[CASSANDRA-10642] cqlsh COPY bulk round trip dtest flaps on Windows

• Type: Sub-task
• Status: Resolved
• Priority: Normal
• Resolution: Duplicate
• Component/s: Legacy/Tools, Test/dtest/python
• Assignee: Stefania Alborghetti
• Reporter: Jim Witschey

Authors

Stefania Alborghetti

• Created: 03/Nov/15 17:38
• Updated: 16/Apr/19 09:30
• Resolved: 07/Dec/15 09:17

Description

cqlsh_tests/cqlsh_copy_tests.py:CqlshCopyTest.test_bulk_round_trip flaps on Windows under both C* 3.0 and 2.2:

http://cassci.datastax.com/view/win32/job/cassandra-3.0_dtest_win32/100/testReport/junit/cqlsh_tests.cqlsh_copy_tests/CqlshCopyTest/test_bulk_round_trip/history/
http://cassci.datastax.com/view/win32/job/cassandra-2.2_dtest_win32/127/testReport/junit/cqlsh_tests.cqlsh_copy_tests/CqlshCopyTest/test_bulk_round_trip/

It fails because, after round-tripping with cqlsh COPY, it fails to find as many values as expected with SELECT COUNT (*). The stderr from cqlsh includes the following error:

(EE)  (EE)  <stdin>:2:(EE)  <ErrorMessage code=1100 [Coordinator node timed out waiting for replica nodes' responses] message="Operation timed out - received only 0 responses.">(EE)  <stdin>:2:Aborting import at record #190. Previously inserted records are still present, and some records after that may be present as well.(EE)  <stdin>:2:(EE)  <ErrorMessage code=1000 [Unavailable exception] message="Cannot achieve consistency level ONE">(EE)  <stdin>:2:Aborting import at record #637. Previously inserted records are still present, and some records after that may be present as well.(EE)  

So, it looks like the load step may be timing out. See this run for the error:

http://cassci.datastax.com/view/win32/job/cassandra-3.0_dtest_win32/100/testReport/junit/cqlsh_tests.cqlsh_copy_tests/CqlshCopyTest/test_bulk_round_trip/

Issue Links

• is blocked by: CASSANDRA-9302 Optimize cqlsh COPY FROM, part 3
• is duplicated by: CASSANDRA-9302 Optimize cqlsh COPY FROM, part 3