Template for deploying a Kubernetes cluster backed by Flux

Welcome to my highly opinionated template for deploying a single Kubernetes (k3s) cluster with Ansible and managing applications with Flux. Upon completion you will be able to expose web applications you choose to the internet with Cloudflare Tunnel.

Overview

👋 Introduction

The following components will be installed in your k3s cluster by default. Most are only included to get a minimum viable cluster up and running.

flux - GitOps operator for managing Kubernetes clusters from a Git repository
kube-vip - Load balancer for the Kubernetes control plane nodes
cert-manager - Operator to request SSL certificates and store them as Kubernetes resources
cilium - Container networking interface for inter pod and service networking
external-dns - Operator to publish DNS records to Cloudflare (and other providers) based on Kubernetes ingresses
k8s_gateway - DNS resolver that provides local DNS to your Kubernetes ingresses
ingress-nginx - Kubernetes ingress controller used for a HTTP reverse proxy of Kubernetes ingresses
local-path-provisioner - provision persistent local storage with Kubernetes

Additional applications can be enabled in the addons configuration file

📝 Prerequisites

Please bring a positive attitude and be ready to learn and fail a lot. The more you fail, the more you can learn from.
This was designed to run in your home network on bare metal machines or VMs NOT in the cloud.
You MUST have a domain you can manage on Cloudflare.
Secrets will be commited to your Git repository AND they will be encrypted by SOPS.
By default your domain name will NOT be visible to the public.
To reach internal-only apps you MUST have a DNS server that supports split DNS (Pi-Hole, Blocky, Dnsmasq, Unbound, etc...) deployed somewhere outside your cluster ON your home network.
In order for this all to work you have to use nodes that have access to the internet. This is not going to work in air-gapped environments.
Only amd64 and/or arm64 nodes are supported.

With that out of the way please continue on if you are still interested...

📚 Reading material

Organizing Cluster Access Using kubeconfig Files

💻 System Preparation

Download Debian 12 (for raspi/arm64 use the tested images)

AMD64

There is a decent guide here on how to get Debian installed.

Deviations from that guide

- Choose "Guided - use entire disk"
- Choose "All files in one partition"
- Delete Swap partition
- Uncheck all Debian desktop environment options
- Keep ssh server checked

[Post install] Remove CD/DVD as apt source

su -
sed -i '/deb cdrom/d' /etc/apt/sources.list
apt update
exit

[Post install] Enable sudo for your non-root user

su -
apt install sudo
usermod -aG sudo ${username}
echo "${username}  ALL=(ALL) NOPASSWD:ALL" | tee /etc/sudoers.d/${username}
exit
newgrp sudo
sudo apt update

[Post install] Add SSH keys (or use ssh-copy-id on the client that is connecting)

mkdir -m 700 ~/.ssh
sudo apt install curl
curl https://github.com/${github_username}.keys > ~/.ssh/authorized_keys
chmod 600 ~/.ssh/authorized_keys

Raspberry Pi / ARM64

If you choose to use a Raspberry Pi for the cluster, it is recommended to have at minimum a Raspberry Pi4 (4GB) and preferably an 8GB model. Additionally, it is also recommended to boot from an external SSD, rather than the SD card. This is supported natively, however if you have an early Raspberry Pi4, you may need to update the bootloader.

According to the documentation here, after you have flashed the image onto a SSD/NVMe you must mount the drive and do the following.

Edit the sysconf.txt
Add/change root_authorized_key to your desired public SSH key
Add/change root_pw to your desired root password
Add/change hostname to your desired hostname

🚀 Lets go

Very first step will be to create a new public repository by clicking the big green Use this template button on this page.

Clone your new repo to you local workstation and cd into it.

📍 All of the below commands are run on your local workstation, not on any of your cluster nodes.

🔧 Workstation Tools

📍 Install the most recent version of the CLI tools below. If you are having trouble with future steps, it is very likely you don't have the most recent version of these CLI tools. The most troublesome are ansible, go-task, and sops.

Install the following CLI tools on your workstation, if you are using Homebrew skip this step and move onto 2 & 3.
- Required: age, ansible, flux, cloudflared, go-task, direnv, kubectl, python3, python-pip3, sops
- Recommended: helm, kustomize, stern
[Homebrew] Install go-task
```
brew install go-task/tap/go-task
```
[Homebrew] Install the other workstation dependencies
```
task brew:deps
```

🔐 Setting up Age

📍 Here we will create a Age Private and Public key. Using SOPS with Age allows us to encrypt secrets and use them in Ansible and Flux.

