[BEAM-5777] Running ParDo in loop with DirectRunners raises RuntimeException

• Type: Bug
• Status: Open
• Priority: P3
• Resolution: Unresolved
• Component/s: testing
• Assignee: Unassigned
• Reporter: Kasia Kucharczyk
• Created: 17/Oct/18 12:11
• Updated: 03/Jun/22 22:50

Description

The Python load test of ParDo operation for SyntheticSources that I created contains parametrized loop of ParDo with no operation inside besides metrics (this issue). With setting the number of iterations to >~200 and running the test on DirectRunner I was encountering test failures. The test outputs whole (really long) pipeline logs. Some test runs raised the following exception:

 

Traceback (most recent call last):

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/testing/load_tests/par_do_test.py", line 144, in testParDo

    result = p.run()

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/testing/test_pipeline.py", line 104, in run

    result = super(TestPipeline, self).run(test_runner_api)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/pipeline.py", line 403, in run

    self.to_runner_api(), self.runner, self._options).run(False)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/pipeline.py", line 416, in run

    return self.runner.run_pipeline(self)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/direct/direct_runner.py", line 139, in run_pipeline

    return runner.run_pipeline(pipeline)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/portability/fn_api_runner.py", line 229, in run_pipeline

    return self.run_via_runner_api(pipeline.to_runner_api())

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/portability/fn_api_runner.py", line 232, in run_via_runner_api

    return self.run_stages(*self.create_stages(pipeline_proto))

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/portability/fn_api_runner.py", line 1015, in run_stages

    pcoll_buffers, safe_coders).process_bundle.metrics

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/portability/fn_api_runner.py", line 1132, in run_stage

    self._progress_frequency).process_bundle(data_input, data_output)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/portability/fn_api_runner.py", line 1388, in process_bundle

    result_future = self._controller.control_handler.push(process_bundle)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/portability/fn_api_runner.py", line 1260, in push

    response = self.worker.do_instruction(request)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/worker/sdk_worker.py", line 212, in do_instruction

    request.instruction_id)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/worker/sdk_worker.py", line 231, in process_bundle

    self.data_channel_factory)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/worker/bundle_processor.py", line 343, in __init__

    self.ops = self.create_execution_tree(self.process_bundle_descriptor)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/worker/bundle_processor.py", line 385, in create_execution_tree

    descriptor.transforms, key=topological_height, reverse=True)])

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/worker/bundle_processor.py", line 320, in wrapper

    result = cache[args] = func(*args)

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/worker/bundle_processor.py", line 368, in get_operation

    in descriptor.transforms[transform_id].outputs.items()

  File "/Users/kasia/Repos/beam/sdks/python/apache_beam/runners/worker/bundle_processor.py", line 367, in <dictcomp>

    for tag, pcoll_id

... (3 last lines repeated for long period)

 
RuntimeError: maximum recursion depth exceeded

 

 

From my observation, I can say the problem appeared with various iteration number depending on computer resources. On my weaker computer started failing on ~150 iterations. The test succeeds on DataFlow with 1000 iterations (I didn't check higher number).

I provide whole test output in Attachements.

Issue Links

• is related to: BEAM-5758 Load tests for SyntheticSources in Python