NVIDIA GPU (Proprietary drivers)

Enabling NVIDIA GPU support on Talos is bound by NVIDIA EULA. The Talos published NVIDIA drivers are bound to a specific Talos release. The extensions versions also needs to be updated when upgrading Talos.

The published versions of the NVIDIA system extensions can be found here:

• nonfree-kmod-nvidia
• nvidia-container-toolkit

To build a NVIDIA driver version not published by SideroLabs jump to Building the NVIDIA extensions and then use those in the steps below instead of the ones published by SideroLabs

Upgrading Talos and enabling the NVIDIA modules and the system extension

Make sure to use talosctl version v1.5.5 or later

First create a patch yaml gpu-worker-patch.yaml to update the machine config similar to below:

- op: add
  path: /machine/install/extensions
  value:
    - image: ghcr.io/siderolabs/nonfree-kmod-nvidia:535.54.03-v1.5.5
    - image: ghcr.io/siderolabs/nvidia-container-toolkit:535.54.03-v1.13.5
- op: add
  path: /machine/kernel
  value:
    modules:
      - name: nvidia
      - name: nvidia_uvm
      - name: nvidia_drm
      - name: nvidia_modeset
- op: add
  path: /machine/sysctls
  value:
    net.core.bpf_jit_harden: 1

Update the driver version and Talos release in the above patch yaml from the published versions if there is a newer one available. Make sure the driver version matches for both the nonfree-kmod-nvidia and nvidia-container-toolkit extensions. The nonfree-kmod-nvidia extension is versioned as <nvidia-driver-version>-<talos-release-version> and the nvidia-container-toolkit extension is versioned as <nvidia-driver-version>-<nvidia-container-toolkit-version>.

Now apply the patch to all Talos nodes in the cluster having NVIDIA GPU’s installed:

talosctl patch mc --patch @gpu-worker-patch.yaml

Now we can proceed to upgrading Talos to the same version to enable the system extension:

talosctl upgrade --image=ghcr.io/siderolabs/installer:v1.5.5

Once the node reboots, the NVIDIA modules should be loaded and the system extension should be installed.

This can be confirmed by running:

talosctl read /proc/modules

which should produce an output similar to below:

nvidia_uvm 1146880 - - Live 0xffffffffc2733000 (PO)
nvidia_drm 69632 - - Live 0xffffffffc2721000 (PO)
nvidia_modeset 1142784 - - Live 0xffffffffc25ea000 (PO)
nvidia 39047168 - - Live 0xffffffffc00ac000 (PO)

talosctl get extensions

which should produce an output similar to below:

NODE           NAMESPACE   TYPE              ID                                                                 VERSION   NAME                       VERSION
172.31.41.27   runtime     ExtensionStatus   000.ghcr.io-frezbo-nvidia-container-toolkit-510.60.02-v1.9.0       1         nvidia-container-toolkit   510.60.02-v1.9.0

talosctl read /proc/driver/nvidia/version

which should produce an output similar to below:

NVRM version: NVIDIA UNIX x86_64 Kernel Module  510.60.02  Wed Mar 16 11:24:05 UTC 2022
GCC version:  gcc version 11.2.0 (GCC)

Deploying NVIDIA device plugin

First we need to create the RuntimeClass

Apply the following manifest to create a runtime class that uses the extension:

---
apiVersion: node.k8s.io/v1
kind: RuntimeClass
metadata:
  name: nvidia
handler: nvidia

Install the NVIDIA device plugin:

helm repo add nvdp https://nvidia.github.io/k8s-device-plugin
helm repo update
helm install nvidia-device-plugin nvdp/nvidia-device-plugin --version=0.13.0 --set=runtimeClassName=nvidia

(Optional) Setting the default runtime class as nvidia

Do note that this will set the default runtime class to nvidia for all pods scheduled on the node.

Create a patch yaml nvidia-default-runtimeclass.yaml to update the machine config similar to below:

- op: add
  path: /machine/files
  value:
    - content: |
        [plugins]
          [plugins."io.containerd.grpc.v1.cri"]
            [plugins."io.containerd.grpc.v1.cri".containerd]
              default_runtime_name = "nvidia"        
      path: /etc/cri/conf.d/20-customization.part
      op: create

Now apply the patch to all Talos nodes in the cluster having NVIDIA GPU’s installed:

talosctl patch mc --patch @nvidia-default-runtimeclass.yaml

Testing the runtime class

Note the spec.runtimeClassName being explicitly set to nvidia in the pod spec.

Run the following command to test the runtime class:

kubectl run \
  nvidia-test \
  --restart=Never \
  -ti --rm \
  --image nvcr.io/nvidia/cuda:12.1.0-base-ubuntu22.04 \
  --overrides '{"spec": {"runtimeClassName": "nvidia"}}' \
  nvidia-smi

Building the NVIDIA extensions

If you want to build the NVIDIA extensions yourself instead of using the extensions published by SideroLabs start by cloning the release-1.5 branch extensions repository.

git clone --depth=1 --branch=release-1.5 https://github.com/siderolabs/extensions.git

Lookup the version of pkgs used for the particular Talos version at https://github.com/siderolabs/talos/blob/<talos-version>/pkg/machinery/gendata/data/pkgs.

Now run the following command to build and push custom NVIDIA extension.

make nonfree-kmod-nvidia PKGS=<pkgs-version-looked-up-above> PLATFORM=linux/amd64 PUSH=true

Replace the platform with linux/arm64 if building for ARM64. To change the NVIDIA driver version modify the build argument in nvidia-gpu/nonfree/kmod-nvidia/vars.yaml accordingly. Make sure to use talosctl version v1.5.5 or later