Task 4: Verifying worker node addition

This guide outlines the verification procedure to confirm that your worker nodes have successfully joined the K3s cluster. After a worker node runs the join script, it's crucial to perform checks from the control-plane node to ensure the node is properly integrated, healthy, and visible to the Kubernetes scheduler.

Before you start

It is essential that your team is prepared for this verification process.

Make sure that:

Task 3 (Addition) has been successfully completed.
All users are in the shared terminal session.

How to validate node addition

The designated Driver will now execute commands, oservers should watch the shared screen and verify the output of each command before proceeding to the next step.

Step1: Check the Agent Service Status (From the worker node)

sudo systemctl status k3s-agent.service

Expected output:

● k3s-agent.service - Lightweight Kubernetes
     Loaded: loaded (/etc/systemd/system/k3s-agent.service; enabled; vendor preset: enabled)
     Active: active (running) since Thu 2025-06-19 14:27:59 CDT; 57min ago
       Docs: https://k3s.io
    Process: 1034785 ExecStartPre=/bin/sh -xc ! /usr/bin/systemctl is-enabled --quiet nm-cloud-setup.service 2>/dev/null (code=exited, status=0/SUCCESS)
    Process: 1034787 ExecStartPre=/sbin/modprobe br_netfilter (code=exited, status=0/SUCCESS)
    Process: 1034788 ExecStartPre=/sbin/modprobe overlay (code=exited, status=0/SUCCESS)
   Main PID: 1034789 (k3s-agent)
      Tasks: 17
     Memory: 277.7M
     CGroup: /system.slice/k3s-agent.service
             └─1034789 /usr/local/bin/k3s agent

To exit: q.

Step2: Switch to control plane

2.1 Exit out of the shared terminal

exit

2.2 Exit out of the SSH session to the AGX node and return to the terminal on your computer.

exit

2.3 Login to the control plane node

ssh <username>@<control plane ip>

2.4 Switch to the shared user

sudo su - k3s

2.5 Start a shared terminal session.

screen -x group_session

Step 3: Validate node addition from the control plane

kubectl get nodes -o wide

Expected output:

NAME      STATUS   ROLES                  AGE     VERSION        INTERNAL-IP   EXTERNAL-IP   OS-IMAGE             KERNEL-VERSION   CONTAINER-RUNTIME
agx       Ready    <none>                 67m     v1.30.6+k3s1   10.10.50.22   <none>        Ubuntu 20.04.6 LTS   5.10.192-tegra   docker://24.0.7
k3smain   Ready    control-plane,master   4h56m   v1.30.6+k3s1   10.20.1.150   <none>        Ubuntu 24.04.2 LTS   6.8.0-1029-aws   containerd://1.7.22-k3s1

Step 4: Assign the newly added node to be a worker.

While the node is already functional and can accept workloads, applying a standard label is a best practice for organization and for tools that might look for this specific role.

kubectl label node agx node-role.kubernetes.io/worker=worker

Expected response:

node/agx labeled

Step 5: Validate tag assignment.

kubectl get nodes -o wide

Expected output:

NAME      STATUS   ROLES                  AGE    VERSION        INTERNAL-IP   EXTERNAL-IP   OS-IMAGE             KERNEL-VERSION   CONTAINER-RUNTIME
agx       Ready    worker                 76m    v1.30.6+k3s1   <AGX IP>      <none>        Ubuntu 20.04.6 LTS   5.10.192-tegra   docker://24.0.7
k3smain   Ready    control-plane,master   5h5m   v1.30.6+k3s1   <CP ip>       <none>        Ubuntu 24.04.2 LTS   6.8.0-1029-aws   containerd://1.7.22-k3s1

What to look for:

The output confirms that the agx node now has the worker role and the correct INTERNAL-IP.

Most importantly, verify that both nodes show a STATUS of Ready, which confirms your multi-node cluster is healthy and fully operational.

Accomplishments

Local Health Verified

The health of the K3s agent was confirmed directly on the worker node, ensuring the systemd service was active (running).

Successful Cluster Registration

The new worker node (agx) was verified from the control plane, confirming it successfully joined the cluster and achieved a Ready status.

Worker Role Assigned

The agx node was formally designated as a worker by applying the standard node-role.kubernetes.io/worker label, making its purpose clear for scheduling and organization.

Multi-Node Cluster Achieved

The single-node instance has been successfully expanded into a healthy, multi-node cluster with a dedicated control plane and a fully integrated worker node ready to host applications.

With the successful addition and verification of the agx worker node, your multi-node cluster is fully validated and operational.

Reflection

Please take a moment to write down any questions, issues, or doubts you encountered during this milestone.
This will help guide the subsequent discussion and ensure everyone is on the same page before moving forward.

Next Steps

exit (ctrl + a then d) from the screen session.
exit from the ssh session on the AGX Node.
Return to the terminal on your computer.
Retain the node token, it will be used in the next step.
Rejoin the common Discord lobby to await further instructions or support your peers.

Last modified: 23 June 2025