[MAPREDUCE-1628] HarFileSystem shows incorrect replication numbers and permissions

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.21.0
• Component/s: harchive
• Hadoop Flags: Reviewed
• Assignee: Tsz-wo Sze
• Reporter: Tsz-wo Sze
• Created: 27/Feb/10 00:02
• Updated: 24/Aug/10 21:21
• Resolved: 31/Mar/10 18:46

Description

In the har dir, the replication # of part-0 is 3.

-bash-3.1$ hadoop fs -ls  ${DIR}.har
Found 3 items
-rw-------   5 tsz users       1141 2010-02-10 18:34 /user/tsz/t20.har/_index
-rw-------   5 tsz users         24 2010-02-10 18:34 /user/tsz/t20.har/_masterindex
-rw-------   3 tsz users      15052 2010-02-10 18:34 /user/tsz/t20.har/part-0

but the replication # of the individual har:// files is shown as 5.

-bash-3.1$ hadoop fs -lsr  ${HAR_FULL}/
drw-------   - tsz users          0 2010-02-10 18:34 /user/tsz/t20.har/t20
-rw-------   5 tsz users        723 2010-02-10 18:34 /user/tsz/t20.har/t20/text-00000000
-rw-------   5 tsz users        779 2010-02-10 18:34 /user/tsz/t20.har/t20/text-00000001
-rw-------   5 tsz users        818 2010-02-10 18:34 /user/tsz/t20.har/t20/text-00000002
...

The permission also has similar problem. Clearly, the permission of t20.har/t20 shown above is incorrect.

Issue Links

• is related to: MAPREDUCE-1548 Hadoop archives should be able to preserve times and other properties from original files