Add NVIDIA container runtime for CRI-O #586
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Collaborator
koct9i
commented
Nov 3, 2025
koct9i
commented
- test/r8r: check nvidia container runtime
- Add nvidia container runtime for CRI-O
Contributor
There was a problem hiding this comment.
Pull Request Overview
This PR adds support for CRI-O (Container Runtime Interface - O) as an alternative to containerd for running containerized jobs in YTsaurus exec nodes. It includes configuration generation for CRI-O with NVIDIA GPU runtime support.
Key changes:
- Implemented
GetCRIOConfig()method to generate CRI-O configuration in TOML format - Refactored CRI service configuration handling to support both containerd and CRI-O through a unified
NewJobsSidecarConfig()function - Added test coverage for CRI-O with and without NVIDIA container runtime
Reviewed Changes
Copilot reviewed 64 out of 64 changed files in this pull request and generated no comments.
Show a summary per file
| File | Description |
|---|---|
| pkg/ytconfig/cri.go | Added CRI-O configuration generation with support for multiple runtimes (runc, crun, nvidia) and replaced environment variable-based configuration with TOML config file |
| pkg/components/exec_node_base.go | Created unified NewJobsSidecarConfig() function to handle both containerd and CRI-O configurations, added CRI-O volume mount and command-line arguments |
| pkg/components/exec_node.go | Refactored to use new NewJobsSidecarConfig() function instead of containerd-specific logic |
| pkg/components/exec_node_remote.go | Applied same refactoring as exec_node.go and added monitoring port configuration for CRI service |
| pkg/consts/cmd.go | Added CRI-O-specific constants for config volume, mount point, and file name |
| pkg/testutil/spec_builders.go | Added WithNvidiaContainerRuntime flag to test builder for NVIDIA runtime testing |
| test/r8r/components_test.go | Added test cases for CRI-O with and without NVIDIA container runtime |
Comments suppressed due to low confidence (1)
pkg/components/exec_node_base.go:102
- The volume is always created with
consts.ContainerdConfigVolumeName, but when CRI-O is used, the volume mount inaddCRIServiceConfig()expectsconsts.CRIOConfigVolumeName(line 116). This mismatch will cause the volume mount to fail for CRI-O. The volume name should be determined based on the CRI service type, similar to how volume mounts are handled. Consider using a switch statement or variable to select the appropriate volume name based onn.criConfig.Service.
podSpec.Volumes = append(podSpec.Volumes, createConfigVolume(consts.ContainerdConfigVolumeName,
n.sidecarConfig.labeller.GetSidecarConfigMapName(consts.JobsContainerName), nil))
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Signed-off-by: Konstantin Khlebnikov <[email protected]>
Add configuration file for CRI-O, because configuration via environment variables is not flexible enough. Signed-off-by: Konstantin Khlebnikov <[email protected]>
koct9i
force-pushed
the
khlebnikov/crio-add-nvidia-runtime
branch
from
7ecdb64 to
8cf530a
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.