[ 
https://issues.apache.org/jira/browse/MAPREDUCE-1641?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12859224#action_12859224
 ] 

Arun C Murthy commented on MAPREDUCE-1641:
------------------------------------------

Dick, some comments:

# I'd rename DistributedCache.validateCachesDontOverlap to 
DistributedCache.validate to keep it simple, in future we may add more 
validations there, so having one public api might prove useful and succint.
# In the same function, and elsewhere, I'd define a public static final String 
for 'mapred.cache.archives' and 'mapred.cache.files' and use them instead of 
the raw strings.
# Request you to add a test-case which uses the Mini* clusters to ensure the 
validation works with fully-qualified hdfs uris (hdfs://nn:8000/user/blah/foo), 
and absolute Paths (/user/blah/foo).

> Job submission should fail if same uri is added for mapred.cache.files and 
> mapred.cache.archives
> ------------------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-1641
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1641
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: distributed-cache
>            Reporter: Amareshwari Sriramadasu
>            Assignee: Dick King
>             Fix For: 0.22.0
>
>         Attachments: BZ-3539321--off-0-20-101--2010-04-20.patch
>
>
> The behavior of mapred.cache.files and mapred.cache.archives is different 
> during localization in the following way:
> If a jar file is added to mapred.cache.files,  it will be localized under 
> TaskTracker under a unique path. 
> If a jar file is added to mapred.cache.archives, it will be localized under a 
> unique path in a directory named the jar file name, and will be unarchived 
> under the same directory.
> If same jar file is passed for both the configurations, the behavior 
> undefined. Thus the job submission should fail.
> Currently, since distributed cache processes files before archives, the jar 
> file will be just localized and not unarchived.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to