[
https://issues.apache.org/jira/browse/CASSANDRA-12092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15432191#comment-15432191
]
Stefania commented on CASSANDRA-12092:
--------------------------------------
I traced the remaining rare failures down to a speculative retry. It was not
the async repair digest mismatch that caused problems, rather in some rare
cases a speculative retry (which is sent after a sleep period) arrives before
the local read. I've disabled speculative retry and I got two 5x200 builds
without any failures, running a modified test that repeated QUORUM/ONE 10
times, see
[here|http://cassci.datastax.com/view/Dev/view/stef1927/job/stef1927-dtest-multiplex/26/]
and
[here|http://cassci.datastax.com/view/Dev/view/stef1927/job/stef1927-dtest-multiplex/28/].
[Pull request|https://github.com/riptano/cassandra-dtest/pull/1269].
> dtest failure in consistency_test.TestAccuracy.test_simple_strategy_counters
> ----------------------------------------------------------------------------
>
> Key: CASSANDRA-12092
> URL: https://issues.apache.org/jira/browse/CASSANDRA-12092
> Project: Cassandra
> Issue Type: Test
> Reporter: Sean McCarthy
> Assignee: Stefania
> Labels: dtest
> Attachments: jenkins-stef1927-dtest-multiplex-23_logs.000.tar.gz,
> node1.log, node2.log, node3.log
>
>
> example failure:
> http://cassci.datastax.com/job/cassandra-2.1_dtest/484/testReport/consistency_test/TestAccuracy/test_simple_strategy_counters
> Failed on CassCI build cassandra-2.1_dtest #484
> {code}
> Standard Error
> Traceback (most recent call last):
> File "/home/automaton/cassandra-dtest/consistency_test.py", line 514, in run
> valid_fcn(v)
> File "/home/automaton/cassandra-dtest/consistency_test.py", line 497, in
> validate_counters
> check_all_sessions(s, n, c)
> File "/home/automaton/cassandra-dtest/consistency_test.py", line 490, in
> check_all_sessions
> "value of %s at key %d, instead got these values: %s" % (write_nodes,
> val, n, results)
> AssertionError: Failed to read value from sufficient number of nodes,
> required 2 nodes to have a counter value of 1 at key 200, instead got these
> values: [0, 0, 1]
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
