Kubernetes Device Plugin for NVIDIA GPUs

This directory contains the code for a Kubernetes device plugin for NVIDIA GPUs.

The daemonset manifest at https://github.com/kubernetes/kubernetes/blob/master/cluster/addons/device-plugins/nvidia-gpu/daemonset.yaml can be used to deploy this device plugin to a cluster (1.9 onwards).

In GKE, from 1.9 onwards, this daemonset is automatically deployed as an addon. Note that daemonset pods are only scheduled on nodes with accelerators attached, they are not scheduled on nodes that don't have any accelerators attached.

This device plugin requires that NVIDIA drivers and libraries are installed in a particular way.

Examples of how driver installation needs to be done can be found at:

For COS:
- Installer code: https://github.com/GoogleCloudPlatform/cos-gpu-installer
- Installer daemonset: https://github.com/GoogleCloudPlatform/container-engine-accelerators/blob/master/daemonset.yaml
For Ubuntu (experimental):
- Installer code: https://github.com/GoogleCloudPlatform/container-engine-accelerators/blob/master/nvidia-driver-installer/ubuntu/entrypoint.sh
- Installer daemonset: https://github.com/GoogleCloudPlatform/container-engine-accelerators/blob/master/nvidia-driver-installer/ubuntu/daemonset.yaml

In short, this device plugins expects that all the nvidia libraries needed by the containers are present under a single directory on the host. You can specify the directory on the host containing nvidia libraries using -host-path. You can specify the location to mount that directory in all the containers using -container-path. For example, let's say on the host all nvidia libraries are present under /var/lib/nvidia/lib64 and you want to make these libraries available to containers under /usr/local/nvidia/lib64, then you would use -host-path=/var/lib/nvidia/lib64 and -container-path=/usr/local/nvidia/lib64.