Is nvidia-ctk --set-as-default flag set as false the default option

[qzoscar]

follow the instruction https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html here to config a k8s gpu node, but it seems failed to config containerd correctly. Seems the instruction

sudo nvidia-ctk runtime configure --runtime=containerd

did not config the containerd's config.toml with the flag --set-as-default set. I wonder if this is a doc mistake or we should make --set-as-default=true default.

[itay-nvn-nv]

+1

the instructions here make it look like its sth to explicitly state:
https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/1.10.1/install-gpu-operator.html#bare-metal-passthrough-with-pre-installed-drivers-and-nvidia-container-toolkit

the containerd config is correct here though:
https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/1.10.1/install-gpu-operator.html#bare-metal-passthrough-with-pre-installed-drivers-and-nvidia-container-toolkit

[github-actions]

This issue is stale because it has been open 90 days with no activity. This issue will be closed in 30 days unless new comments are made or the stale label is removed. To skip these checks, apply the "lifecycle/frozen" label.

[elezar]

We intentionally do not set the default runtime when configuring containerd. Note that this config is usually used in the context of k8s where there are other mechanisms for ensuring that the runtime is used (e.g. RuntimeClass).