Facing issue with a distributed computing setup using Kubernetes, where GPUs (3080 Ti) are recognized on control plane nodes, but GPUs (1660 Super and 1660 Ti) on 18 worker nodes are not being detected.

Safal_Shrestha · November 30, 2025, 7:22pm

I am currently studying B.Tech in AI at Kathmandu University, Nepal. While setting up distributed computing through Kubernetes, I’ve encountered an issue where the GPU is not being detected on the worker nodes. Specifically, I have configured one load balancer and three control plane nodes. The control planes successfully recognize the GPU (3080 Ti), but none of the 18 worker nodes, which have 1660 Super and 1660 Ti GPUs, show the GPU.

Topic		Replies	Views
Local test cluster - hardware requirements General Discussions development	1	1254	October 15, 2020
How can I know which gpu is scheduled General Discussions development	0	350	May 9, 2023
How i can use multiple gpus with kubernetes General Discussions development	0	401	May 26, 2023
Limit GPUs on Node available to Pods microk8s	0	768	July 28, 2019
Containers not allocated exclusive GPU when multiple Pods are deployed General Discussions microk8s	0	173	June 18, 2025

Facing issue with a distributed computing setup using Kubernetes, where GPUs (3080 Ti) are recognized on control plane nodes, but GPUs (1660 Super and 1660 Ti) on 18 worker nodes are not being detected.

Related topics