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 2020/12/15 20:28:05 UTC
[GitHub] [beam] chamikaramj commented on a change in pull request #13536: [BEAM-11360] Updates Dataflow Python multi-language pipelines to use portable job submission by default
chamikaramj commented on a change in pull request #13536:
URL: https://github.com/apache/beam/pull/13536#discussion_r543659576
##########
File path: sdks/python/apache_beam/pipeline.py
##########
@@ -509,19 +514,25 @@ def run(self, test_runner_api='AUTO'):
self.runner.__class__.__name__ == 'SwitchingDirectRunner' and
not self._options.view_as(StandardOptions).streaming)
+ # Multi-language pipelines that contain external pipeline segments may
+ # not be able to create a Python pipeline object graph. Hence following
+ # runner API check should be skipped for such pipelines.
+ external_transform_finder = ExternalTransformFinder()
+ self.visit(external_transform_finder)
Review comment:
This was redundant. Updated.
Also, updated the visitor class by adding a static method like you mentioned.
##########
File path: sdks/python/apache_beam/runners/dataflow/dataflow_runner.py
##########
@@ -580,6 +533,11 @@ def run_pipeline(self, pipeline, options):
debug_options.add_experiment(
'min_cpu_platform=' + worker_options.min_cpu_platform)
+ if pipeline.contains_external_transforms:
+ # All Dataflow multi-language pipelines use portable job submission by
+ # default.
+ debug_options.add_experiment("use_portable_job_submission")
Review comment:
Yes, updated to also include a Runner v2 check.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org