NVIDIA AI Enterprise

NVIDIA AI Enterprise is an end-to-end, cloud-native suite of AI and data analytics software, optimized, certified, and supported by NVIDIA with  NVIDIA-Certified  Systems. Additional information can be found at the NVIDIA AI Enterprise web page.

NVIDIA AI Enterprise customers have access to a pre-configured GPU Operator within the NVIDIA Enterprise Catalog. The GPU Operator is pre-configured to simplify the provisioning experience with NVIDIA AI Enterprise deployments.

The pre-configured GPU Operator differs from the GPU Operator in the public NGC catalog. The differences are:

  • It is configured to use a prebuilt vGPU driver image (Only available to NVIDIA AI Enterprise customers)

  • It is configured to use the NVIDIA License System (NLS)

This page provides documentation for the following configurations:

  • Kubernetes on bare metal and on vSphere VMs with GPU passthrough and vGPU.

  • VMware vSphere with Tanzu.

  • Red Hat Openshift on bare metal and on vSphere VMs with GPU passthrough and vGPU.

The following sections are applicable to the first two configurations and describe how to deploy the GPU Operator using the Helm Chart. For Red Hat Openshift configurations, please follow this procedure NVIDIA AI Enterprise with OpenShift <nvaie-ocp>.

Installing GPU Operator

To install GPU Operator with NVIDIA AI Enterprise, apply the following steps.

Note

You can also use the following script, which automates the below installation instructions.

Create the gpu-operator namespace:

$ kubectl create namespace gpu-operator

Create an empty vGPU license configuration file:

$ sudo touch gridd.conf

Generate and download a NLS client license token. Please refer to Section 4.6 of the NLS User Guide for instructions.

Rename the NLS client license token that you downloaded to client_configuration_token.tok.

Create the licensing-config ConfigMap object in the gpu-operator namespace. Both the vGPU license configuration file and the NLS client license token will be added to this ConfigMap:

$ kubectl create configmap licensing-config \
    -n gpu-operator --from-file=gridd.conf --from-file=<path>/client_configuration_token.tok

Create an image pull secret in the gpu-operator namespace for the private registry that contains the containerized NVIDIA vGPU software graphics driver for Linux for use with NVIDIA GPU Operator:

  • Set the registry secret name:

$ export REGISTRY_SECRET_NAME=ngc-secret

  • Set the private registry name:

$ export PRIVATE_REGISTRY=nvcr.io/nvaie

  • Create an image pull secret in the gpu-operator namespace with the registry secret name and the private registry name that you set. Replace password, and email-address with your NGC API key and email address respectively:

$ kubectl create secret docker-registry ${REGISTRY_SECRET_NAME} \
    --docker-server=${PRIVATE_REGISTRY} \
    --docker-username='$oauthtoken' \
    --docker-password='<password>' \
    --docker-email='<email-address>' \
    -n gpu-operator

Add the NVIDIA AI Enterprise Helm repository, where password is the NGC API key for accessing the NVIDIA Enterprise Collection that you generated:

$ helm repo add nvaie https://helm.ngc.nvidia.com/nvaie \
  --username='$oauthtoken' --password='<password>' \
  && helm repo update

Install the NVIDIA GPU Operator:

$ helm install --wait gpu-operator nvaie/gpu-operator-2-0 -n gpu-operator

Note

in case you need to deploy the Helm chart with some customizations, please refer to this link: Chart Customization Options

Installing GPU Operator with the NVIDIA Datacenter Driver

To install GPU Operator on baremetal with the NVIDIA Datacenter Driver, apply the following steps.

Note

You can also use the following script, which automates the below installation instructions.

Create the gpu-operator namespace:

$ kubectl create namespace gpu-operator

Create an image pull secret in the gpu-operator namespace for the private registry that contains the NVIDIA GPU Operator:

  • Set the registry secret name:

$ export REGISTRY_SECRET_NAME=ngc-secret

  • Set the private registry name:

$ export PRIVATE_REGISTRY=nvcr.io/nvaie

  • Create an image pull secret in the gpu-operator namespace with the registry secret name and the private registry name that you set. Replace password, and email-address with your NGC API key and email address respectively:

Add the NVIDIA AI Enterprise Helm repository, where password is the NGC API key for accessing the NVIDIA Enterprise Collection that you generated:

$ helm repo add nvaie https://helm.ngc.nvidia.com/nvaie \
  --username='$oauthtoken' --password='<password>' \
  && helm repo update

Install the NVIDIA GPU Operator:

$ helm install --wait gpu-operator nvaie/gpu-operator-2-0 -n gpu-operator \
  --set driver.repository=nvcr.io/nvidia \
  --set driver.image=driver \
  --set driver.version=510.47.03 \
  --set driver.licensingConfig.config.name=""

Note

In case you need to deploy the Helm chart with some customizations, please refer to this link: Chart Customization Options

Updating NLS client license token

In case the NLS client license token needs to be updated, please use the following procedure:

Create an empty vGPU license configuration file:

$ sudo touch gridd.conf

Generate and download a new NLS client license token. Please refer to Section 4.6 of the NLS User Guide for instructions.

Rename the NLS client license token that you downloaded to client_configuration_token.tok.

Create a new licensing-config-new ConfigMap object in the gpu-operator namespace (make sure the name of the configmap is not already used in the kubernetes cluster). Both the vGPU license configuration file and the NLS client license token will be added to this ConfigMap:

$ kubectl create configmap licensing-config-new \
    -n gpu-operator --from-file=gridd.conf --from-file=<path>/client_configuration_token.tok

Edit the clusterpolicies by using the command:

$ kubectl edit clusterpolicies.nvidia.com

Go to the driver section and replace the following argument:

licensingConfig:
    configMapName: licensing-config

with

licensingConfig:
    configMapName: licensing-config-new

Write and exit from the kubectl edit session (you can use :qw for instance if vi utility is used)

GPU Operator will redeploy sequentially all the driver pods with this new licensing information.