Create a Age Private / Public Key
```
age-keygen -o age.agekey
```

Set up the directory for the Age key and move the Age file to it

mkdir -p ~/.config/sops/age
mv age.agekey ~/.config/sops/age/keys.txt

Export the SOPS_AGE_KEY_FILE variable in your bashrc, zshrc or config.fish and source it, e.g.
```
export SOPS_AGE_KEY_FILE=~/.config/sops/age/keys.txt
source ~/.bashrc
```
Fill out the Age public key in the appropriate variable in configuration section below, note the public key should start with age...

☁️ Cloudflare API Token

In order to use cert-manager with the Cloudflare DNS challenge you will need to create a API Token.

Head over to Cloudflare and create a API Token by going here.
Under the API Tokens section, click the blue "Create Token" button.
Click the "Use template" blue button for the Edit zone DNS template.
Give your token a name like home-k8s-cluster
Under Permissions, click + Add More and add each permission below:

Zone - DNS - Edit # should be there already if using the template from the previous step
Account - Cloudflare Tunnel - Read

📍 Feel free to limit the permissions to a specific account and zone resources.

Use the API Token in the appropriate variable in configuration section below.

☁️ Cloudflare Tunnel

In order to expose services to the internet you will need to create a Cloudflare Tunnel.

Authenticate cloudflared to your domain
```
cloudflared tunnel login
```
Create the tunnel
```
cloudflared tunnel create k8s
```
In the ~/.cloudflared directory there will be a json file with details you need to populate in configuration section below. You can ignore the cert.pem file.

📄 Configuration

📍 The template/vars/config.yaml file contains necessary configuration that is needed by Ansible and Flux. The template/vars/addons.yaml file allows you to customize which additional apps you want deployed in your cluster. These files are added to the .gitignore file and will not be tracked by Git.

Copy the configuration and addons files and start filling out all the variables.
```
task init
```
Once done run the following command which will verify and generate all the files needed to continue.
```
task configure
```

📂 Repository structure

The configure script will have created the following directories under ./kubernetes.

📁 kubernetes      # Kubernetes cluster defined as code
├─📁 bootstrap     # Flux installation (not tracked by Flux)
├─📁 flux          # Main Flux configuration of repository
└─📁 apps          # Apps deployed into the cluster grouped by namespace

⚡ Preparing Debian Server with Ansible

📍 Here we will be running an Ansible Playbook to prepare Debian server for running a Kubernetes cluster.

📍 Nodes are not security hardened by default, you can do this with dev-sec/ansible-collection-hardening or similar if supported. This is an advanced configuration and generally not recommended unless you want to DevSecOps your cluster and nodes.

Ensure you are able to SSH into your nodes from your workstation using a private SSH key without a passphrase. This is how Ansible is able to connect to your remote nodes.
Install the Ansible deps
```
task ansible:deps
```
Verify Ansible can view your config
```
task ansible:list
```
Verify Ansible can ping your nodes
```
task ansible:ping
```
Run the Ansible prepare playbook
```
task ansible:prepare
```
Reboot the nodes (if not done in step 5)
```
task ansible:force-reboot
```

⛵ Installing k3s with Ansible

📍 Here we will be running a Ansible Playbook to install k3s with this wonderful k3s Ansible galaxy role. After completion, Ansible will drop a kubeconfig in ./kubeconfig for use with interacting with your cluster with kubectl.

☢️ If you run into problems, you can run task ansible:nuke to destroy the k3s cluster and start over.

Verify Ansible can view your config
```
task ansible:list
```
Verify Ansible can ping your nodes
```
task ansible:ping
```
Install k3s with Ansible
```
task ansible:install
```

Verify the nodes are online

task cluster:nodes
# NAME           STATUS   ROLES                       AGE     VERSION
# k8s-0          Ready    control-plane,etcd,master   1h      v1.27.3+k3s1
# k8s-1          Ready    control-plane,etcd,master   1h      v1.27.3+k3s1
# k8s-2          Ready    control-plane,etcd,master   1h      v1.27.3+k3s1
# k8s-3          Ready    worker                      1h      v1.27.3+k3s1

🔹 GitOps with Flux

📍 Here we will be installing flux after some quick bootstrap steps.

