hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Robert Joseph Evans (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-4194) ConcurrentModificationError in DirectoryCollection
Date Wed, 25 Apr 2012 15:04:17 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-4194?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13261692#comment-13261692
] 

Robert Joseph Evans commented on MAPREDUCE-4194:
------------------------------------------------

The test will delete the ./target dir, which is not really something that I want removed after
running tests because it may have all of my compiled .class files in it, along with who knows
what else.  I would rather see something like target/<testName>/ as the testDir.

Also getGoodDirs() and getFailedDirs() are still returning internal data structures.  This
is not really what this bug was trying to address because this is just to avoid a concurrent
modification exception, but I is still bad practice, because I can delete an entry from the
list returned and that will be reflected for all people with the list.  I would like to see
the Lists returned to be immutable, or if they are changed they will have no impact on DirectoryCollection.
 That can be done on a separate JIRA 
                
> ConcurrentModificationError in DirectoryCollection
> --------------------------------------------------
>
>                 Key: MAPREDUCE-4194
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-4194
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: mrv2
>    Affects Versions: 0.23.3, 2.0.0
>            Reporter: Jonathan Eagles
>            Assignee: Jonathan Eagles
>         Attachments: MAPREDUCE-4194.patch
>
>
> As found as part of work on MAPREDUCE-4169, it is possible for a ConcurrentModificationException
to be thrown upon disk failure. DirectoryCollection hands out its internal list structure
that is accessed across multiple threads. Upon disk failure its internal list is modified,
invalidating all current iterators to that structure.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message