Skip to content

Conversation

junaruga
Copy link
Member

Note that the default configure option ./configure cppflags=-DRUBY_DEBUG
causes the following Ractor test and other tests failing. So, we don't set the
option in ppc64le cases.

$ make btest BTESTS=bootstraptest/test_ractor.rb

You can check the following ticket for details.
https://bugs.ruby-lang.org/issues/21534

@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch from 0babe52 to 8fe3709 Compare August 15, 2025 15:25
@junaruga
Copy link
Member Author

junaruga commented Aug 15, 2025

The GitHub Actions ppc64le sometimes delays to start the job. In my latest case, I need to wait for the job to start more than 25 minutes. This is a blocker to merge this PR in my opinion. I reported the issue IBM/actionspz#34.

@junaruga
Copy link
Member Author

junaruga commented Aug 15, 2025

Some tests in the make test-all failed with timeout on the ppc64le case below.
https://github.com/ruby/ruby/actions/runs/16993265630/job/48177676417?pr=14222#step:16:1446

@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch 4 times, most recently from 5181a1f to c5795eb Compare August 20, 2025 19:46
@junaruga
Copy link
Member Author

junaruga commented Aug 21, 2025

I am seeing many timeout related failures, while the the tests passe on RubyCI Ubuntu (ppc64le) server. I suspect the reason of the failures is a CI machine's slow performance. So, I added the logic to measure the performance of the make check by mpstat, vmstat, and iostat commands on the latest commit in this PR.

@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch 4 times, most recently from 97bf09a to 603e5ae Compare August 21, 2025 17:40
@junaruga
Copy link
Member Author

I reported the GitHub Actions ppc64le (and s390x) performance issue IBM/actionspz#38 coming from the wrong CPU settings in the containers.

@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch 5 times, most recently from 80efe5a to b437998 Compare August 25, 2025 19:17
@junaruga junaruga marked this pull request as ready for review August 25, 2025 20:42
@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch from b437998 to d3a0ca2 Compare August 25, 2025 21:44
@junaruga
Copy link
Member Author

I found the root cause IBM/actionspz#38 (comment).

@junaruga junaruga marked this pull request as draft August 26, 2025 08:21
@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch 2 times, most recently from ede819f to 323cfea Compare August 26, 2025 15:20
@junaruga
Copy link
Member Author

junaruga commented Aug 26, 2025

As the issue IBM/actionspz#34 still blocks to add GitHub Actions ppc64le to ruby/ruby, I sent the #14345 to merge only the 1st commit related to the nproc.

@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch 3 times, most recently from 570eee4 to c617b31 Compare August 27, 2025 17:19
@junaruga junaruga marked this pull request as ready for review August 27, 2025 23:34
@junaruga junaruga changed the title CI: ubuntu.yml: Add GitHub Actions ppc64le case CI: ubuntu-ibm.yml: Add GitHub Actions ppc64le case Aug 27, 2025
@junaruga
Copy link
Member Author

People at IBM said they they fixed the issue IBM/actionspz#34 that jobs sometimes not starting, and I confirmed the GitHub Actions ppc64le/390x jobs started without delay on this PR's latest CI. So, let's merge this PR adding the s390x and ppc64le, and we will see how these work.

@junaruga junaruga marked this pull request as draft August 28, 2025 09:40
@junaruga
Copy link
Member Author

I see my fork repository's CI: https://github.com/junaruga/ruby/actions/runs/17273836937 is failing because there is not cases in the .github/workflows/ubuntu-ibm.yml. I want to make the CI green in the fork repository case too.

@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch 2 times, most recently from 20b9a15 to 42b4b2b Compare August 28, 2025 11:30
* Remove logic that was used for the ubuntu.yml, but not used for
  ubuntu-ibm.yml.
* Add a dummy Ubuntu x86_64 case to make this CI pass on fork repositories.
  This case only runs on fork repositories.
Note that the default configure option `./configure cppflags=-DRUBY_DEBUG` with
the default optimization level `-O3`, causes the following Ractor test and other
tests failing. So, we don't set the option in ppc64le case.

```
$ make btest BTESTS=bootstraptest/test_ractor.rb
```

See https://bugs.ruby-lang.org/issues/21534 for details.
@junaruga junaruga force-pushed the wip/gha-add-ppc64le branch from 42b4b2b to afca9c4 Compare August 28, 2025 11:43
@junaruga junaruga marked this pull request as ready for review August 28, 2025 11:43
@junaruga junaruga merged commit 4fc0db7 into ruby:master Aug 28, 2025
87 checks passed
@junaruga junaruga deleted the wip/gha-add-ppc64le branch August 28, 2025 13:55
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.

1 participant