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-2419) Risk of counter over-count when recovering commit log
Date Tue, 05 Apr 2011 21:53:05 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13016142#comment-13016142
] 

Sylvain Lebresne commented on CASSANDRA-2419:
---------------------------------------------

Actually I don't think this really solves the race condition. We really shouldn't compact
the newly flushed sstable until we marked the commit log, because if we compact it, even if
we're able to detect that 'some parts' of a sstable will be replay during recover, there is
nothing we can do about it.

> Risk of counter over-count when recovering commit log
> -----------------------------------------------------
>
>                 Key: CASSANDRA-2419
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2419
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 0.8
>            Reporter: Sylvain Lebresne
>            Assignee: Sylvain Lebresne
>              Labels: counters
>   Original Estimate: 8h
>  Remaining Estimate: 8h
>
> When a memtable was flush, there is a small delay before the commit log replay position
gets updated. If the node fails during this delay, all the updates of this memtable will be
replay during commit log recovery and will end-up being over-counts.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message