You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@beam.apache.org by GitBox <gi...@apache.org> on 2021/06/30 18:31:46 UTC

[GitHub] [beam] TheNeuralBit commented on a change in pull request #15104: [BEAM-12515] Mark test_display_data as no_xdist

TheNeuralBit commented on a change in pull request #15104:
URL: https://github.com/apache/beam/pull/15104#discussion_r661718852



##########
File path: sdks/python/apache_beam/options/pipeline_options_test.py
##########
@@ -215,6 +216,7 @@ def _add_argparse_args(cls, parser):
       parser.add_argument(
           '--fake_multi_option', action='append', help='fake multi option')
 
+  @pytest.mark.no_xdist

Review comment:
       I think it's very unlikely that parallelization will cause flakiness due to some shared state and a race condition. xdist works by starting up multiple separate python processes each running a partition of the tests. The GIL would eliminate any benefit in a threaded approach.
   
   My guess would be that xdist makes this flaky by sometimes executing some other test that modifies state in the same python process/worker. I'm not sure what exactly no_xdist does, presumably it just causes the test to be run in the main process? It makes sense that that would help since that problematic test is then guaranteed to run in a separate process. 




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@beam.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org