Reply To: Setting the group of a MR job

Hi Alex,

In HDFS, when a file or sub-directory is created within a directory, the rule is that owner is set to the calling user and the group is inherited from the parent directory. See the last sentence of the second paragraph here.

http://hadoop.apache.org/docs/r2.7.0/hadoop-project-dist/hadoop-hdfs/HdfsPermissionsGuide.html#Overview

Based on that, would it be helpful if you do a one-time creation of a directory, then do a one-time chown to set it to whatever group you want for MapReduce job output, and then specify the MapReduce job’s output location as a sub-directory of this directory?