Skip to main content

Creating clusters

Step 1: Applying the Cluster Stack#

A defines a complete, opinionated set of Kubernetes resources including cluster templates, node image configurations, and add-on manifests which are needed to create a production-ready cluster. For each release, these resources and configurations are heavily tested by Syself to guarantee integrity and interoperability.

The first step for cluster creation is making a Cluster Stack Release available. Create a file named clusterstack.yaml, with the following content:

clusterstack.yamlyaml
		apiVersion: clusterstack.x-k8s.io/v1alpha1
kind: ClusterStack
metadata:
  name: hetzner-apalla-1-33
spec:
  provider: hetzner
  name: apalla
  kubernetesVersion: "1.33"
  channel: stable
  autoSubscribe: true
  providerRef:
    name: hetzner-apalla-1-33
    kind: HetznerClusterStackReleaseTemplate
    apiVersion: infrastructure.clusterstack.x-k8s.io/v1alpha1
---
apiVersion: infrastructure.clusterstack.x-k8s.io/v1alpha1
kind: HetznerClusterStackReleaseTemplate
metadata:
  name: hetzner-apalla-1-33
spec:
  template:
    spec:
      nodeImages:
        - controlplaneamd64hcloud
        - workeramd64hcloud

And apply it to the management cluster:

		$ kubectl apply -f clusterstack.yaml
clusterstack.clusterstack.x-k8s.io/clusterstack created
hetznerclusterstackreleasetemplate.infrastructure.clusterstack.x-k8s.io/clusterstack created

When a Cluster Stack is applied, a ClusterStackRelease is created from it, and the necessary components starts to be prepared in your Hetzner account. In our case, we applied a Cluster Stack with autoSubscribe set to true, so it'll fetch the latest release and create the ClusterStackRelease object based on that. You don't need to interact directly with the release.

It takes around 5 minutes for the ClusterStackRelease to be ready, as the node images are being built. You can check the status of these node images by running kubectl get HetznerNodeImageReleases. These objects have a one-to-one relationship with the snapshots in your Hetzner project, which you can view under Servers / Snapshots in the Hetzner Console.

Note

If you already have a Cluster Stack applied from a previous cluster, there is no need to apply it again since the HetznerNodeImageReleases and ClusterStackRelease are already there. You can use the same release and node images for any number of clusters.

After the node images are successfully built, the ClusterStackRelease hetzner-apalla-1-33-v13 will be ready for you to use in your cluster. Check with:

		$ kubectl get clusterstackrelease
NAME                      K8S VERSION   READY   AGE   REASON   MESSAGE
hetzner-apalla-1-33-v13   v1.33.13      true    1m

Step 2: Creating your cluster#

Tip

If you need your clusters to be SCS-compatible, you can refer to the page.

The SCS (Sovereign Cloud Stack)⁠ is a European initiative for an open, transparent and vendor-neutral cloud ecosystem that guarantees sovereignty.

Now that we have a ClusterStackRelease ready, we can create a cluster from it. For that, create a cluster.yaml file with the contents below:

cluster.yamlyaml
		apiVersion: cluster.x-k8s.io/v1beta2
kind: Cluster
metadata:
  name: mycluster
spec:
  clusterNetwork:
    services:
      cidrBlocks: ["10.128.0.0/12"]
    pods:
      cidrBlocks: ["192.168.0.0/16"]
    serviceDomain: "cluster.local"
  topology:
    classRef:
      name: hetzner-apalla-1-33-v13
    version: v1.33.13
    controlPlane:
      class: hcloud
      replicas: 3
    workers:
      machineDeployments:
        - class: workeramd64hcloud
          name: md-0
          replicas: 1
          failureDomain: nbg1
          variables:
            overrides:
              - name: workerMachineTypeHcloud
                value: cpx42
    variables:
      - name: region
        value: nbg1
      - name: controlPlaneMachineTypeHcloud
        value: cpx42

And apply it to the management cluster with:

		$ kubectl apply -f cluster.yaml
cluster.cluster.x-k8s.io/mycluster created

Cluster creation will take a few more minutes. You can monitor the process by looking at the machine objects:

		$ kubectl get machines
NAME                                 CLUSTER       NODENAME                          PROVIDERID            PHASE          AGE    VERSION
mycluster-jndgf-r4k7v                mycluster     mycluster-hb2wb-pb6d4             hcloud://12345678     Running        10m    v1.33.13
mycluster-jndgf-sb6b8                mycluster     mycluster-hb2wb-2g9wq             hcloud://87654321     Running        8m     v1.33.13
mycluster-jndgf-vhbsb                mycluster     mycluster-hb2wb-bnkjh             hcloud://12348765     Running        5m     v1.33.13
mycluster-md-0-zw8ln-pxztl-j4stb     mycluster     mycluster-md-0-bp4v2-7r2sv        hcloud://43215678     Running        4m     v1.33.13

These represent actual machines in your Hetzner project. If all of them are in the Running phase, it means your cluster is ready!

Note

If a control plane node replacement or rolling update is stuck provisioning, this is most likely a regional HCloud capacity issue and the operation will remain blocked until capacity returns.

Baremetal control planes are not subject to this, as updates and recovery operate on existing inventory. See for setup instructions.

Step 3: Accessing your cluster#

To get the kubeconfig of your workload cluster, you can use the command:

		$ kubectl get secrets mycluster-kubeconfig -o=jsonpath='{.data.value}' \
| base64 -d \
> mycluster-kubeconfig.yaml

Now, you can change the KUBECONFIG environment variable to point to your new cluster:

		$ export KUBECONFIG=mycluster-kubeconfig.yaml

If you want more information about workload and management clusters, check the section.

When accessing your cluster, you may see some pods still pending. This is normal, you just have to wait for it to be completely initialized. When all pods are running and the four nodes are in ready state, it means your cluster is completely provisioned.

Your cluster is up, managed by Syself Autopilot. Because everything is defined as Kubernetes manifests, you can keep the whole cluster configuration in Git and manage it with GitOps.

Optional: Run sample workload#

Deploy the sample pod to your cluster:

		$ cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
  name: autopilot-sample
spec:
  containers:
    - image: ghcr.io/syself/autopilot-sample:latest
      name: autopilot-sample
      ports:
        - containerPort: 80
EOF
pod/autopilot-sample created

And create a port forwarding to access it:

		$ kubectl port-forward autopilot-sample 8080:80
Forwarding from 127.0.0.1:8080 -> 80
Forwarding from [::1]:8080 -> 80
Handling connection for 8080

Now, go to http://localhost:8080⁠ in your browser.

Further Reading#

In the section you can find tutorials for common tasks you might want to do in your cluster. We recommend you go over it after creating your first cluster.

For a list of all servers supported by Syself Autopilot and recommendations, you can refer to the page.

Hetzner account preparation

Prepare your Hetzner account to run Kubernetes nodes on cloud and bare metal servers with Syself Autopilot.

Management and Workload Clusters

Understand the role of management and workload clusters in Kubernetes and how Syself Autopilot uses Cluster API to provision and manage infrastructure.