Zoltan Borok-Nagy has posted comments on this change. ( http://gerrit.cloudera.org:8080/17508 )
Change subject: IMPALA-10713: Use PARTITION-level locking for static partition INSERTs for ACID tables ...................................................................... Patch Set 2: (2 comments) http://gerrit.cloudera.org:8080/#/c/17508/1//COMMIT_MSG Commit Message: http://gerrit.cloudera.org:8080/#/c/17508/1//COMMIT_MSG@9 PS1, Line 9: create > nit: 'creates' Done http://gerrit.cloudera.org:8080/#/c/17508/2/tests/stress/test_acid_stress.py File tests/stress/test_acid_stress.py: http://gerrit.cloudera.org:8080/#/c/17508/2/tests/stress/test_acid_stress.py@215 PS2, Line 215: 3 > How are these duration values chosen ? Could it lead to non-deterministic t For some reason we invoke sleep() twice during an INSERT statement. Once for UnionNode::MaterializeExprs(), and once for the HDFS table sink. This means a single INSERT INTO ... SELECT sleep(5000) takes at least 2 * 5 sec. This test is limited to HDFS, so I think 3 * 5 sec should be enough on a healthy cluster. In my local dev env it usually takes ~10.6s. -- To view, visit http://gerrit.cloudera.org:8080/17508 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I0f5f998b908d836567c9bf7498d2caa0d231df35 Gerrit-Change-Number: 17508 Gerrit-PatchSet: 2 Gerrit-Owner: Zoltan Borok-Nagy <[email protected]> Gerrit-Reviewer: Aman Sinha <[email protected]> Gerrit-Reviewer: Csaba Ringhofer <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Zoltan Borok-Nagy <[email protected]> Gerrit-Comment-Date: Mon, 31 May 2021 09:29:31 +0000 Gerrit-HasComments: Yes
