AI / Agents

Skills
llms.txt
llms-full.txt

Community

Full license

Homelab cluster with k3s and Flux

This repo configures a single Kubernetes (k3s) cluster with Ansible and uses the GitOps tool Flux to manage its state.

✨ Features

Automated, reproducible, customizable setup through Ansible templates and playbooks
Opinionated implementation of Flux from the Home Operations Community’s template
Encrypted secrets with SOPS and Age
Web application firewall provided by Cloudflare Tunnels
SSL certificates from Cloudflare and cert-manager
HA control plane capability via kube-vip
Next-gen networking using Cilium
A Renovate-ready repository with pull request diffs provided by flux-local
Integrated GitHub Actions

… and more!

📝 Prerequisites

A domain managed on Cloudflare.
A DNS server that supports split DNS (e.g., Pi-hole) deployed somewhere outside your cluster on your home network.

🚀 Installation

📍 Set up your local environment

Ensure just, direnv, and uv are available.
This repository is configured with a nix flake to create a devshell with the necessary tools. Use nix develop to enter it manually, or configure direnv to activate it automatically.

Manual activation:
```
nix develop ...
```
If running with nix develop fails, add --extra-experimental-features flags:
```
nix --extra-experimental-features 'nix-command flakes' develop ...
```
Automatic activation with direnv:

Add the following to .envrc (or update it if it already exists):
```
use flake
```
Then run direnv allow . to enable auto-activation.

The devshell provides the required CLI tools (kubectl, flux, sops, age, cloudflared, etc.) and keeps versions consistent.

🔧 Initial configuration

Set up Age private/public key

📍 Using SOPS with Age allows us to encrypt secrets and use them in Ansible and Flux.
1. Create an Age private/public key (this file is gitignored)
```
age-keygen -o age.key
```
1. Ensure that this key is available as an environment variable.
Add the following to the .envrc:
```
# export SOPS_AGE_KEY_FILE="$(expand_path "${HOME}/Library/Application Support/sops/age/keys.txt")"
export SOPS_AGE_KEY_FILE="$(expand_path "${HOME}/.config/sops/age/keys.txt")"
export AGE_PUBLIC_KEY="$(grep "public key" "$SOPS_AGE_KEY_FILE" | awk '{ print $4 }')"
```
Then run direnv allow . to refresh the environment.
Create Cloudflare API Token

📍 To use cert-manager with the Cloudflare DNS challenge you will need to create an API token.
1. Create a Cloudflare API Token here.
2. Under the API Tokens section click the blue Create Token button.
3. Click the blue Use template button for the Edit zone DNS template.
4. Name your token something like flux-system
5. Under Permissions, click + Add More and add each permission below:
```
Zone - DNS - Edit
Account - Cloudflare Tunnel - Read
```
6. Limit the permissions to a specific account and zone resources.
7. Fill out the appropriate vars in .env file:
```
CLOUDFLARE_EMAIL=''
CLOUDFLARE_TOKEN=''
CLOUDFLARE_ACCOUNT=''
CLOUDFLARE_TUNNELID=''
CLOUDFLARE_TUNNEL_SECRET=''
```
Create Cloudflare Tunnel

📍 To expose services to the internet you will need to create a Cloudflare Tunnel.
1. Authenticate cloudflared to your domain
```
cloudflared tunnel login
```
2. Create the tunnel
```
cloudflared tunnel create k8s
```
3. Fill out the appropriate Cloudflare Tunnel vars in the .env file: CLOUDFLARE_ACCOUNT, CLOUDFLARE_TUNNELID, CLOUDFLARE_TUNNEL_SECRET
Cloudflare Tunnel info can be found with cat ~/.cloudflared/*.json | jq -r.

⚡ Prepare your nodes for k3s

📍 Here we will be running an Ansible playbook to prepare your nodes for running a Kubernetes cluster.

Ensure you are able to SSH into your nodes from your workstation using a private SSH key without a passphrase (for example using a SSH agent). This lets Ansible interact with your nodes.
Verify Ansible can view your config
```
just ansible hosts
```
Verify Ansible can ping your nodes
```
just ansible ping
```
Run the Ansible prepare playbook (nodes will reboot when done)
```
just ansible prepare
```

🛰️ Build your k3s cluster with Ansible

📍 Here we will be running a Ansible Playbook to install k3s with this Ansible galaxy role.

If you run into problems, you can run just kube nuke to destroy the k3s cluster and start over from this point.

Verify Ansible can view your config
```
just ansible hosts
```
Verify Ansible can ping your nodes
```
just ansible ping
```
Install k3s (may need to run this twice to pass the k3s systemd restart)
```
just kube install
```
The kubeconfig for interacting with your cluster should have been created in the root of your repository.

