[
https://issues.apache.org/jira/browse/IMPALA-13893?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17941790#comment-17941790
]
ASF subversion and git services commented on IMPALA-13893:
----------------------------------------------------------
Commit bad56af62c66c9a717d5d41d83e7e5ec74a1836a in impala's branch
refs/heads/master from Yida Wu
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=bad56af62 ]
IMPALA-13893: Add node id to the file name in tuple cache correctness
verification
Currently, the file name for tuple cache correctness verification
was generated using only the fragment instance id. However, in
some cases, multiple tuple cache nodes within the same fragment
instance may share the same cache key, leading to conflicts
during file comparison.
This patch resolves the issue by appending the node id to the file
name, make sure it is unique among nodes within the same fragment
instance.
Old Format:
file: {fragment_instance_id}
Ref file: {fragment_instance_id}_{org_fragment_instance_id}_ref
New Format:
file: {fragment_instance_id}_{node_id}
Ref file: {fragment_instance_id}_{node_id}
_{org_fragment_instance_id}_{node_id}_ref
Tests:
Passed query_test.test_queries.TestQueriesTextTables.test_random,
which previously failed due to this issue.
Change-Id: I6ce5ed05623761ace7b7a1027b762736a63e97fc
Reviewed-on: http://gerrit.cloudera.org:8080/22671
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> Tuple caching correctness verification fails for
> query_test.test_queries.TestQueriesTextTables.test_random
> ----------------------------------------------------------------------------------------------------------
>
> Key: IMPALA-13893
> URL: https://issues.apache.org/jira/browse/IMPALA-13893
> Project: IMPALA
> Issue Type: Bug
> Components: Backend
> Affects Versions: Impala 5.0.0
> Reporter: Joe McDonnell
> Assignee: Yida Wu
> Priority: Major
>
> When running with tuple cache correctness verification,
> query_test.test_queries.TestQueriesTextTables.test_random fails with the
> following error:
> {noformat}
> query_test/test_queries.py:256: in test_random
> self.run_test_case('QueryTest/random', vector)
> common/impala_test_suite.py:874: in run_test_case
> result = exec_fn(query, user=test_section.get('USER', '').strip() or None)
> common/impala_test_suite.py:804: in __exec_in_impala
> result = self.__execute_query(target_impalad_client, query, user=user)
> common/impala_test_suite.py:1286: in __execute_query
> return impalad_client.execute(query, user=user)
> common/impala_connection.py:462: in execute
> fetch_profile_after_close=fetch_profile_after_close)
> beeswax/impala_beeswax.py:193: in execute
> handle = self.__execute_query(query_string.strip(), user=user)
> beeswax/impala_beeswax.py:289: in __execute_query
> self.wait_for_finished(handle)
> beeswax/impala_beeswax.py:315: in wait_for_finished
> raise ImpalaBeeswaxException(error_log, None)
> E ImpalaBeeswaxException: Query 534df8c1bbfdca0e:57bc605100000000 failed:
> E Inconsistent tuple cache found: Failed to open file
> '/data/jenkins/workspace/tmp/impala-tuplecache-debugdump-0/tuple-cache-debug-dump/c5ee5de51167923509373f3adfd81a17_1709992867/534df8c1bbfdca0e:57bc605100000000_bb4715c6f083597c:0cbf03f200000000_ref.bad'.{noformat}
> What's happening is that two tuple caching locations have the same cache key
> because they are scanning the same table. Tuple caching correctness
> verification fails because one location has already removed locations when
> the other wants to do correctness verification. One way to fix this is to
> incorporate the plan node id into the filenames.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
