[Bug]: gradle-check action ran for different PR

[andrross]

Describe the bug

I submitted a revision to PR 4869 which appeared to trigger a gradle check action 5204. This reported a failure in my PR, but that gradle check action was actually for a different PR and ran against a different commit, leading to quite a lot of confusion. Seems like maybe a race condition?

More details here: opensearch-project/OpenSearch#4869 (comment)

To reproduce

Probably requires two PR submitting changes nearly simultaneously? Not sure how to deterministically reproduce.

Expected behavior

The gradle check action should run against my commit in my PR, not a different commit.

Screenshots

No response

Host / Environment

No response

Additional context

No response

Relevant log output

No response

[peterzhuamazon]

It is possible that it is a race condition tho I have not seen such cases happen until this point.
Will need to keep this open for a while as this might be a very rare case.

Thanks.

[andrross]

I just observed a similar problem, but in this case it wasn't a PR workflow and instead was the gradle check that runs on merge. See this commit. The little red x links to this GitHub Action Workflow, which links to this gradle check jenkins run, which shows that it ran for an entirely different commit on a user fork

[jordarlu]

Could this be the casue ? actions/checkout#27
Also, both the wrong commit (sandeshkr419/OpenSearch@d3df6ee) and correct commit (opensearch-project/OpenSearch@1cdff3b) happened very closer to each other, could it be the GHA was first triggered by either one but them another one got pushed in before the GHA has completed?
Maybe we can open an issue with Github?

thanks,

[jordarlu]

Hi, I am cloing this issue as it has been a while with no further updating from all stack holders in case... pls feel free to re-open or create a new one if you face the same issue ... thanks,