[
https://issues.apache.org/jira/browse/BEAM-5467?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16675253#comment-16675253
]
Thomas Weise commented on BEAM-5467:
------------------------------------
I did a little more investigation here:
[https://github.com/apache/beam/pull/6937]
The seg faults happen after the test run is complete (regardless if there are
test failures or not). Looks like the Python process fails just before it exits.
The specific test case failure is indeed a separate issue, but also
non-deterministic.
{code:java}
======================================================================
ERROR: test_pardo_unfusable_side_inputs (__main__.FlinkRunnerTest)
----------------------------------------------------------------------
Traceback (most recent call last):
File "apache_beam/runners/portability/fn_api_runner_test.py", line 254, in
test_pardo_unfusable_side_inputs
equal_to([('a', 'a'), ('a', 'b'), ('b', 'a'), ('b', 'b')]))
File "apache_beam/pipeline.py", line 425, in __exit__
self.run().wait_until_finish()
==================== Timed out after 30 seconds. ===================={code}
> Python Flink ValidatesRunner job issues
> ---------------------------------------
>
> Key: BEAM-5467
> URL: https://issues.apache.org/jira/browse/BEAM-5467
> Project: Beam
> Issue Type: Improvement
> Components: runner-flink
> Reporter: Thomas Weise
> Priority: Minor
> Labels: portability-flink
> Time Spent: 15h 20m
> Remaining Estimate: 0h
>
>
> The job is currently flaky and does not provide a useful signal.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