Verify Flux can be installed

task cluster:verify
# ► checking prerequisites
# ✔ kubectl 1.27.3 >=1.18.0-0
# ✔ Kubernetes 1.27.3+k3s1 >=1.16.0-0
# ✔ prerequisites checks passed

Push you changes to git

📍 Verify all the *.sops.yaml and *.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

task cluster:install
# namespace/flux-system configured
# customresourcedefinition.apiextensions.k8s.io/alerts.notification.toolkit.fluxcd.io created

Verify Flux components are running in the cluster

task cluster:pods -- -n flux-system
# 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

Mic check, 1, 2 - In a few moments applications should be lighting up like a Christmas tree 🎄

You are able to run all the commands below with one task

task cluster:resources

View the Flux Git Repositories
```
task cluster:gitrepositories
```
View the Flux kustomizations
```
task cluster:kustomizations
```
View all the Flux Helm Releases
```
task cluster:helmreleases
```
View all the Flux Helm Repositories
```
task cluster:helmrepositories
```
View all the Pods
```
task cluster:pods
```
View all the certificates and certificate requests
```
task cluster:certificates
```
View all the ingresses
```
task cluster:ingresses
```

🏆 Congratulations if all goes smooth you'll have a Kubernetes cluster managed by Flux, your Git repository is driving the state of your cluster.

☢️ If you run into problems, you can run task ansible:nuke to destroy the k3s cluster and start over.

🧠 Now it's time to pause and go get some coffee ☕ because next is describing additional things like how DNS is handled.

📣 Post installation

🌱 Environment

direnv will make it so anytime you cd to your repo's directory it export the required environment variables (e.g. KUBECONFIG). To set this up make sure you hook it into your shell and after that is done, run direnv allow while in your repos directory.

🌐 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 public sub-domains exposed. In order to make additional applications public you must set an ingress annotation (external-dns.alpha.kubernetes.io/target) like done in the HelmRelease for echo-server.

For split DNS to work it is required to have ${SECRET_DOMAIN} point to the ${K8S_GATEWAY_ADDR} load balancer IP address on your home DNS server. This will ensure DNS requests for ${SECRET_DOMAIN} will only get routed to your k8s_gateway service thus providing internal DNS resolution to your cluster applications/ingresses from any device that uses your home DNS server.

For and example with Pi-Hole apply the following file and restart dnsmasq:

# /etc/dnsmasq.d/99-k8s-gateway-forward.conf
server=/${SECRET_DOMAIN}/${K8S_GATEWAY_ADDR}

Now try to resolve an internal-only domain with dig @${pi-hole-ip} hajimari.${SECRET_DOMAIN} it should resolve to your ${INGRESS_NGINX_ADDR} IP.

If having trouble you can ask for help in this Github discussion.

If nothing is working, that is expected. This is DNS after all!

📜 Certificates

By default this template will deploy a wildcard certificate with the Let's Encrypt staging servers. This is to prevent you from getting rate-limited on configuration that might not be valid on bootstrap using the production server. Once you have confirmed the certificate is created and valid, make sure to switch to the Let's Encrypt production servers as outlined below. Do not enable the production certificate until you are sure you will keep the cluster up for more than a few hours.

Update the resources to use the production Let's Encrypt server:

diff --git a/kubernetes/base/apps/networking/ingress-nginx/app/helmrelease.yaml b/kubernetes/base/apps/networking/ingress-nginx/app/helmrelease.yaml
index e582d4a..0f80700 100644
--- a/kubernetes/base/apps/networking/ingress-nginx/app/helmrelease.yaml
+++ b/kubernetes/base/apps/networking/ingress-nginx/app/helmrelease.yaml
@@ -60,7 +60,7 @@ spec:
           namespaceSelector:
             any: true
       extraArgs:
-        default-ssl-certificate: "networking/${SECRET_DOMAIN/./-}-staging-tls"
+        default-ssl-certificate: "networking/${SECRET_DOMAIN/./-}-production-tls"
       resources:
         requests:
           cpu: 10m
diff --git a/kubernetes/base/apps/networking/ingress-nginx/certificates/kustomization.yaml b/kubernetes/base/apps/networking/ingress-nginx/certificates/kustomization.yaml
index d57147d..f58e4a7 100644
--- a/kubernetes/base/apps/networking/ingress-nginx/certificates/kustomization.yaml
+++ b/kubernetes/base/apps/networking/ingress-nginx/certificates/kustomization.yaml
@@ -3,4 +3,4 @@ apiVersion: kustomize.config.k8s.io/v1beta1
 kind: Kustomization
 resources:
   - ./staging.yaml
-  # - ./production.yaml
+  - ./production.yaml

Commit and push your changes:

git add -A
git commit -m "fix: use production LE certificates"
git push

Force Flux to pick up the changes:
```
task cluster:reconcile
```

To view the certificate request run kubectl -n networking get certificaterequests
To verify the certificate is created run kubectl -n networking get certificates

You should start to see your applications using the new certificate.

🤖 Renovatebot

Renovatebot will scan your repository and offer PRs when it finds dependencies out of date. Common dependencies it will discover and update are Flux, Ansible Galaxy Roles, Terraform Providers, Kubernetes Helm Charts, Kubernetes Container Images, and more!

