This guide walks you through the end-to-end process of creating and verifying an Edge Node Cluster using the ZEDEDA platform UI. Follow the steps to ensure a successful deployment.

An Edge Node Cluster groups three physical edge nodes into a high-availability compute unit. All three nodes are equal cluster members with full functionality — there is no controller/worker distinction between them.

The underlying cluster infrastructure (k3s, networking, storage) is fully managed by the ZEDEDA platform. Customers interact only through the ZEDEDA UI workflow — no manual Kubernetes configuration or deployment is required. If one node is designated a Witness (tie-breaker), apply the tag tie-breaker = true to that node before clustering.

<WRAP center round important 80%> ⚠️ Important — Wait Before Clustering
After onboarding the 3 nodes, allow them to run for 8–10 minutes before creating the cluster. This ensures initial setup tasks complete and prevents cluster formation errors. </WRAP>

Ensure all of the following are in place before proceeding:

Follow these steps in the ZEDEDA UI to create the cluster:

• Navigate to the side menu and select Edge Node Clusters:

• Click the + button in the top-right corner of the screen to begin creating a new cluster:

Fill in the cluster information form:

• Name — Enter a unique name for the cluster.
• Title — Enter a descriptive title.
• Project — Select the correct project from the dropdown.
• Description — (Optional) Enter a description.
• Cluster Tags — Leave at default unless a Witness node is required.

Click Next to continue.

• On the Edge Node Selection screen, check the boxes next to the 3 nodes you want to include in the cluster.
• Available nodes will appear in the list below the selection area.

Click Next to continue.

• On the Cluster Network screen, configure the network interface for each node:
• Expand each member node to reveal its network adapters.
• Select the correct Network Adapter to be used for cluster communication:

Click Next to continue.

• On the Review & Create screen, verify all selected settings.
• Enter the Cluster Prefix — this is the IP prefix used for inter-node cluster communications (e.g., 10.200.0.0/24):

Click Create to provision the cluster.

After creation, navigate back to Edge Node Clusters and select the cluster you just created. The Cluster Run State indicator will cycle through the following stages:

<WRAP center round tip 80%> ⏱ Timing
This process typically takes 3–5 minutes for all nodes to cluster and reach the Online state. </WRAP>

<WRAP center round info 80%> ℹ️ Fully Managed Service
The cluster setup and k3s configuration is handled entirely by the ZEDEDA platform — no manual intervention is needed. SSH access is provided here as an optional way to observe the cluster formation process in real time. </WRAP>

To watch the cluster come online via live log output, SSH into an edge node and tail the install log.

Follow the SSH enable procedure to enable SSH access and add your public key to the edge node, then connect.

Once connected to the EVE shell, run:

eve enter kube

This drops you into the cluster management control plane shell.

Run the following command to stream the cluster setup log in real time:

tail -f /var/log/k3s-install.log

You should see output similar to the following as the cluster forms:

2026-05-07 15:42:01 : Reapply Multus CNI for clusternodeip: 10.200.0.3
2026-05-07 15:42:01 : Cluster Node IP prefix to multus: 10.200.0.0/24 with node-ip 10.200.0.3
2026-05-07 15:42:03 : provision config file for node to cluster mode
2026-05-07 15:42:03 : config.yaml configured with Join-ServerIP 10.200.0.1
2026-05-07 15:42:03 : Check if the Endpoint https://10.200.0.1:6443 is in cluster mode...
2026-05-07 15:42:08 : Cluster transition in progress, waiting before starting k3s

Once the cluster reaches online state, the ZEDEDA platform will have automatically configured and verified all required cluster components. If you are monitoring via SSH, the log will confirm each component is ready:

Component: multus    ready on existing version
Component: kubevirt  ready on existing version
Component: cdi       ready on existing version
Component: longhorn  ready on existing version

End of Document