All documents

Managed – FPT Kubernetes Engine

    Adding GPU Worker Group
    Adding GPU Worker Group
    Updated on 29 Nov 2024

    Prerequisites:

    • CPU, GPU, RAM, Storage, and Instances Quota: enough for the extra Worker Group specifications. If you plan to use auto-scale, you must have enough GPU quota for the maximum amount of nodes that you desire.

    • 01 Network Subnet: Network for the Kubernetes Nodes, with a static IP Pool.

    Steps to configure

    Step 1: Login to the FPT Cloud Portal. From the sidebar menu, choose Kubernetes > Choose the cluster to config. In Detail Cluster > Worker Group > click the Config Icon

    Step 2: Click Add Worker Group to add a new Worker Group.

    Step 3: Select GPU specs in the Worker Group.

    • Choose Instance type: GPU

    • Choose GPU type: A30, A100, etc…

    • Choose GPU Sharing configuration: None, Single, Mixed, etc…

    • Choose Specs type: CPU – RAM – GPU

    Step 4: Click Save and verify the initialization information.

    Step 5: Wait until the cluster succeeds initialize (status: Running) > Ready to use.