Run:AI Installation

Contributors kevin-hoke

To install Run:AI, complete the following steps:

  1. Install the Kubernetes cluster using DeepOps and configure the NetApp default storage class.

  2. Prepare GPU nodes:

    1. Verify that NVIDIA drivers are installed on GPU nodes.

    2. Verify that nvidia-docker is installed and configured as the default docker runtime.

  3. Install Run:AI:

    1. Log into the Run:AI Admin UI to create the cluster.

    2. Download the created runai-operator-<clustername>.yaml file.

    3. Apply the operator configuration to the Kubernetes cluster.

      kubectl apply -f runai-operator-<clustername>.yaml
  4. Verify the installation:

    1. Go to https://app.run.ai/.

    2. Go to the Overview dashboard.

    3. Verify that the number of GPUs on the top right reflects the expected number of GPUs and the GPU nodes are all in the list of servers.For more information about Run:AI deployment, see installing Run:AI on an on-premise Kubernetes cluster and installing the Run:AI CLI.