[IMPALA-9354] The E2E test test_concurrent_create_kudu_table seems flaky

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Duplicate
• Labels: broken-build
• Target Version: Impala 3.4.0
• Epic Color: ghx-label-10
• Assignee: Quanlong Huang
• Reporter: Fang-Yu Rao
• Created: 02/Feb/20 07:14
• Updated: 03/Feb/20 02:21
• Resolved: 03/Feb/20 02:21

Description

The E2E test test_concurrent_create_kudu_table seems flaky when ERASURE_CODING is set to true. This test was recently added in the patch for https://issues.apache.org/jira/browse/IMPALA-8984. According to the error message, it seems that some error occurred when https://github.com/apache/impala/blob/master/tests/custom_cluster/test_concurrent_kudu_create.py#L62 is being called. The stacktrace is also provided in the following.

Stacktrace
custom_cluster/test_concurrent_kudu_create.py:62: in test_concurrent_create_kudu_table
    r3.get()
/usr/lib64/python2.7/multiprocessing/pool.py:554: in get
    raise self._value
E   ImpalaBeeswaxException: ImpalaBeeswaxException:
E    Not connected

In addition, according to the line in the console output, the error occurred at 20:06:47.

20:06:47 custom_cluster/test_concurrent_kudu_create.py::TestConcurrentKuduCreate::test_concurrent_create_kudu_table FAILED

The related 4 log files when this error occurred are also attached. Please refer to them for further details.

Since stigahuang and vihangk1 reviewed the patch for https://issues.apache.org/jira/browse/IMPALA-8984, maybe you could provide some insight into this issue? The JIRA is assigned to stigahuang for now, but please re-assign it to others as appropriate. Thanks!

Issue Links

• duplicates: IMPALA-9289 TestConcurrentKuduCreate is flaky