comes with it’s own way to get SSL certs (in addition to ingresses and cert manager), so you can look into that as well, but I decided to use the standardized ingress + cert-manager method because it was also compatible with other ingress software.

Although it seems complex, I’ve come to really, really love Kubernetes because of features mentioned here. Especially the declarative part, where all my services can be code in a git repo.

Hey there,

I made a similar journey a few years ago. But I only have one home server and do not run my services in high availability (HA). As @non_burglar@lemmy.world mentioned, to run a service in HA, you need more than “just scaling up”. You need to exactly know what talks when to whom. For example, database entries or file writes will be difficult when scaling up a service not ready for HA.

Here are my solutions for your challenges:

• No, you are not supposed to run kubectl apply -f for each file. I would strongly recommend helm. Then you just have to run helm install per service. If you want to write each service by yourself, you will end up with multiple .yaml files. I do it this way. Normally, you create one repository per service, which holds all YAML files. Alternatively, you could use a predefined Helm Chart and just customize the settings. This is comparable to DockerHub.
• If you want to deploy to a cluster, you just have to deploy to one server. If in your .yaml configuration multiple replicas are defined, k8s will automatically balance these replicas on multiple servers and split the entire load on all servers in the same cluster. If you just look for configuration examples, look into Helm Charts. Often service provide examples only for Docker (and Docker Compose) and not for K8s.
• As I see it, you only have to run a single line of install script on your first server and afterward join the cluster with the second server. Then you have k3s deployed. Traefik will be installed alongside k3s. If you want to access the dashboard of Traefik and install rancher and longhorn, yes, you will have to run multiple installations. Since you already have experience with Ansible, I suggest putting everything for the “base installation” into one playbook and then executing this playbook one.

Changelog:

• Removeing k3s install command. If you want to use it, look it up on the official website. Do not copy paste the command from a random user on lemmy ;) Thanks to @atzanteol@sh.itjust.works for bringing up this topic.

Yeah - k8s has a bit of a steep learning curve. I recentlyish make the conversion from “a bunch of docker-compose files” to microk8s myself. So here are some thoughts for you (in no particular order).

I would avoid helm like the plague. Everybody is going to recommend it to you but it just puts a wrapper on a wrapper and is MUCH more complicated than what you’re going to need because you’re not spinning up hundreds of similar-but-different services. Making things into templates adds a ton of complexity and overhead. It’s something for a vendor to do, not a home-gamer. And you’re going to need to understand the basics before you can create helm charts anyway.

The actual yml files you need are actually relatively simple compared to a helm chart that needs to be parameterized and support a bazillion features.

So yes - you’re going to create a handful of yml files and kubectl apply -f them. But - you can do that with Ansible if you want, or you can combine them into a single yml (separate sections with ----).

What I do is - for each service I create a directory. In it I have name_deployment.yml, name_service.yml, name_ingress.ymlandname_pvc.yml`. I just apply them when I change them, which isn’t frequent. Each application I deploy generally has its own namespace for all its resources. I’ll combine deployments into a NS if they’re closely related (e.g. prometheus and grafana are in the same NS).

Do yourself a favor and install kubens which lets you easily see and change your namespace globally. Gawd I hate having to type out my namespace for everything. 99% of the time when you can’t find a thing with kubectl get you’re not looking in the right namespace.

You’re going to need to sort out your storage situation. I use NFS for long-term storage for my pods and have microk8s configured to automatically create space on my NFS server when pods request a PV (persistent volume). You can also use local directories but that won’t cluster.

There are two basic types of “ingress” load balancing. “ClusterIp” means the cluster controller will act like a hostname-based router for HTTP. You can point your DNS entries at that server and it will route to your pods on their internal IP address based on the DNS name of the request. It’s easy to use and works very well - but it only works for HTTP traffic. The other is to use LoadBalancerIp that will give your pods an IP address on the network that you can connect to directly. The former only works for HTTP, the latter will let you use any ports (e.g. ssh for a forgejo instance).

For question 1: You can have multiple resource objects in a single file, each resource object just needs to be separated by ---. The small resource definitions help keep things organized when you’re working with dozens of precisely configured services. It’s a lot more readable than the other solutions out there.

For question 2, unfortunately Docker Compose is much more common than Kubernetes. There are definitely some apps that provide kubernetes documentation, especially Kubernetes operators and enterprise stuff, but Docker-Compose definitely has bigger market share for self-hosted apps. You’ll have to get experienced with turning a docker compose example into deployment+service+pvc.

Kubernetes does take a lot of the headaches out of managing self-hosted clusters though. The self-healing, smart networking, and batteries-included operators for reverse-proxy/database/ACME all save so much hassle and maintenance. Definitely Install ingress-nginx, cert-manager, ArgoCD, and CNPG (in order of difficulty).

Try to write yaml resources yourself instead of fiddling with Helm values.yaml. Usually the developer experience is MUCH nicer.

Feel free to take inspiration/copy from my 500+ container cluster: https://codeberg.org/jlh/h5b/src/branch/main/argo

In my repo, custom_applications are directories with hand-written/copy-pasted yaml files auto-synced via ArgoCD Operator, while external_applications are helm installations, managed via ArgoCD Operator Applications.

Everyone talks about helm charts.
I tried them and hate writing them.
I found garden.io, and it makes a really nice way to consume repos (of helm charts, manifests etc) and apply them in a sensible way to a k8s cluster.
Only thing is, it seems to be very tailored to a team of developers. I kinda muddled through with it, and it made everything so much easier.
Although I massively appreciate that helm charts are used for most projects, they make sense for something you are going to share.
But if it’s a solo project or consuming other people’s projects, I don’t think it really solves a problem.

Which is why I used garden.io. Designed for deploying kubernetes manifests, I found it had just enough tooling to make things easier.
Though, if you are used to ansible, it might make more sense to use ansible.
Pretty sure ansible will be able to do it all in a way you are familiar with.

As for writing the manifests themselves, I find it rare I need to (unless it’s something I’ve made myself). Most software has a k8s helm chart. So I just reference that in a garden file, set any variables I need to, and all good.
If there aren’t helm charts or kustomize files, then it’s adapting a docker compose file into manifests. Which is manual.
Occasionally I have to write some CRDs, config maps or secrets (CMs and secrets are easily made in garden).

I also prefer to install operators, instead of the raw service. For example, I use Cloudnative Postgres to set up postgres databases.
I create a CRD that defines the database, and CNPG automatically provisions all the storage, pods, services, config maps and secrets.

The way I use kubernetes for the projects I do is:
Apply all the infrastructure stuff (gateways, metallb, storage provisioners etc) from helm files (or similar).
Then apply all my pods, services, certificates etc from hand written manifests.
Using garden, I can make sure things are deployed in the correct order: operators are installed before trying to apply a CRD, secrets/cms created before being referenced etc.
If I ever have to wipe and reinstall a cluster, it takes me 30 minutes or so from a clean TalosOS install to the project up and running, with just 3 or 4 commands.

Any on-the-fly changes I make, I ensure I back port to the project configs so when I wipe, reset, reinstall I still get what I expect.

However, I have recently found https://cdk8s.io/ and I’m meaning to investigate that for creating the manifests themselves.
Write code using a typed language, and have cdk8s create the raw yaml manifests. Seems like a dream!
I hate writing yaml. Auto complete is useless (the editor has no idea what format the yaml doc should take), auto formatting is useless (mostly because yaml is whitespace sensitive, and the editor has no idea what things are a child or a new parent). It just feels ugly and clunky.