[
https://issues.apache.org/jira/browse/CASSANDRA-11341?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15191060#comment-15191060
]
Jim Witschey commented on CASSANDRA-11341:
------------------------------------------
The more I read this, the more I think it actually looks pretty bad? What's
happening here is
* value stored at RF=3 is written at CL=3 over an exclusive connection
* value is repeatedly written and read at CL=1 over the same exclusive
connection
* value [is read again
here|https://github.com/riptano/cassandra-dtest/blob/master/upgrade_tests/cql_tests.py#L4294]
and we get back {{None}} from the driver
I don't know how we can chalk this up to environmental instability -- it's not
like the request times out, or we get a value that would indicate we're getting
a value written to a different node. First off, we're using an exclusive
connection, so we shouldn't be getting values from other nodes. Second, the
value written to the cluster should be present on all nodes anyway. Third, we
just wrote it to the current node and got a 'success' response. I know clusters
with multiple nodes on the same machine are sort of a degenerate environment,
but I don't see why this would cause such incorrect behavior.
> dtest failure in
> upgrade_tests.cql_tests.TestCQLNodes3RF3_2_1_HEAD_UpTo_2_2.whole_list_conditional_test
> -------------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-11341
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11341
> Project: Cassandra
> Issue Type: Test
> Reporter: Jim Witschey
> Assignee: Russ Hatch
> Labels: dtest
>
> example failure:
> http://cassci.datastax.com/job/upgrade_tests-all/22/testReport/upgrade_tests.cql_tests/TestCQLNodes3RF3_2_1_HEAD_UpTo_2_2/whole_list_conditional_test
> Failed on CassCI build upgrade_tests-all #22
> There's only one flap in the history currently. This was the failure:
> {code}
> Expected [[0, ['foo', 'bar', 'foobar']]] from SELECT * FROM tlist, but got
> [[0, None]]
> -------------------- >> begin captured logging << --------------------
> dtest: DEBUG: cluster ccm directory: /mnt/tmp/dtest-SF2dOV
> dtest: DEBUG: Custom init_config not found. Setting defaults.
> dtest: DEBUG: Done setting configuration options:
> { 'initial_token': None,
> 'num_tokens': '32',
> 'phi_convict_threshold': 5,
> 'range_request_timeout_in_ms': 10000,
> 'read_request_timeout_in_ms': 10000,
> 'request_timeout_in_ms': 10000,
> 'truncate_request_timeout_in_ms': 10000,
> 'write_request_timeout_in_ms': 10000}
> dtest: DEBUG: upgrading node1 to 2.2.5
> dtest: DEBUG: Querying upgraded node
> --------------------- >> end captured logging << ---------------------
> Stacktrace
> File "/usr/lib/python2.7/unittest/case.py", line 329, in run
> testMethod()
> File "/home/automaton/cassandra-dtest/tools.py", line 253, in wrapped
> f(obj)
> File "/home/automaton/cassandra-dtest/upgrade_tests/cql_tests.py", line
> 4294, in whole_list_conditional_test
> check_applies("l != null AND l IN (['foo', 'bar', 'foobar'])")
> File "/home/automaton/cassandra-dtest/upgrade_tests/cql_tests.py", line
> 4282, in check_applies
> assert_one(cursor, "SELECT * FROM %s" % (table,), [0, ['foo', 'bar',
> 'foobar']])
> File "/home/automaton/cassandra-dtest/assertions.py", line 50, in assert_one
> assert list_res == [expected], "Expected %s from %s, but got %s" %
> ([expected], query, list_res)
> "Expected [[0, ['foo', 'bar', 'foobar']]] from SELECT * FROM tlist, but got
> [[0, None]]\n-------------------- >> begin captured logging <<
> --------------------\ndtest: DEBUG: cluster ccm directory:
> /mnt/tmp/dtest-SF2dOV\ndtest: DEBUG: Custom init_config not found. Setting
> defaults.\ndtest: DEBUG: Done setting configuration options:\n{
> 'initial_token': None,\n 'num_tokens': '32',\n 'phi_convict_threshold':
> 5,\n 'range_request_timeout_in_ms': 10000,\n
> 'read_request_timeout_in_ms': 10000,\n 'request_timeout_in_ms': 10000,\n
> 'truncate_request_timeout_in_ms': 10000,\n 'write_request_timeout_in_ms':
> 10000}\ndtest: DEBUG: upgrading node1 to 2.2.5\ndtest: DEBUG: Querying
> upgraded node\n--------------------- >> end captured logging <<
> ---------------------"
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
