OpenShift Cluster

Overview

Armada Bridge supports Red Hat OpenShift as a Kubernetes distribution alongside upstream Kubernetes. OpenShift clusters are created and managed through the same cluster workflow, with OpenShift-specific configuration — including an OpenShift version selector, base domain, and a Red Hat pull secret for image registry access.

This page covers:

• Creating an OpenShift cluster
• Monitoring cluster creation and accessing the cluster
• Onboarding an existing external OpenShift cluster

Prerequisites

• Tenant Admin access — Log in as a Tenant Admin to create or onboard clusters.
• Compute resources — Bare Metal resources allocated to your tenant.
• Red Hat Pull Secret — Required to pull OpenShift images from the Red Hat registry. Obtain your pull secret from Red Hat Cloud Console.
• Port-forward on Bridge node — For cluster creation to succeed, port-forward must be running on the Bridge-deployed node.

Port-forward for cluster creation

Run the port forwarding on the Bridge node so that cluster creation can complete.

kubectl -n amcop-system port-forward --address 0.0.0.0 svc/ingress-ingress-nginx-controller 443:443

Keep this command running during cluster creation. If you do not have access to the Bridge node, contact your Bridge Super Administrator.

Create an OpenShift Cluster

Step 1: Start Cluster Creation

1. Log in to Armada Bridge as a Tenant Admin.
2. In the left sidebar, open Compute → Cluster.
3. Click Create Cluster.

Step 2: Configure Cluster Setup

1. Set Type to OpenShift from the dropdown.
2. Enter a Cluster Name (e.g., nac-openshift-cluster).
3. Optionally enter a Description.
4. Paste your Pull Secret — the JSON-format authentication token from your Red Hat account. This is required for OpenShift to pull images from the Red Hat registry.
5. Click Next.

Pull Secret format

The pull secret is a JSON object obtained from the Red Hat Hybrid Cloud Console. It contains authentication credentials for registry.redhat.io, quay.io, and other Red Hat registries.

Step 3: Configure OpenShift Version and Domain

1. Select the OpenShift Version from the dropdown (supported: 4.18).
2. Enter the Base Domain for the cluster (e.g., nac-cluster.openshift.clusters). This is used to generate the cluster's API and console endpoints.
3. Toggle Install NVIDIA GPU tools on if you want NVIDIA GPU tooling installed on the cluster.
4. Click Next.

Step 4: Select a Cluster Template

Select the template that matches your workload requirements. The same templates available for standard Kubernetes are supported for OpenShift:

Template	Use case
None	Plain OpenShift cluster with no additional stack.
Ray-Enabled	OpenShift cluster with KubeRay for distributed computing workloads.
JupyterHub with KAI Scheduler	JupyterHub and KAI Scheduler for interactive notebooks and job scheduling.
NVIDIA NIM	NVIDIA NIM for high-performance GPU-accelerated inference.

Click Next after selecting a template.

Step 5: Select Nodes

1. Select the Bare Metal nodes to assign to the cluster from the available node pool.
2. Click Create to start cluster provisioning.

Monitor Cluster Creation

Cluster creation progresses through the following states. Wait until the status shows Running.

Status	Description
Processing	Cluster provisioning has started.
Initializing: Control Plane	OpenShift control plane is being set up.
Running	Cluster is fully provisioned and ready for workloads.

Access the Cluster

Once the cluster is Running, click the cluster name to open the cluster detail view.

Overview Tab

The Overview tab shows cluster details and available actions.

Field	Description
Type	OpenShift
Status	Running
CNI Plugin	Network plugin used by the cluster
Nodes	Number of nodes allocated
GPUs	GPU resources assigned to the cluster

Available actions:

• Pause Cluster — Suspend the cluster without deleting it.
• Kubeconfig — Download the kubeconfig file to access the cluster from your local machine or external tools.
• Connect — Open the cluster console.
• Delete — Permanently remove the cluster.

Nodes Tab

The Nodes tab lists all nodes assigned to the cluster — hostname, IP address, architecture, and type (BareMetal).

Kubectl Terminal

Click the Kubectl Terminal tab to open an in-browser terminal connected to the OpenShift cluster. You can run kubectl or oc commands directly from the Bridge UI without needing a separate terminal session.

To verify the cluster is healthy, run:

kubectl get pods -A

OpenShift-specific system pods you will see include:

• openshift-ovn-kubernetes — OVN-Kubernetes CNI networking
• openshift-operator-lifecycle-manager — Operator Lifecycle Manager
• openshift-network-operator — Network operator
• openshift-route-controller — Route controller manager
• openshift-service-ca — Service CA operator

note

Ensure all pods are in Running state before deploying workloads.

Onboard an External OpenShift Cluster

You can onboard an existing OpenShift cluster managed outside of Armada Bridge and bring it under Bridge management.

Step 1: Open Onboard Dialog

1. In the left sidebar, open Compute → Cluster.
2. Click Onboard External Cluster.

Step 2: Configure the External Cluster

1. Set Cluster type to OpenShift.
2. Enter a Cluster Name (e.g., ocp-external-cluster).
3. Optionally enter a Description.
4. Upload the K8s Config (kubeconfig file) for the external cluster.
5. Click Onboard cluster.

Step 3: Monitor Onboarding

The cluster appears in the list with an External badge. Wait for the status to reach Running.

External Cluster Detail

Once onboarded, click the cluster name to view the external cluster detail. The External badge is shown next to the cluster name.

Available actions for external clusters:

Action	Description
Share Cluster	Share access to the cluster with other tenants or users.
Kubeconfig	Download the kubeconfig for external access.
Connect	Open the cluster console.
Delete	Remove the cluster from Bridge management.

The Nodes tab shows the nodes registered from the external cluster.

note

Deleting an external cluster removes it from Bridge management only. The underlying cluster infrastructure is not affected.

Delete a Cluster

1. In the cluster detail view or cluster list, click Delete.
2. In the confirmation dialog, type the cluster name exactly to confirm deletion.
3. Click Delete. The cluster status changes to Deleting.

warning

Deleting a cluster permanently removes all associated resources. This action cannot be undone.

Next Steps

• Cluster Templates — Learn about available cluster templates.
• Cluster Scaling — Scale cluster nodes up or down.
• Cluster Dashboard — Open the cluster view and use the Kubectl Terminal from the Bridge UI.