Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fix: Avoid CDI device injection error by reverting to legacy GPU device specification
Description: This update addresses a runtime error related to unresolved CDI devices (nvidia.com/gpu=all) during container launch on systems where CDI support is not properly configured or enabled.
Reason for change: Some environments, especially Jetson-based platforms or hosts with older versions of the NVIDIA Container Toolkit, do not support the CDI device naming convention (nvidia.com/gpu=all). This leads to a fatal error during container startup:
OCI runtime create failed: failed to inject CDI devices: unresolvable CDI devices nvidia.com/gpu=all: unknown
Changes made:
Replaced NVIDIA_VISIBLE_DEVICES=nvidia.com/gpu=all,nvidia.com/pva=all with the more broadly compatible NVIDIA_VISIBLE_DEVICES=all on ARM64 (Jetson).
This ensures compatibility with both CDI-enabled and legacy NVIDIA container runtimes.
Impact: Improves container launch stability across a wider range of Jetson and desktop environments by falling back to the standard GPU device exposure mechanism.