I'd like to allow my users to work with docker and gpu in pod. The only way to have DinD I know is to start pod as privileged:
apiVersion: v1
kind: Pod
metadata:
name: mini-pod
spec:
containers:
- name: main
image: ubuntu
args:
- sleep infinity
command:
- /bin/bash
- -c
securityContext:
privileged: true
resources:
limits:
amd.com/gpu: 1
Usually I'm checking /dev/dri folder to find out what GPU is available in pod, but because of privileged mode I've all 8 gpu visible in this folder.
How can I find out which gpu is reserved for my pod?
I'd like to allow my users to work with docker and gpu in pod. The only way to have DinD I know is to start pod as privileged:
Usually I'm checking /dev/dri folder to find out what GPU is available in pod, but because of privileged mode I've all 8 gpu visible in this folder.
How can I find out which gpu is reserved for my pod?