Verify the nodes are online

kubectl get nodes -o wide
# NAME           STATUS   ROLES                       AGE     VERSION
# k8s-0          Ready    control-plane,etcd,master   1h      v1.27.3+k3s1
# k8s-1          Ready    worker                      1h      v1.27.3+k3s1

Review the pods currently running in the cluster
```
kubectl get pods -A -o wide
```

🔹 Install Flux in your cluster

Verify Flux can be installed

flux check --pre
# ► checking prerequisites
# ✔ kubectl 1.27.3 >=1.18.0-0
# ✔ Kubernetes 1.27.3+k3s1 >=1.16.0-0
# ✔ prerequisites checks passed

Push your changes to git

📍 Verify all the *.sops.yaml files under the ./ansible and ./kubernetes directories are encrypted with SOPS
```
git add -A
git commit -m "Initial commit :rocket:"
git push
```

Install Flux and sync the cluster to the Git repository

just flux bootstrap
# namespace/flux-system configured
# customresourcedefinition.apiextensions.k8s.io/alerts.notification.toolkit.fluxcd.io created
# ...

Verify Flux components are running in the cluster

kubectl -n flux-system get pods -o wide
# NAME                                       READY   STATUS    RESTARTS   AGE
# helm-controller-5bbd94c75-89sb4            1/1     Running   0          1h
# kustomize-controller-7b67b6b77d-nqc67      1/1     Running   0          1h
# notification-controller-7c46575844-k4bvr   1/1     Running   0          1h
# source-controller-7d6875bcb4-zqw9f         1/1     Running   0          1h

☑️ Verification Steps

Output all the common resources in your cluster.

📍 Suggestion: Use the provided just recipes for validation of cluster resources or continue to get familiar with the kubectl and flux CLI tools.
```
just kube resources
```
⚠️ It might take cert-manager a while to generate certificates, this is normal so be patient.

📣 Post installation

🌐 Public DNS

The external-dns application created in the networking namespace will handle creating public DNS records. By default, echo-server and the flux-webhook are the only subdomains reachable from the public internet. In order to make additional applications public you must set the correct ingress class name and ingress annotations like in the HelmRelease for echo-server.

🏠 Home DNS

k8s_gateway will provide DNS resolution to external Kubernetes resources (i.e. points of entry to the cluster) from any device that uses your home DNS server. For this to work, your home DNS server must be configured to forward DNS queries for <cloudflare_domain> to <k8s_gateway_addr> instead of the upstream DNS server(s) it normally uses. This is a form of split DNS (aka split-horizon DNS / conditional forwarding).

[!TIP] Below is how to configure a Pi-hole for split DNS. Other platforms should be similar.
Apply this file on the server
# /etc/dnsmasq.d/99-k8s-gateway-forward.conf
server=/${bootstrap_cloudflare_domain}/${bootstrap_k8s_gateway_addr}
Restart dnsmasq on the server.

Query an internal-only subdomain from your workstation (any internal class ingresses): dig @${home-dns-server-ip} hubble.${bootstrap_cloudflare_domain}. It should resolve to ${bootstrap_internal_ingress_addr}.

If you’re having trouble with DNS be sure to check out these two GitHub discussions: Internal DNS and Pod DNS resolution broken.

… Nothing working? That is expected, this is DNS after all!

📜 Certificates

By default this template will deploy a wildcard certificate using the Let’s Encrypt staging environment, which prevents you from getting rate-limited by the Let’s Encrypt production servers if your cluster doesn’t deploy properly (for example, due to a misconfiguration). Once you are sure you will keep the cluster up for more than a few hours, be sure to switch to the production servers as outlined in config.yaml.

📍 You will need a production certificate to reach internet-exposed applications through cloudflared.

🪝 GitHub Webhook

By default Flux will periodically check your git repository for changes. In order to have Flux reconcile on git push you must configure GitHub to send push events.

Follow FluxCD instructions to generate a token.

Obtain the webhook path

📍 Hook id and path should look like /hook/123abc123abc...

kubectl -n flux-system get receiver github-webhook -o jsonpath='{.status.webhookPath}'

Piece together the full URL with the webhook path appended

https://flux-webhook.${bootstrap_cloudflare_domain}/hook/123abc123abc...

Navigate to the settings of your repository on GitHub, under “Settings/Webhooks” press the “Add webhook” button. Fill in the webhook url and your bootstrap_flux_github_webhook_token secret and save.

