[
https://issues.apache.org/jira/browse/BEAM-6280?focusedWorklogId=183882&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-183882
]
ASF GitHub Bot logged work on BEAM-6280:
----------------------------------------
Author: ASF GitHub Bot
Created on: 10/Jan/19 19:35
Start Date: 10/Jan/19 19:35
Worklog Time Spent: 10m
Work Description: robertwb commented on pull request #7433: [BEAM-6280]
Refactors Python portability tests to be multi-threaded aware
URL: https://github.com/apache/beam/pull/7433#discussion_r246893622
##########
File path: sdks/python/apache_beam/runners/portability/local_job_service.py
##########
@@ -185,55 +152,89 @@ class BeamJob(threading.Thread):
The current state of the pipeline is available as self.state.
"""
- def __init__(self,
- job_id,
- pipeline_options,
- pipeline_proto):
+ def __init__(self, pipeline_proto):
super(BeamJob, self).__init__()
- self._job_id = job_id
- self._pipeline_options = pipeline_options
self._pipeline_proto = pipeline_proto
self._state = None
- self._state_change_callbacks = []
- self._last_log_message = None
- self._log_callbacks = [lambda msg: setattr(self, '_last_log_message', msg)]
+ self._logs = []
Review comment:
I still don't think that we should be storing all logs. What if we allowed
connecting *before* the job is started to ensure all logs were seen. Otherwise,
one gets the logs from the point at which one connected (which better matches
the intention).
Similarly for states, as long as one gets a response for the current state,
there's no need to replay all past states. In this case we would go back to the
callback logic, but maybe remove the last_log_message which is the fishy one,
and fix the runner to start listening to logs before it starts the job (which
we'll want to do for other jobs servers as well).
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 183882)
Time Spent: 2h (was: 1h 50m)
> Failure in PortableRunnerTest.test_error_traceback_includes_user_code
> ---------------------------------------------------------------------
>
> Key: BEAM-6280
> URL: https://issues.apache.org/jira/browse/BEAM-6280
> Project: Beam
> Issue Type: Bug
> Components: sdk-py-core, test-failures
> Reporter: Kenneth Knowles
> Assignee: Sam Rohde
> Priority: Critical
> Labels: flaky-test
> Time Spent: 2h
> Remaining Estimate: 0h
>
> [https://builds.apache.org/job/beam_PreCommit_Python_Cron/732/]
> [https://builds.apache.org/job/beam_PreCommit_Python_Cron/732/testReport/apache_beam.runners.portability.portable_runner_test/PortableRunnerTest/test_error_traceback_includes_user_code/]
> [https://scans.gradle.com/s/do3hjulee3gaa/console-log?task=:beam-sdks-python:testPython3]
> {code:java}
> 'second' not found in 'Traceback (most recent call last):\n File
> "/home/jenkins/jenkins-slave/workspace/beam_PreCommit_Python_Cron/src/sdks/python/apache_beam/runners/portability/fn_api_runner_test.py",
> line 466, in test_error_traceback_includes_user_code\n p |
> beam.Create([0]) | beam.Map(first) # pylint:
> disable=expression-not-assigned\n File
> "/home/jenkins/jenkins-slave/workspace/beam_PreCommit_Python_Cron/src/sdks/python/apache_beam/pipeline.py",
> line 425, in __exit__\n self.run().wait_until_finish()\n File
> "/home/jenkins/jenkins-slave/workspace/beam_PreCommit_Python_Cron/src/sdks/python/apache_beam/runners/portability/portable_runner.py",
> line 314, in wait_until_finish\n self._job_id, self._state,
> self._last_error_message()))\nRuntimeError: Pipeline
> job-cdcefe6d-1caa-4487-9e63-e971f67ec68c failed in state FAILED: start
> <DataInputOperation receivers=[ConsumerSet[Create/Read/Impulse.out0,
> coder=WindowedValueCoder[BytesCoder], len(consumers)=1]]>\n'{code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
