[
https://issues.apache.org/jira/browse/IMPALA-9766?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17112768#comment-17112768
]
ASF subversion and git services commented on IMPALA-9766:
---------------------------------------------------------
Commit 29591eb7ad5a31c9cff201550d257aa74ce09013 in impala's branch
refs/heads/master from Tim Armstrong
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=29591eb ]
IMPALA-9766: make TestParquet tests deterministic
test_page_size_counters and test_bytes_read_per_column were
written in a way so that queries were short-circuited by
the limit and could terminate early before the scans made
much progress. Add an order by to make the queries deterministic.
Testing:
Looped the tests locally for ~10 minutes. Before the change
the issue reproduced very quickly.
Change-Id: I49ac7895542e4f797b16a71b3d24abb34e282e67
Reviewed-on: http://gerrit.cloudera.org:8080/15967
Reviewed-by: Tim Armstrong <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> TestParquet.test_bytes_read_per_column is flaky after IMPALA-6984
> -----------------------------------------------------------------
>
> Key: IMPALA-9766
> URL: https://issues.apache.org/jira/browse/IMPALA-9766
> Project: IMPALA
> Issue Type: Improvement
> Components: Infrastructure
> Reporter: Tim Armstrong
> Assignee: Tim Armstrong
> Priority: Major
> Labels: flaky
>
> The test matches the pattern of problematic queries for IMPALA-6984 - it has
> a non-deterministic select and reads profile counters.
> {noformat}
> query_test.test_scanners.TestParquet.test_bytes_read_per_column[protocol:
> beeswax | exec_option: {'batch_size': 0, 'num_nodes': 0,
> 'disable_codegen_rows_threshold': 0, 'disable_codegen': False,
> 'abort_on_error': 1, 'debug_action':
> '-1:OPEN:[email protected]',
> 'exec_single_node_rows_threshold': 0} | table_format: parquet/none] (from
> pytest)
> Failing for the past 1 build (Since Failed#7180 )
> Took 87 ms.
> add description
> Error Message
> query_test/test_scanners.py:928: in test_bytes_read_per_column assert not
> self._is_summary_stats_counter_empty(summary) E assert not True E +
> where True = <bound method TestParquet._is_summary_stats_counter_empty of
> <test_scanners.TestParquet object at
> 0x7f65204ed810>>(TSummaryStatsCounter(total_num_values=0, name=None,
> max_value=0, sum=0, min_value=0, unit=None)) E + where <bound method
> TestParquet._is_summary_stats_counter_empty of <test_scanners.TestParquet
> object at 0x7f65204ed810>> = <test_scanners.TestParquet object at
> 0x7f65204ed810>._is_summary_stats_counter_empty
> Stacktrace
> query_test/test_scanners.py:928: in test_bytes_read_per_column
> assert not self._is_summary_stats_counter_empty(summary)
> E assert not True
> E + where True = <bound method
> TestParquet._is_summary_stats_counter_empty of <test_scanners.TestParquet
> object at 0x7f65204ed810>>(TSummaryStatsCounter(total_num_values=0,
> name=None, max_value=0, sum=0, min_value=0, unit=None))
> E + where <bound method TestParquet._is_summary_stats_counter_empty of
> <test_scanners.TestParquet object at 0x7f65204ed810>> =
> <test_scanners.TestParquet object at
> 0x7f65204ed810>._is_summary_stats_counter_empty
> {noformat}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
