• About
• Prerequisites
• Quick Start
  • Preparing your GPU Nodes
  • Enabling GPU Support in Kubernetes
  • Running GPU Jobs
• Docs
• Changelog
• Issues and Contributing

The NVIDIA device plugin for Kubernetes is a Daemonset that allows you to automatically:

• Expose the number of GPUs on each nodes of your cluster
• Keep track of the health of your GPUs
• Run GPU enabled containers in your Kubernetes cluster.

This repository contains NVIDIA's official implementation of the Kubernetes device plugin.

The list of prerequisites for running the NVIDIA device plugin is described below:

• NVIDIA drivers ~= 384.81
• nvidia-docker version > 2.0 (see how to install and it's prerequisites)
• docker configured with nvidia as the default runtime.
• Kubernetes version >= 1.10

The following steps need to be executed on all your GPU nodes. This README assumes that the NVIDIA drivers and nvidia-docker have been installed.

Note that you need to install the nvidia-docker2 package and not the nvidia-container-toolkit. This is because the new --gpus options hasn't reached kubernetes yet. Example:

# Add the package repositories
$ distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
$ curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
$ curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list

$ sudo apt-get update && sudo apt-get install -y nvidia-docker2
$ sudo systemctl restart docker

You will need to enable the nvidia runtime as your default runtime on your node. We will be editing the docker daemon config file which is usually present at /etc/docker/daemon.json:

{
    "default-runtime": "nvidia",
    "runtimes": {
        "nvidia": {
            "path": "/usr/bin/nvidia-container-runtime",
            "runtimeArgs": []
        }
    }
}

if runtimes is not already present, head to the install page of nvidia-docker

Once you have enabled this option on all the GPU nodes you wish to use, you can then enable GPU support in your cluster by deploying the following Daemonset:

$ kubectl create -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/1.0.0-beta6/nvidia-device-plugin.yml

NVIDIA GPUs can now be consumed via container level resource requirements using the resource name nvidia.com/gpu:

apiVersion: v1
kind: Pod
metadata:
  name: gpu-pod
spec:
  containers:
    - name: cuda-container
      image: nvidia/cuda:9.0-devel
      resources:
        limits:
          nvidia.com/gpu: 2 # requesting 2 GPUs
    - name: digits-container
      image: nvidia/digits:6.0
      resources:
        limits:
          nvidia.com/gpu: 2 # requesting 2 GPUs

WARNING: if you don't request GPUs when using the device plugin with NVIDIA images all the GPUs on the machine will be exposed inside your container.

Please note that:

• the device plugin feature is beta as of Kubernetes v1.11.
• the NVIDIA device plugin is still considered beta and is missing
  • More comprehensive GPU health checking features
  • GPU cleanup features
  • ...
• support will only be provided for the official NVIDIA device plugin.

The next sections are focused on building the device plugin and running it.

Option 1, pull the prebuilt image from Docker Hub:

$ docker pull nvidia/k8s-device-plugin:1.0.0-beta6

Option 2, build without cloning the repository:

$ docker build -t nvidia/k8s-device-plugin:1.0.0-beta6 https://github.com/NVIDIA/k8s-device-plugin.git#1.0.0-beta6

Option 3, if you want to modify the code:

$ git clone https://github.com/NVIDIA/k8s-device-plugin.git && cd k8s-device-plugin
$ git checkout 1.0.0-beta6
$ docker build -t nvidia/k8s-device-plugin:1.0.0-beta6 .

$ docker run --security-opt=no-new-privileges --cap-drop=ALL --network=none -it -v /var/lib/kubelet/device-plugins:/var/lib/kubelet/device-plugins nvidia/k8s-device-plugin:1.0.0-beta6

$ kubectl create -f nvidia-device-plugin.yml

$ C_INCLUDE_PATH=/usr/local/cuda/include LIBRARY_PATH=/usr/local/cuda/lib64 go build

$ ./k8s-device-plugin

• Add a new plugin.yml variant that is compatible with the CPUManager
• Change CMD in Dockerfile to ENTRYPOINT
• Add flag to optionally return list of device nodes in Allocate() call
• Refactor device plugin to eventually handle multiple resource types
• Move plugin error retry to event loop so we can exit with a signal
• Update all vendored dependencies to their latest versions
• Fix bug that was inadvertently always disabling health checks
• Update minimal driver version to 384.81

• Fixes a bug with a nil pointer dereference around getDevices:CPUAffinity

• Manifest is updated for Kubernetes 1.16+ (apps/v1)
• Adds more logging information

• Adds the Topology field for Kubernetes 1.16+

• If gRPC throws an error, the device plugin no longer ends up in a non responsive state.

• Reversioned to SEMVER as device plugins aren't tied to a specific version of kubernetes anymore.

• No change.

• The device Plugin API is now v1beta1

• The device Plugin API changed and is no longer compatible with 1.8
• Error messages were added

Checkout the Contributing document!

• You can report a bug by filing a new issue
• You can contribute by opening a pull request

Before 1.10 the versioning scheme of the device plugin had to match exactly the version of Kubernetes. After the promotion of device plugins to beta this condition was was no longer required. We quickly noticed that this versioning scheme was very confusing for users as they still expected to see a version of the device plugin for each version of Kubernetes.

We recently decided to reversion to follow a SEMVER scheme. This means that we are currently a beta project (as we depend on the device plugin API which is beta). If you have a version of Kubernetes > 1.10 you can deploy this device plugin.

Upgrading Kubernetes when you have a device plugin deployed doesn't require you to do any, particular changes to your workflow. The API is versioned and is pretty stable (though it is not guaranteed to be non breaking), you can therefore use the 1.0.0-beta3 version starting from kubernetes version 1.10, upgrading kubernetes won't require you to deploy a different version of the device plugin and you will see GPUs re-registering themselves after your node comes back online.

Upgrading the device plugin is a more complex task. It is recommended to drain GPU tasks as we cannot guarantee that GPU tasks will survive a rolling upgrade. However we make best efforts to preserve GPU tasks during an upgrade.