You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When working with podman on fedora and using nvidia GPU using the CDI backend, everytime the kernel is updated or the driver is updated, nvidia-ctk cdi generate ... to be required to be re-run for the CDI container to find the appropriate libraries.
Is this a fedora issue or would it be possible to have a hook to run this command automatically once the kernel module is rebuilt?
The text was updated successfully, but these errors were encountered:
I have to run it after every reboot on a fresh fedora 41. I'm uncertain of this whole flow tbh. Or what I'm doing wrong. I'm missing /dev/nvidia-uvm until I "generate" below.
I just run both of these even though the files persist.
nvidia-ctk config default > /etc/nvidia-container-runtime/config.toml
nvidia-ctk cdi generate --output=/etc/cdi/nvidia.yaml
Hello,
When working with podman on fedora and using nvidia GPU using the CDI backend, everytime the kernel is updated or the driver is updated,
nvidia-ctk cdi generate ...
to be required to be re-run for the CDI container to find the appropriate libraries.Is this a fedora issue or would it be possible to have a hook to run this command automatically once the kernel module is rebuilt?
The text was updated successfully, but these errors were encountered: