hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Allen Wittenauer (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HADOOP-9022) Hadoop distcp tool fails to copy file if -m 0 specified
Date Sat, 10 Nov 2012 13:37:16 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-9022?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13494685#comment-13494685

Allen Wittenauer commented on HADOOP-9022:

When -m 0 is specified, it should probably just call copy in a serial fashion across the list
and not run a MapReduce job.  After all, setting the number of maps to zero implies that one
doesn't want a job executed at all.

> Hadoop distcp tool fails to copy file if -m 0 specified
> -------------------------------------------------------
>                 Key: HADOOP-9022
>                 URL: https://issues.apache.org/jira/browse/HADOOP-9022
>             Project: Hadoop Common
>          Issue Type: Bug
>    Affects Versions: 0.23.1, 0.23.3, 0.23.4
>            Reporter: Haiyang Jiang
> When trying to copy file using distcp on H23, if -m 0 is specified, distcp will just
spawn 0 mapper tasks and the file will not be copied.
> But this used to work before H23, even when -m 0 specified, distcp will always copy the
> Checked the code of DistCp.java
> Before the rewrite, it set the number maps at least to 1
> job.setNumMapTasks(Math.max(numMaps, 1));
> But in the newest code, it just takes the input from user:
> job.getConfiguration().set(JobContext.NUM_MAPS,
>                   String.valueOf(inputOptions.getMaxMaps()));

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

View raw message