Install NVIDIA GPU Support on CoreOS based Kubernetes Cluster
- CoreOS based Kubernetes cluster with GPU nodes (e.g. AWS P2 instances)
Install helm via
helm init --service-account tiller
kubectl apply -f https://raw.githubusercontent.com/gardener/gpu-installer/master/manifests/helm-service-account.yaml
Install the provided helm chart
helm install --name gpu-installer --namespace kube-system charts/nvidia-installer
First install the nvidia driver via this daemonset
kubectl apply -f https://raw.githubusercontent.com/gardener/gpu-installer/master/manifests/k8s-nvidia-driver.yaml
Wait until the init container finishes on each node and install the device plugin
kubectl apply -f https://raw.githubusercontent.com/gardener/gpu-installer/master/manifests/k8s-nvidia-deviceplugin.yaml
- Build and install NVIDIA driver on CoreOS: https://github.com/squat/modulus
- Nvidia Device Plugin: https://github.com/kubernetes/kubernetes/blob/master/cluster/addons/device-plugins/nvidia-gpu/daemonset.yaml