NVIDIA beta driver is installed

Bug #1909933 reported by David van der Spek
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Containerd Subordinate Charm
Triaged
Low
Unassigned

Bug Description

I have been experiencing some issues with CDK on channel 1.20/stable and containerd charm revision 100. While I admit that my current setup is not ideal as I have my GPU passed through to a VM which is then configured as a worker node, I have recently had the issue that `nvidia-smi` is not available when the node comes up with the error:
NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

Rebooting fixes the nvidia-smi issue (I don't remember needing to do this before), but I still have some other issues with running GPU pods. Nvidia-smi shows: NVIDIA-SMI 460.27.04 Driver Version: 460.27.04 CUDA Version: 11.2

According to https://www.nvidia.com/Download/driverResults.aspx/167671/en-us this is a BETA driver.

While I am not sure what the exact cause of issues I am experiencing is, I do not believe a beta driver should EVER be installed by a charm, MAAS or JuJu without the user manually choosing to do so.

Tags: gpu nvidia
George Kraft (cynerva)
Changed in charm-containerd:
importance: Undecided → Low
status: New → Triaged
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.