🤖 Renovate

Renovate is a tool that automates dependency management. It is designed to scan your repository around the clock and open PRs for out-of-date dependencies it finds. Common dependencies it can discover are Helm charts, container images, GitHub Actions, Ansible roles… even Flux itself! Merging a PR will cause Flux to apply the update to your cluster.

To enable Renovate, click the ‘Configure’ button over at their GitHub app page and select your repository. Renovate creates a “Dependency Dashboard” as an issue in your repository, giving an overview of the status of all updates. The dashboard has interactive checkboxes that let you do things like advance scheduling or reattempt update PRs you closed without merging.

The base Renovate configuration in your repository can be viewed at .github/renovate.json5. By default it is scheduled to be active with PRs every weekend, but you can change the schedule to anything you want, or remove it if you want Renovate to open PRs right away.

🐛 Debugging

Below is a general guide on trying to debug an issue with a resource or application. For example, if a workload/resource is not showing up or a pod has started but is in a CrashLoopBackOff or Pending state.

Start by checking all Flux Kustomizations & Git Repository & OCI Repository and verify they are healthy.
```
flux get sources oci -A
flux get sources git -A
flux get ks -A
```
Then check all the Flux Helm Releases and verify they are healthy.
```
flux get hr -A
```
Then check if the pod is present.
```
kubectl -n <namespace> get pods -o wide
```

Then check the logs of the pod if it’s there.

kubectl -n <namespace> logs <pod-name> -f
# or
stern -n <namespace> <fuzzy-name>

If a resource exists try to describe it to see what problems it might have.
```
kubectl -n <namespace> describe <resource> <name>
```

Check the namespace events

kubectl -n <namespace> get events --sort-by='.metadata.creationTimestamp'

Resolving problems that you have could take some tweaking of your YAML manifests in order to get things working, other times it could be an external factor like permissions on NFS. If you are unable to figure out your problem see the help section below.

Authenticate Flux over SSH

Authenticating Flux to your git repository has a benefits like using a private git repository and/or using the Flux Image Automation Controllers.

By default this template only works on a public GitHub repository, it is advised to keep your repository public.

The benefits of a public repository include:

When debugging or asking for help, you can provide a link to a resource you are having issues with.
Adding a topic to your repository of k8s-at-home to be included in the k8s-at-home-search. This search helps people discover different configurations of Helm charts across other Flux-based repositories.

Expand to read guide on adding Flux SSH authentication (private repos only)

This cluster syncs a public repository over anonymous HTTPS, so no git credentials are needed and no deploy key is applied during bootstrap. Follow this section only if you make your repository private and want Flux to pull it over SSH.

Generate a new SSH key:

ssh-keygen -t ecdsa -b 521 -C "github-deploy-key" -f ./bootstrap/github-deploy.key -q -P ""

Paste the public key into the Deploy keys section of your repository settings.

Create a sops secret at ./bootstrap/github-deploy-key.sops.yaml:

apiVersion: v1
kind: Secret
metadata:
  name: github-deploy-key
  namespace: flux-system
stringData:
  # 3a. Contents of ./bootstrap/github-deploy.key (the private key)
  identity: |
    -----BEGIN OPENSSH ... -----
        ...
    -----END OPENSSH ... -----
  # 3b. Output of: curl --silent https://api.github.com/meta | jq --raw-output '"github.com "+.ssh_keys[]'
  known_hosts: |
    github.com ssh-ed25519 ...
    github.com ecdsa-sha2-nistp256 ...
    github.com ssh-rsa ...

(Alternatively, seed it from 1Password via op inject in bootstrap/resources.yaml.j2, consistent with the rest of this repo.)

Encrypt the secret:

sops --encrypt --in-place ./bootstrap/github-deploy-key.sops.yaml

Re-add the secret to the bootstrap SOPS apply list in scripts/bootstrap/apply_sops_secrets.sh so it is created in flux-system (or apply it manually with sops exec-file ... kubectl apply).

Point the FluxInstance sync at SSH and reference the secret in ./kubernetes/apps/flux-system/flux-instance/app/helmrelease.yaml:

spec:
  values:
    instance:
      sync:
        kind: GitRepository
        # 6a. Change this to your user and repo names
        url: ssh://git@github.com/$user/$repo
        ref: refs/heads/main
        path: kubernetes/flux/cluster
        # 6b. Reference the deploy-key secret
        pullSecret: github-deploy-key

Commit and push changes.

Force flux to reconcile your changes:

flux reconcile -n flux-system kustomization flux-system --with-source

Verify the git repository is now using SSH:
```
flux get sources git -A
```
Optionally set your repository to Private in your repository settings.

🤝 Thanks

This would not be possible without onedr0p and the k8s-at-home community!