[jira] [Commented] (MAPREDUCE-3097) archive does not archive if the content specified is a file

["Vasco (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/MAPREDUCE-3097?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13692005#comment-13692005 ] 

Vasco commented on MAPREDUCE-3097:
----------------------------------

The contents is actually archived. What goes wrong here is that listStatus() in HarFileSystem does not list the archive's top level content unless the index contains a root directory (i.e., "/").

There are two possibilities to fix this unwanted behaviour:
1) make sure the hadoop archive tool always creates a root directory entry in the index.
2) make listStatus() work even when there is no root directory entry in the index.

There is something to be said in favor of both options. Maybe the HAR file and index format is formally specified somewhere (if so, can someone point me to this?). Does the specification imply either of these two options? Does anyone have an opinion on either? 
                
> archive does not archive if the content specified is a file
> -----------------------------------------------------------
>
>                 Key: MAPREDUCE-3097
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-3097
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>    Affects Versions: 0.20.203.0, 0.20.205.0
>            Reporter: Arpit Gupta
>            Priority: Minor
>
> archive command only archives directories. when the content specified is a file it proceeds with the archive job but does not archive the content this can be misleading as the user might think that archive was successful. We should change it to either throw an error or make it archive files.

--
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