hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lipcon (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-4029) GenerationStamp should use an AtomicLong
Date Tue, 16 Oct 2012 21:25:03 GMT

    [ https://issues.apache.org/jira/browse/HDFS-4029?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13477363#comment-13477363
] 

Todd Lipcon commented on HDFS-4029:
-----------------------------------

+1, this looks reasonable to me.
                
> GenerationStamp should use an AtomicLong 
> -----------------------------------------
>
>                 Key: HDFS-4029
>                 URL: https://issues.apache.org/jira/browse/HDFS-4029
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: name-node
>    Affects Versions: 2.0.0-alpha
>            Reporter: Eli Collins
>            Assignee: Eli Collins
>         Attachments: hdfs-4029.txt
>
>
> GenerationStamp#genstamp is a volatile long. Incrementing volatiles is not thread safe,
which is currently not an issue because the only increment is done in nextStamp which is synchronized
(theoretically setStamp could race with nextStamp though that wouldn't happen given the current
code). It would be cleaner to use an AtomicLong so that genstamp itself is atomic and we don't
have to synchronize the various accesses to it.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message