Open chalin opened 5 months ago
This discussion links to the official GH documentation about the topic:
When you use the repository's GITHUB_TOKEN to perform tasks, events triggered by the GITHUB_TOKEN, with the exception of workflow_dispatch and repository_dispatch, will not create a new workflow run. This prevents you from accidentally creating recursive workflow runs.
And:
If you do want to trigger a workflow from within a workflow run, you can use a GitHub App installation access token or a personal access token instead of GITHUB_TOKEN to trigger events that require a token.
Using a PAT seems to be a reasonable solution for me.
Using a PAT seems to be a reasonable solution for me.
We have one already, and maybe using it consistently will help. Let me raise a PR for that.
looks like #4345 did not solve it. I tested it with https://github.com/open-telemetry/opentelemetry.io/pull/4349 and now checks were run afterwards.
So the "good" news is that we are not the only ones running in those issues, so there is some info out there that is helpful, but this is a rabbit hole, looks like the checkout action adds additional trouble:
@svrnm - we're hitting permission issues now, see https://github.com/open-telemetry/opentelemetry.io/actions/runs/8822774576/job/24221719586:
@chalin that's odd because this step was not updated by any of the 2 latest changes:
Reverting the last change, as you did was what I had in mind too. Thanks for doing that and investigating further 🙏
Reverting the last change, as you did was what I had in mind too. Thanks for doing that and investigating further 🙏
I am worried we are in a position where neither token gives us all the things we need, as it looks like the PAT is not allowed to push to forks. In my research I came across a page that listed solutions to this problem but I closed it so I have to find it once again will keep you posted.
For context, see:
4335
Unless we get a separate bot to handle
/fix:*
commands, we're stuck with this behavior: GH doesn't rerun actions if a workflow over the same repo committed a change to a PR. I usually work around this by forcing a change (via rebase, or by making a trivial edit) over the PR, but that isn't always practical.Other possible solutions:
/fix:*
commands so that GH runs the full action checks. (This would be a lot of work but might be worth it in the long run.)main
branch. (This is easier to setup, and will at least notify us if build checks are broken).Thoughts @open-telemetry/docs-maintainers?