All Products
Search

This Product

    Container Service for Kubernetes:Add GPU nodes to a cluster

    Document Center

    Container Service for Kubernetes:Add GPU nodes to a cluster

    Last Updated:Jun 25, 2026

    NVIDIA GPU devices are often used to accelerate compute-intensive applications, such as scientific computing and graphics rendering. Container Service for Kubernetes lets you centrally schedule and manage various types of GPU resources, improving GPU cluster utilization. This topic shows you how to add GPU nodes to a cluster.

    Prerequisites

    You must have an existing ACK Pro cluster or ACK dedicated cluster (no longer available for creation).

    Create a GPU node pool

    1. Log on to the ACK console. In the left navigation pane, click Clusters.

    2. On the Clusters page, click the name of your cluster. In the left navigation pane, click Nodes > Node Pools.

    3. Click Create Node Pool. Set Instance Type to Elastic GPU Service and Expected Nodes to the number of nodes you want. For more information about other parameters, see Create and manage a node pool.

      For a list of supported GPU-accelerated ECS instance types, see GPU instance types supported by ACK.

      Note

      • If no available instance types appear in the list, try selecting a different vSwitch.

      • If your node operating system is Ubuntu 22.04 or Red Hat Enterprise Linux (RHEL) 9.3 64-bit, the NVIDIA Device Plugin component sets the NVIDIA_VISIBLE_DEVICES=all environment variable for pods by default. After you run the systemctl daemon-reload or systemctl daemon-reexec command on the node, it might lose access to its GPU devices, causing the NVIDIA Device Plugin to stop working. For more information, see What do I do if the "Failed to initialize NVML: Unknown Error" error occurs when I run a GPU container?.

    View attached GPU devices

    After the node pool is created, you can view the GPU devices attached to the nodes.

    1. Log on to the ACK console. In the left navigation pane, click Clusters.

    2. On the Clusters page, click the name of your cluster. In the left navigation pane, click Nodes > Nodes.

    3. For the target node, click Details in the Actions column to view its attached GPU devices. To find the specific GPU model, click the node's ECS instance ID to go to the instance details page in the ECS console. On the Instance Details tab, view the GPU field in the Configuration Information section. For example, the field might show GPU: NVIDIA T4 for an ecs.gn6i-c4g1.xlarge instance type.