Skip to main content
Skip table of contents

GPUs in an AWS environment

Understanding GPUs

Before working with GPUs, ensure you are familiar with the following:

Kommander also accesses Kommander GPU configuration resources.

For using GPUs in an air-gapped on-premises environment, D2iQ recommends setting up Pod Disruption Budget before Update Cluster Nodepools.

Install GPU Support for Supported Distributions on AWS

Using the Konvoy Image Builder, you can build an image that has support to use NVIDIA GPU hardware to support GPU workloads. DKP supported NVIDIA driver version is 470.x.

  1. In your overrides/nvidia.yaml file add the following to enable GPU builds. You can also access and use the overrides repo.

        - nvidia

    Build your image using the following Konvoy image builder commands:

    konvoy-image build --region us-west-2 --source-ami=ami-12345abcdef images/ami/centos-7.yaml --overrides overrides/nvidia.yaml

    By default, your image builds in the us-west-2 region. To specify another region, set the --region flag:

    konvoy-image build --region us-east-1 --overrides override-source-ami.yaml images/ami/<Your OS>.yaml

Ensure that an AMI file is available for your OS selection.

  1. When the command is complete the ami id is printed and written to ./manifest.json.

    To use the built ami with Konvoy Image Builder, specify it with the --ami flag when calling cluster create.

    dkp create cluster aws --cluster-name=$(whoami)-aws-cluster --region us-west-2 --ami <ami>

If your environment uses HTTP/HTTPS proxies, you must include the flags --http-proxy, --https-proxy, and --no-proxy and their related values in this command for it to be successful. More information is available in Configuring an HTTP/HTTPS Proxy.

Additional helpful information can be found in the NVIDIA Device Plug-in for Kubernetes instructions and the Installation Guide of Supported Platforms.

See also: NVIDIA documentation

JavaScript errors detected

Please note, these errors can depend on your browser setup.

If this problem persists, please contact our support.