Run:ai Installation

Use the Run:ai BCM installation guide for the Run:ai version included in your deployment. Complete its preparations and follow its air-gapped deployment procedure.

That procedure provisions the user Kubernetes cluster before deploying Run:ai onto it. Use the k8s-system-user nodes for the k8s-user control plane and the GPU worker node category for its worker nodes. The k8s-admin cluster created earlier hosts administrative services; it is not the Run:ai deployment target.

Prepare the offline artifacts on the internet-connected host as directed by the Run:ai guide. Run the deployment wizard from the active BCM head node. The host where you run the wizard is distinct from the Kubernetes nodes you select for the user cluster.

Refer to NVIDIA Run:ai Installation for the BCM category prerequisites.