The base Renovate configuration provided in your repository can be view at .github/renovate.json5. If you notice this only runs on weekends and you can change the schedule to anything you want or simply remove it.

To enable Renovate on your repository, click the 'Configure' button over at their Github app page and choose your repository. Over time Renovate will create PRs for out-of-date dependencies it finds. Any merged PRs that are in the kubernetes directory Flux will deploy.

🪝 Github Webhook

Flux is pull-based by design meaning it will periodically check your git repository for changes, using a webhook you can enable Flux to update your cluster on git push. In order to configure Github to send push events from your repository to the Flux webhook receiver you will need two things:

Webhook URL - Your webhook receiver will be deployed on https://flux-webhook.${bootstrap_cloudflare_domain}/hook/:hookId. In order to find out your hook id you can run the following command:

kubectl -n flux-system get receiver/github-receiver
# NAME              AGE    READY   STATUS
# github-receiver   6h8m   True    Receiver initialized with URL: /hook/12ebd1e363c641dc3c2e430ecf3cee2b3c7a5ac9e1234506f6f5f3ce1230e123

So if my domain was onedr0p.com the full url would look like this:

https://flux-webhook.onedr0p.com/hook/12ebd1e363c641dc3c2e430ecf3cee2b3c7a5ac9e1234506f6f5f3ce1230e123

Webhook secret - Your webhook secret can be found by decrypting the secret.sops.yaml using the following command:
```
sops -d ./kubernetes/base/apps/flux-system/addons/webhooks/github/secret.sops.yaml | yq .stringData.token
```
Note: Don't forget to update the bootstrap_flux_github_webhook_token variable in the config.yaml file so it matches the generated secret if applicable

Now that you have the webhook url and secret, it's time to set everything up on the Github repository side. Navigate to the settings of your repository on Github, under "Settings/Webhooks" press the "Add webhook" button. Fill in the webhook url and your secret.

💾 Storage

Rancher's local-path-provisioner is a great start for storage but soon you might find you need more features like replicated block storage, or to connect to a NFS/SMB/iSCSI server. Check out the projects below to read up more on some storage solutions that might work for you.

🔏 Authenticate Flux over SSH

Authenticating Flux to your git repository has a couple 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:

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 others Flux based repositories.

Expand to read guide on adding Flux SSH authentication

Generate new SSH key:

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

Paste public key in the deploy keys section of your repository settings

Create sops secret in ./kubernetes/bootstrap/github-deploy-key.sops.yaml with the contents of:

apiVersion: v1
kind: Secret
metadata:
  name: github-deploy-key
  namespace: flux-system
stringData:
  # 3a. Contents of github-deploy-key
  identity: |
    -----BEGIN OPENSSH PRIVATE KEY-----
        ...
    -----END OPENSSH PRIVATE KEY-----
  # 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 ...

Encrypt secret:

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

Apply secret to cluster:

sops --decrypt ./kubernetes/bootstrap/github-deploy-key.sops.yaml | kubectl apply -f -

Update ./kubernetes/flux/config/cluster.yaml:

apiVersion: source.toolkit.fluxcd.io/v1beta2
kind: GitRepository
metadata:
  name: home-kubernetes
  namespace: flux-system
spec:
  interval: 10m
  # 6a: Change this to your user and repo names
  url: ssh://git@github.com/$user/$repo
  ref:
    branch: main
  secretRef:
    name: github-deploy-key

Commit and push changes
Force flux to reconcile your changes
```
task cluster:reconcile
```
Verify git repository is now using SSH:
```
task cluster:gitrepositories
```
Optionally set your repository to Private in your repository settings.

💨 Kubernetes Dashboard

Included in your cluster is the Kubernetes Dashboard. Inorder to log into this you will have to get the secret token from the cluster using the command below.

kubectl -n monitoring get secret kubernetes-dashboard -o jsonpath='{.data.token}' | base64 -d

You should be able to access the dashboard at https://kubernetes.${SECRET_DOMAIN}

🐛 Debugging

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

kubectl -n <namespace> get pods

Then check the logs of the pod if its there.

kubectl -n <namespace> logs <pod-name> -f

Note: If a resource exists, running kubectl -n <namespace> describe <resource> <name> might give you insight into what the problem(s) could be.

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

👉 Help

Make a post in this repository's GitHub Discussions.
Start a thread in the support or flux-cluster-template channel in the k8s@home Discord server.

❔ What's next

The world is your cluster, have at it!

🤝 Thanks

Big shout out to all the authors and contributors to the projects that we are using in this repository.

@whazor created this website as a creative way to search Helm Releases across GitHub. You may use it as a means to get ideas on how to configure an applications' Helm values.

denis-ev / k8s-home