Challenges running multiple instances of device-plugin pods?

NVIDIA / k8s-device-plugin

NVIDIA device plugin for Kubernetes

Apache License 2.0

2.65k stars 606 forks source link

Challenges running multiple instances of device-plugin pods? #455

Open kodurunaresh opened 9 months ago

kodurunaresh commented 9 months ago

Hello,

I have a question: Besides scheduling and resource conflicts, are there any additional challenges if we run multiple instances of the Nvidia-device-plugin pod on a node for a duration of 5-10 minutes?

Version: 1.0.0-beta4

Thanks.

kodurunaresh commented 9 months ago

/label question

klueska commented 9 months ago

Regardless of your question, version 1.0.0-beta4 was released in 2019. The latest version is 0.14.2 (released last month). I would highly recommend upgrading.

Please see the following regarding versioning: https://github.com/NVIDIA/k8s-device-plugin#versioning