GPU nodes

Available in VPC

GPU nodes may be used as worker nodes of the Ncloud Kubernetes Service.

Caution

GPU servers provided by NAVER Cloud Platform do not support NVIDIA's Multi-Instance GPU (MIG) feature. For stable service operation, you should use the default GPU settings provided.
GPU nodes can only be used in the KR region.
Clusters exclusively made up of GPU nodes have usage restrictions.
As any default object in Ncloud Kubernetes Service requires general nodes, you need to set up both general and GPU node pools.

Note

In GPU nodes, you can upgrade your SMI/CUDA driver.
You can upgrade your driver using runfile. For detailed instructions see Reinstall and upgrade GPU driver/CUDA.

Place NVIDIA GPU device plugin

You can't use GPU resources in the Kubernetes cluster before placing the GPU plugin.
In order to place the NVIDIA device plugin after the GPU node switches to Running at the cluster, execute the following command:

1. Install Helm

curl -fsSL -o get_helm.sh https://raw.githubusercontent.com/helm/helm/master/scripts/get-helm-3 \
   && chmod 700 get_helm.sh \
   && ./get_helm.sh

2. Add NVIDIA device plugin helm storage

helm repo add nvdp https://nvidia.github.io/k8s-device-plugin \
   && helm repo update

3. Release NVIDIA device plugin

Now, deploy the NVIDIA device plugin in the kube-system namespace using the following command.

helm install --generate-name nvdp/nvidia-device-plugin -n kube-system

Note

The HELM Chart is provided by NVIDIA. If you encounter any issues, check the official NVIDIA website to see if it has changed.

For a quick start, you can directly deploy a simple Damonset to activate your GPU resources.

$ kubectl create -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/v0.17.0/deployments/static/nvidia-device-plugin.yml

4. Check GPU usage

You can check the GPU resource availability by viewing the details of the node using the following command.

$ kubectl describe [node-name]

...
Capacity:
  cpu:                4
  ephemeral-storage:  51341792Ki
  hugepages-1Gi:      0
  hugepages-2Mi:      0
  memory:             20474628Ki
  nvidia.com/gpu:     1
  pods:               110
Allocatable:
  cpu:                3920m
  ephemeral-storage:  47316595429
  hugepages-1Gi:      0
  hugepages-2Mi:      0
  memory:             17425156Ki
  nvidia.com/gpu:     1
  pods:               110
...

Use device plugin

Kubernetes implements device plugins and enables its pods to access special hardware features such as GPU. For more information on how to use the plugin, see the official website.