[
https://issues.apache.org/jira/browse/MAPREDUCE-1480?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12841641#action_12841641
]
Aaron Kimball commented on MAPREDUCE-1480:
------------------------------------------
Amareshwari,
Thanks for looking over this patch.
The previous progress calculator was strictly based on the number of sub-splits
processed. The underlying RecordReader's getProgress() function was never
called, which means that the granularity of progress was only based around the
number of subsplits and did not take intra-split progress into account. A
review from Dhruba is definitely welcome.
I'll add another testcase as you suggest and post this in the next couple of
days.
> CombineFileRecordReader does not properly initialize child RecordReader
> -----------------------------------------------------------------------
>
> Key: MAPREDUCE-1480
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1480
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Reporter: Aaron Kimball
> Assignee: Aaron Kimball
> Attachments: MAPREDUCE-1480.2.patch, MAPREDUCE-1480.patch
>
>
> CombineFileRecordReader instantiates child RecordReader instances but never
> calls their initialize() method to give them the proper TaskAttemptContext.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
