cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-11016) Fix flapping dtests with "unavailable" and "cannot achieve consistency" errors
Date Fri, 15 Jan 2016 16:02:39 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-11016?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15101972#comment-15101972
] 

Sylvain Lebresne commented on CASSANDRA-11016:
----------------------------------------------

I'll note that the truncate failures in the upgrade test are pretty easy to fix, see https://github.com/riptano/cassandra-dtest/pull/749
(I haven't checked this fix it tbh honest, but it's almost surely the problem).

Haven't looked at the failure of replace_first_boot_test but from the error message that looks
like a similar problem (the test not ensure a restart node has been detected by other ones
when it should).

> Fix flapping dtests with "unavailable" and "cannot achieve consistency" errors
> ------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-11016
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-11016
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Jim Witschey
>
> A number of dtests flap when they run a query and get an unavailable exception or can't
meet a particular CL. Examples include:
> http://cassci.datastax.com/job/cassandra-2.2_dtest/470/testReport/replace_address_test/TestReplaceAddress/replace_first_boot_test/
> http://cassci.datastax.com/view/cassandra-3.0/job/cassandra-3.0_dtest/488/testReport/upgrade_tests.cql_tests/TestCQLNodes3RF3/range_tombstones_test/
> This seems to be a different issue than CASSANDRA-10730 -- bumping instance sizes to
m3.2xl stopped that class of error, but bumping instance sizes to i2.4xl did not stop the
errors described here.
> I'm probably the best person to tackle this, but won't get to it for a bit, so [~philipthompson]
might take it if he has time sooner.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message