You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-issues@hadoop.apache.org by "Harsh J (Commented) (JIRA)" <ji...@apache.org> on 2012/01/24 10:40:40 UTC

[jira] [Commented] (MAPREDUCE-3715) FileSplit's Writable implementation discards hosts information

    [ https://issues.apache.org/jira/browse/MAPREDUCE-3715?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13192038#comment-13192038 ] 

Harsh J commented on MAPREDUCE-3715:
------------------------------------

We store out the locations separately during submission anyway, so I guess this was done for saving the size of metadata written during submission. It has been this way since 0.18.

We can document it, but were you looking for location info from inside a record reader? How were you planning to use it?
                
> FileSplit's Writable implementation discards hosts information
> --------------------------------------------------------------
>
>                 Key: MAPREDUCE-3715
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-3715
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>    Affects Versions: 0.20.2, 1.0.0
>            Reporter: Pavel Podgoretsky
>            Priority: Minor
>
> org.apache.hadoop.mapreduce.lib.input.FileSplit's write() does not write out hosts, and readFields() sets them to null.
> If this is intended then it should be documented very clearly.  But it looks more like an oversight to me.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira