[ https://issues.apache.org/jira/browse/IMPALA-7828?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Michael Ho updated IMPALA-7828: ------------------------------- Description: Before IMPALA-4063, the error message detected during {{FragmentInstanceState::Close()}} was always lost. After IMPALA-4063, we may sometimes get the error message in {{FragmentInstanceState::Close()}}. It's non-deterministic as the fragment instance thread may race with the query state thread which reports the final status. The test currently tries to handle this non-determinism by using "row_regex:.*" in the {{--ERRORS--}} section but it doesn't seem to work. Let's unbreak the test for now. Longer run, we need to move the point in which all fragment instances are done after {{FragmentInstanceState::Close()}} so things will become deterministic. This may have implication to query completion time so performance may need to be evaluated. {noformat} Stacktrace query_test/test_udfs.py:313: in test_ir_functions self.run_test_case('QueryTest/udf-no-expr-rewrite', vector, use_db=unique_database) common/impala_test_suite.py:496: in run_test_case self.__verify_results_and_errors(vector, test_section, result, use_db) common/impala_test_suite.py:358: in __verify_results_and_errors replace_filenames_with_placeholder) common/test_result_verifier.py:362: in verify_raw_results verify_errors(expected_errors, actual_errors) common/test_result_verifier.py:314: in verify_errors VERIFIER_MAP['VERIFY_IS_EQUAL'](expected, actual) common/test_result_verifier.py:271: in verify_query_result_is_equal assert expected_results == actual_results E assert Comparing QueryTestResults (expected vs actual): E row_regex:.* != None E Number of rows returned (expected vs actual): 1 != 0 {noformat} was: Before IMPALA-4063, the error message detected during {{FragmentInstanceState::Close()} was always lost. After IMPALA-4063, we may sometimes get the error message in {{FragmentInstanceState::Close()}}. It's non-deterministic as the fragment instance thread may race with the query state thread which reports the final status. The test currently tries to handle this non-determinism by using "row_regex:.*" in the {{--ERRORS--}} section but it doesn't seem to work. Let's unbreak the test for now. Longer run, we need to move the point in which all fragment instances are done after {{FragmentInstanceState::Close()}} so things will become deterministic. This may have implication to query completion time so performance may need to be evaluated. {noformat} Stacktrace query_test/test_udfs.py:313: in test_ir_functions self.run_test_case('QueryTest/udf-no-expr-rewrite', vector, use_db=unique_database) common/impala_test_suite.py:496: in run_test_case self.__verify_results_and_errors(vector, test_section, result, use_db) common/impala_test_suite.py:358: in __verify_results_and_errors replace_filenames_with_placeholder) common/test_result_verifier.py:362: in verify_raw_results verify_errors(expected_errors, actual_errors) common/test_result_verifier.py:314: in verify_errors VERIFIER_MAP['VERIFY_IS_EQUAL'](expected, actual) common/test_result_verifier.py:271: in verify_query_result_is_equal assert expected_results == actual_results E assert Comparing QueryTestResults (expected vs actual): E row_regex:.* != None E Number of rows returned (expected vs actual): 1 != 0 {noformat} > test_mem_leak() is flaky > ------------------------ > > Key: IMPALA-7828 > URL: https://issues.apache.org/jira/browse/IMPALA-7828 > Project: IMPALA > Issue Type: Bug > Components: Backend > Affects Versions: Impala 3.2.0 > Reporter: Michael Ho > Assignee: Michael Ho > Priority: Blocker > Labels: broken-build > > Before IMPALA-4063, the error message detected during > {{FragmentInstanceState::Close()}} was always lost. After IMPALA-4063, we may > sometimes get the error message in {{FragmentInstanceState::Close()}}. It's > non-deterministic as the fragment instance thread may race with the query > state thread which reports the final status. The test currently tries to > handle this non-determinism by using "row_regex:.*" in the {{--ERRORS--}} > section but it doesn't seem to work. > Let's unbreak the test for now. Longer run, we need to move the point in > which all fragment instances are done after > {{FragmentInstanceState::Close()}} so things will become deterministic. This > may have implication to query completion time so performance may need to be > evaluated. > {noformat} > Stacktrace > query_test/test_udfs.py:313: in test_ir_functions > self.run_test_case('QueryTest/udf-no-expr-rewrite', vector, > use_db=unique_database) > common/impala_test_suite.py:496: in run_test_case > self.__verify_results_and_errors(vector, test_section, result, use_db) > common/impala_test_suite.py:358: in __verify_results_and_errors > replace_filenames_with_placeholder) > common/test_result_verifier.py:362: in verify_raw_results > verify_errors(expected_errors, actual_errors) > common/test_result_verifier.py:314: in verify_errors > VERIFIER_MAP['VERIFY_IS_EQUAL'](expected, actual) > common/test_result_verifier.py:271: in verify_query_result_is_equal > assert expected_results == actual_results > E assert Comparing QueryTestResults (expected vs actual): > E row_regex:.* != None > E Number of rows returned (expected vs actual): 1 != 0 > {noformat} -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-all-unsubscr...@impala.apache.org For additional commands, e-mail: issues-all-h...@impala.apache.org