cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Branimir Lambov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-12917) Increase error margin in SplitterTest
Date Mon, 13 Feb 2017 13:49:41 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-12917?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15863702#comment-15863702
] 

Branimir Lambov commented on CASSANDRA-12917:
---------------------------------------------

Code looks good.

However, raising the margin decreases the flakiness, but I'm not sure it eliminates it. In
such cases I personally prefer to have a large number of points in the individual test (which
this already has), a fixed seed, and a lower error margin.

> Increase error margin in SplitterTest
> -------------------------------------
>
>                 Key: CASSANDRA-12917
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-12917
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Marcus Eriksson
>            Assignee: Marcus Eriksson
>              Labels: lhf
>             Fix For: 3.11.x
>
>
> SplitterTest is a randomized test - it generates random tokens and splits the ranges
in equal parts. Since it is random we sometimes get very big vnodes right where we want a
split and that makes the split unbalanced
> Bumping the error margin a bit will avoid these false positives.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message