Skip to content

Conversation

@liamhuber
Copy link
Member

As discussed in #191, it looks like the macos failures are fundamentally not our fault; ideally the runners would handle MPI nicely, but in the meantime I think it is preferable to test on an older version than not to test at all (not to mention much more convenient to not always be restarting tests).

@liamhuber
Copy link
Member Author

@jan-janssen this trick has been working perfectly in pyiron_workflow to avoid the macos failures. As an aside, there's also enough flexibility in the centralized CI now to specify a particular osx version.

@jan-janssen
Copy link
Member

I feel we might miss the point when the latest images get fixed. So I am wondering if we just want to add both the latest images and the last known working images.

@liamhuber
Copy link
Member Author

I would recommend doing latest as a separate cron job then, so it doesn't keep giving failures to the PRs

I'll remove this once the test is behaving as expected. I just can't do the manual trigger on a workflow that's only on a branch
It worked just fine with the macos-11 tests, so there is nothing wrong with the workflow file.
@liamhuber
Copy link
Member Author

@jan-janssen I added a weekly cron job to try the openmpi tests on macos-latest. In an intermediate state I had it run on: pull-request and with macos-11 and it worked perfectly.

@jan-janssen jan-janssen merged commit 0d54d56 into main Dec 8, 2023
@jan-janssen jan-janssen deleted the test_macos11 branch December 8, 2023 05:26
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants