cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Benedict (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (CASSANDRA-8744) Ensure SSTableReader.first/last are honoured universally
Date Mon, 09 Feb 2015 10:19:34 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-8744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14309389#comment-14309389
] 

Benedict edited comment on CASSANDRA-8744 at 2/9/15 10:19 AM:
--------------------------------------------------------------

Without this, CASSANDRA-8683 is not really fixed, and you should find at the very least LeveledCompactionStrategyTest.testValidationMultipleSSTablePerLevel
fails periodically. What other implications there are, I'm not sure.


was (Author: benedict):
Without this, CASSANDRA-6863 is not fixed, and you should find at the very least LeveledCompactionStrategyTest.testValidationMultipleSSTablePerLevel
fails periodically. What other implications there are, I'm not sure.

> Ensure SSTableReader.first/last are honoured universally
> --------------------------------------------------------
>
>                 Key: CASSANDRA-8744
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-8744
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>            Reporter: Benedict
>            Assignee: Benedict
>            Priority: Critical
>             Fix For: 2.1.4
>
>
> Split out from CASSANDRA-8683; we don't honour the first/last properties of an sstablereader,
and we tend to assume that we do. This can cause problems in LCS validation compactions, for
instance, where a scanner is assumed to only cover the defined range, but may return data
either side of that range. In general it is only wasteful to not honour these ranges anyway.



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

Mime
View raw message