Skip to main content

Adding nodes to the cluster

With the Kubernetes control node set up, you can now add your remaining dedicated servers to your cluster. Before proceeding with this section, you should:

  • Have set up the Kubernetes control node by following the instructions in Creating your cluster
  • Have one or more additional dedicated servers at the same local provider
  • Have secured those additional dedicated servers by following the instructions in Securing your network

For each additional dedicated server, run through the steps below.

Set the hostname on the machine​

Ensure the dedicated server has it's hostname set to something standardized for your cluster (e.g. west-europe-02). To set the hostname on the dedicated server, make sure you're the root user (sudo bash if you aren't), then run:

hostname west-europe-02
echo "west-europe-02" > /etc/hostname
echo "127.0.1.1 west-europe-02" >> /etc/hosts

Each hostname must be unique in the cluster.

Fix up DNS resolution​

By default, Ubuntu uses systemd's DNS resolver. This conflicts with Kubernetes, so we need to turn it off.

On the dedicated server, make sure you're the root user (sudo bash if you aren't), then run:

systemctl stop systemd-resolved
systemctl disable systemd-resolved
rm /etc/resolv.conf
echo "nameserver 1.1.1.1 1.0.0.1" > /etc/resolv.conf

Get the K3S node token​

On the Kubernetes control node, where you originally installed K3S, you'll need to retrieve the node token. This allows Kubernetes worker nodes to register themselves with the control node when the K3S agent starts.

Retrieve the node token by running cat /var/lib/rancher/k3s/server/node-token; you should see output similar to the following:

root@west-europe-01:/home/jrhodes# cat /var/lib/rancher/k3s/server/node-token
K10feefb.....9b324fdef746675afced214e07cf6

Install the K3S agent​

On the dedicated server that you want to connect to the existing cluster, install the K3S agent by running the following command. Make sure you set the following values, replacing the examples provided:

  • INSTALL_K3S_VERSION: This should match the version when you installed K3S on the control node.
  • K3S_URL: This should be https://...:6443, replacing ... with the private IP address of the control node.
  • K3S_TOKEN: This should be set to the token you copied from the control node.
curl -sfL https://get.k3s.io | INSTALL_K3S_SKIP_START=true INSTALL_K3S_VERSION=v1.22.7+k3s1 K3S_URL=https://10.0.0.1:6443 K3S_TOKEN=K10...cf6 sh -

Like the control node, we don't start the K3S agent immediately as we need to update it's configuration.

Determine your network interfaces​

Follow the instructions on Securing your network to identify the private and public interfaces on this dedicated server.

In the examples below, eth0 is the private network interface, and eth1 is the public network interface. 10.0.0.2 is the private IP address and 126.118.159.101 is the public IP address.

Override the command-line arguments for K3S​

We need to tell K3S agent where it should listen for internal connections and where it should listen for public connections.

On the dedicated server, make sure you're the root user (sudo bash if you aren't), then run:

mkdir /etc/systemd/system/k3s-agent.service.d || true
nano /etc/systemd/system/k3s-agent.service.d/override.conf

Using the text editor, add the following contents to that file:

[Service]
ExecStart=
ExecStart=/usr/local/bin/k3s \
    agent \
    --node-name west-europe-02 \
    --node-ip 10.0.0.2 \
    --node-external-ip 126.118.159.101 \
    --flannel-iface eth0

Then run:

systemctl daemon-reload

Check that the worker node has been registered successfully​

On the control node, run k3s kubectl get nodes. You should see output similar to the following:

root@west-europe-01:/home/jrhodes# k3s kubectl get nodes
NAME             STATUS   ROLES                  AGE   VERSION
west-europe-01   Ready    control-plane,master   63m   v1.22.7+k3s1
west-europe-02   Ready    <none>                 48s   v1.22.7+k3s1

If you see the dedicated server in the list, it has been added to the cluster successfully.