hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joydeep Sen Sarma (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-153) skip records that throw exceptions
Date Tue, 29 Apr 2008 17:25:56 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-153?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12593053#action_12593053
] 

Joydeep Sen Sarma commented on HADOOP-153:
------------------------------------------

coversely, if the recordreader can avoid throwing exceptions - then u don't really need new
api's from the recordreader. u can just void calling map()/reduce() on specific records that
had a problem.

one small implementation request - one of the things i have noticed (based on some profiling)
is that a lot of processing overhead (in the map side)  in hadoop comes from function calls
that are invoked for every processed record. it would be nice if this jira does not lead to
more of the same. considering that this is not a normal case - it might make sense to have
different code paths for the first and subsequent retries (so that in the normal case - we
don't incur overhead looking for records to be skipped). 

> skip records that throw exceptions
> ----------------------------------
>
>                 Key: HADOOP-153
>                 URL: https://issues.apache.org/jira/browse/HADOOP-153
>             Project: Hadoop Core
>          Issue Type: New Feature
>          Components: mapred
>    Affects Versions: 0.2.0
>            Reporter: Doug Cutting
>            Assignee: Devaraj Das
>
> MapReduce should skip records that throw exceptions.
> If the exception is thrown under RecordReader.next() then RecordReader implementations
should automatically skip to the start of a subsequent record.
> Exceptions in map and reduce implementations can simply be logged, unless they happen
under RecordWriter.write().  Cancelling partial output could be hard.  So such output errors
will still result in task failure.
> This behaviour should be optional, but enabled by default.  A count of errors per task
and job should be maintained and displayed in the web ui.  Perhaps if some percentage of records
(>50%?) result in exceptions then the task should fail.  This would stop jobs early that
are misconfigured or have buggy code.
> Thoughts?

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


Mime
View raw message