Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat(ansible): revert "upgrade for CUDA, TensorRT and CUDNN" #5729

Merged
merged 1 commit into from
Feb 5, 2025

Conversation

youtalk
Copy link
Member

@youtalk youtalk commented Feb 5, 2025

Reverts #5608

It was found that due to the impact of this PR, the build started failing, just like with this job.
https://github.com/autowarefoundation/autoware/actions/runs/13132774294/job/36641359188?pr=5688#step:5:20045

It is necessary to revert it for now.

@youtalk youtalk self-assigned this Feb 5, 2025
Copy link

github-actions bot commented Feb 5, 2025

Thank you for contributing to the Autoware project!

🚧 If your pull request is in progress, switch it to draft mode.

Please ensure:

@youtalk youtalk added the tag:run-health-check Run health-check label Feb 5, 2025
@youtalk youtalk marked this pull request as ready for review February 5, 2025 00:23
@youtalk youtalk enabled auto-merge (squash) February 5, 2025 00:23
@youtalk youtalk requested a review from amadeuszsz February 5, 2025 00:24
@youtalk youtalk disabled auto-merge February 5, 2025 01:02
@amadeuszsz
Copy link
Contributor

Could we instead check deps fix? #5730

@amadeuszsz
Copy link
Contributor

The linked PR fixes missing dependencies for arm64 platform. However, without required 0.41.0 release, only nightly build will be valid due to necessary API changes in autoware.universe.

I only wonder about linking autoware main branch with autoware.universe latest tagged version. Shouldn't health-check use both autoware and autoware.universe latest tagged version, and health-check-nightly both main branches with making it required workflow check? Otherwise, if any autoware.universe change will require breaking change in autoware, this check will fail.
I'm not aware of image deployment pipeline in detail, but probably it will also be affected.

@youtalk
Copy link
Member Author

youtalk commented Feb 5, 2025

First, we revert this.
@amadeuszsz Then, pealse create a new PR that includes the fix patch. Otherwise, the problem will become more complicated.

@amadeuszsz
Copy link
Contributor

@youtalk
Ok 👍🏻 Then after revert I'll create draft with CUDA upgrade already containing arm64 fix and we can open & merge it after 0.41.0 release.

@amadeuszsz
Copy link
Contributor

@youtalk

Actually, if we deploy autoware-base by manual trigger with old deps, we don't need to revert PR since container will be build using cached (downgraded) image. Just arm64 fix merge will be needed and later manual trigger of autoware-base after 0.41.0 release. If I'm mistaken, please ignore this comment.

@youtalk youtalk enabled auto-merge (squash) February 5, 2025 03:21
@youtalk
Copy link
Member Author

youtalk commented Feb 5, 2025

the problem will become more complicated.

@mitsudome-r
Copy link
Member

@amadeuszsz Just to confirm, we don't have to revert this one , correct?

@amadeuszsz
Copy link
Contributor

@amadeuszsz Just to confirm, we don't have to revert this one , correct?

@mitsudome-r
Correct, this PR keeps compatibility with old and new dependencies.

@youtalk
Copy link
Member Author

youtalk commented Feb 5, 2025

The base image was changed, so it’s taking time to run the health-check from scratch.
However the problematic section has been resolved, so the health-check should pass.

https://github.com/autowarefoundation/autoware/actions/runs/13147476391/job/36694435889?pr=5729#step:5:19953

#98 19.26 Finished <<< autoware_tensorrt_common [16.1s]

@youtalk youtalk merged commit 8da1a3e into main Feb 5, 2025
28 checks passed
@youtalk youtalk deleted the revert-5608-feat/cuda-trt-upgrade branch February 5, 2025 05:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
tag:run-health-check Run health-check
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants