Create and Install Clusters

Once system requirements are installed on the tenant's Kubernetes environment, you must create the cluster and initiate the installation process.

This step involves:

• Registering the cluster using the /api/v1/clusters API. You’ll provide a cluster name, the associated fully qualified domain name (FQDN), and the required NVIDIA Run:ai version.
• Installing the NVIDIA Run:ai cluster components using the provided Helm chart, which connects the tenant’s cluster to the platform.
• Establishing connectivity between the tenant’s environment and the control plane to enable monitoring, workload submission, and scheduling.

The cluster domain must match the FQDN configured during the system requirements step. After completing this step, the tenant environment becomes fully operational within the NVIDIA Run:ai platform.

Create and Register the Cluster

After installing the system requirements on the tenant’s Kubernetes environment, you must register the cluster in the NVIDIA Run:ai platform. This step creates a unique cluster object and associates it with the correct tenant using the provided domain.

To create a cluster:

• Send a POST request to the /api/v1/clusters endpoint. See Clusters API for more details.
• Provide the cluster’s name, domain (FQDN) and the target NVIDIA Run:ai version.
• Once registered, the platform will return a cluster UUID. This UUID will be used to complete the installation and connect the tenant’s cluster to the NVIDIA Run:ai control plane.

This registration step does not install anything on the tenant’s cluster yet. It only prepares the cluster for installation and establishes the tenant-cluster association.

Example request:

curl -L 'https://console.<DOMAIN>/api/v1/clusters' \ 
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer <TOKEN>' \ 
{
  "name": "string",
  "domain": "string",
  "version": "string",
  "tenantId": 1001
}

Example response:

{
  "uuid": "A0EEBC99-9C0B-4EF8-BB6D-6BB9BD380A11",
  "name": "example",
  "tenantId": 1001,
  "domain": "my.company.com",
  "status": { },
  "createdAt": "2020-01-01T00:00:00Z",
  "updatedAt": "2020-01-02T00:00:00Z",
  "lastLiveness": "2020-01-02T00:00:00Z",
  "version": "2.15.0"
}

Retrieve Cluster Installation Instructions

Once the cluster is registered, the next step is to install the NVIDIA Run:ai components on the tenant’s Kubernetes environment. This is done using the Helm command provided by the platform.

To retrieve the installation command and required parameters:

• Send a GET request to the /v1/clusters/{clusterUuid}/cluster-install-info?version=<version> endpoint using the UUID returned when the cluster was created.
• The response includes a preconfigured installationStr, Helm repository details, and a clientSecret to be used in the install command.

You will run the returned installationStr on the tenant’s cluster to complete the installation and connect it to the NVIDIA Run:ai platform.

Example response:

{
  "installationStr": "helm update --update repo/runai-cluster -n runai --set cluster.url=test_cluster",
  "repositoryName": "runai",
  "chartRepoURL": "https://runai.jfrog.io/artifactory/charts",
  "clientSecret": "ABC333DDD"
}

Install the Cluster

Run the installationStr in the terminal on the tenant’s Kubernetes cluster. This installs the NVIDIA Run:ai components and links the cluster to the tenant in the control plane.

Upgrade the Cluster

To upgrade a tenant's cluster to a newer version of the NVIDIA Run:ai platform, repeat the installation process using the updated target version.