Delete the GPU OperatorΒΆ

This is a pre-release feature and may not function as described in StarlingX 5 documentation.

Use the commands in this section to delete the GPU Operator, if required.

Prerequisites

Ensure that all user generated pods with access to nvidia.com/gpu resources are deleted first.

Procedure

  1. Remove the GPU Operator pods from the system using the following commands:

    ~(keystone_admin)]$ helm delete gpu-operator
~(keystone_admin)]$ kubectl delete runtimeclasses.node.k8s.io nvidia

  2. Remove the GPU Operator, and remove the service parameter platform container_runtime custom_container_runtime from the system, using the following commands:

    1. Lock the host(s).

      ~(keystone_admin)]$ system host-lock <hostname>

    2. List the service parameter using the following command.

      ~(keystone_admin)]$ system service-parameter-list

    3. Remove the service parameter platform container_runtime custom_container_runtime from the system, using the following command.

      ~(keystone_admin)]$ system service-parameter-delete <service param ID>

      where <service param ID> is the ID of the service parameter, for example, 3c509c97-92a6-4882-a365-98f1599a8f56.

    4. Unlock the hosts(s).

      ~(keystone_admin)]$ system host-unlock <hostname>

    For information on configuring the GPU Operator, see Configure NVIDIA GPU Operator for PCI Passthrough Operator.