Document setup with CDI annotations
In our deployment, we found that CDI annotations were more stable, and avoided the issue (other workarounds were not effective) described at https://github.com/NVIDIA/gpu-operator/issues/485.
The setup is a bit tricky, and requires configuring an explicit runtime class for the device plugin itself. Document this here for other users.
Migrated from https://gitlab.com/nvidia/kubernetes/device-plugin/-/merge_requests/298
@mikemckiernan, could you have a peak of these documentation changes? thanks!
This PR is stale because it has been open 90 days with no activity. This PR will be closed in 30 days unless new comments are made or the stale label is removed.
no stale