Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
MapReduce >> mail # user >> Setting the group for output files

Copy link to this message
Re: Setting the group for output files

On 3/11/10 11:05 AM, "Gregory Lawrence" <[EMAIL PROTECTED]> wrote:
> Is there a way to set the output group for a mapreduce (or hdfs fs operation)
> job? For example -Ddfs.umaskmode=027 successfully sets the permissions. I
> would think the -Dgroup.name=GROUP would do a similar thing for the file's
> group. However, this does not appear to be the case. Any help would be
> appreciated.

HDFS uses sticky groups for directories.  In other words, on creation, the
group of the dir is what the group of the file will be.