k8s-hetzner-cloud

Kubernetes Terraform installer for Hetzner bare-metal.

Initial setup

Clone the repository and install the dependencies:

$ git clone https://github.com/felipecruz91/k8s-hetzner.git
$ cd k8s-hetzner
$ terraform init

Note that you'll need Terraform v0.13 or newer to run this project.

Note that, before running the project, you'll have to create an access token for Terraform to connect to the Hetzner Cloud API.

Once you have the token, create the hetzner-api-token.txt file and paste it in there.

Using the token, create the TF_VAR_hcloud_token environment variable:

$ export TF_VAR_hcloud_token=$(cat hetzner-api-token.txt)

Building the Kubernetes VM image with Packer

Prepare snapshot image, so the terraform will use it to create and up nodes of the cluster.

Run once the command below:

$ cd images && \
    packer build image-master.json

For containerd image use command packer build image-master-containerd.json.

Now we have a snapshot and we need to know it's ID.

To do this run:

$ curl -H "Authorization: Bearer $TF_VAR_hcloud_token" 'https://api.hetzner.cloud/v1/images'

and find the image's ID with the name described in the image-master-containerd.json and/or image-master.json files. Put pointed ID in the variables.tf config file.

Usage

Create a bare-metal Kubernetes cluster with one master and 3 worker nodes:

$ terraform apply \
 -var master_location=nbg1 \
 -var master_server_type=cpx11 \
 -var worker_nodes_count=3 \
 -var worker_server_type=cx11 \
 -var worker_location=nbg1

This will do the following:

• provisions three bare-metal servers with Ubuntu 18.04 LTS (the size of the master and the node may be different but must remain in the same type of architecture)
• connects to the master server via SSH and installs Docker CE and kubeadm apt packages
• runs kubeadm init on the master server and configures kubectl
• downloads the kubectl admin config file on your local machine and replaces the private IP with the public one
• (obsolete) creates a Kubernetes secret with the Weave Net password
• (obsolete) installs Weave Net with encrypted overlay
• installs calico CNI
• starts the nodes in parallel and installs Docker CE and kubeadm
• joins the nodes in the cluster using the kubeadm token obtained from the master

Scale up by increasing the number of nodes:

$ terraform apply \
 -var worker_nodes_count=4

There are two useful scripts which can be used to create and destroy a cluster:

upcl.sh
docl.sh

Tear down the whole infrastructure with:

terraform destroy -force

Remote control

After applying the Terraform plan you'll see several output variables like the master public IP, the kubeadm join command and the current workspace admin config.

In order to run kubectl commands against the Hetzner cluster you can use the kubectl_config output variable:

$ export KUBECONFIG="$(pwd)/$(terraform output --raw kubectl_config)"

$ kubectl --insecure-skip-tls-verify get nodes -o wide

NAME       STATUS   ROLES    AGE     VERSION   INTERNAL-IP     EXTERNAL-IP   OS-IMAGE             KERNEL-VERSION       CONTAINER-RUNTIME
master     Ready    master   3m45s   v1.24.5   78.47.249.188   <none>        Ubuntu 18.04.5 LTS   4.15.0-124-generic   docker://19.3.6
worker-0   Ready    <none>   3m28s   v1.24.5   94.130.73.91    <none>        Ubuntu 18.04.5 LTS   4.15.0-124-generic   docker://19.3.6
worker-1   Ready    <none>   3m23s   v1.24.5   157.90.17.247   <none>        Ubuntu 18.04.5 LTS   4.15.0-124-generic   docker://19.3.6
worker-2   Ready    <none>   3m26s   v1.24.5   157.90.24.227   <none>        Ubuntu 18.04.5 LTS   4.15.0-124-generic   docker://19.3.6