[ 
https://issues.apache.org/jira/browse/MAPREDUCE-1073?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Dick King updated MAPREDUCE-1073:
---------------------------------

    Attachment: MAPREDUCE-1073--yhadoop20--2010-07-22--1530.patch

I revised the patch to not add an API to read and set set the property that 
tells MapTask.TrackedRecordReader to not record
progress as it reads the input; just read and set the property "by hand" in the 
code.  Since this is a
pipes-specific feature, it should be handled only by a focused attribute, which 
I then renamed to
mapred.pipes.disable.record.reader.progress .

In 
https://issues.apache.org/jira/browse/MAPREDUCE-1073?focusedCommentId=12891327&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#action_12891327
 , the API for marking a job as having mappers who will indicate their own 
progress is now to just set the property, which I have renamed from 
{{mapred.job.disable.record.reader.progress}} to 
{{mapred.pipes.disable.record.reader.progress}} , because this is a pipes-only 
concept.

> Progress reported for pipes tasks is incorrect.
> -----------------------------------------------
>
>                 Key: MAPREDUCE-1073
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-1073
>             Project: Hadoop Map/Reduce
>          Issue Type: Bug
>          Components: pipes
>    Affects Versions: 0.20.1
>            Reporter: Sreekanth Ramakrishnan
>            Assignee: Dick King
>         Attachments: mapreduce-1073--2010-03-31.patch, 
> mapreduce-1073--2010-04-06.patch, 
> MAPREDUCE-1073--yhadoop20--2010-07-22--1530.patch, 
> MAPREDUCE-1073--yhadoop20--2010-07-22.patch, MAPREDUCE-1073_yhadoop20.patch
>
>
> Currently in pipes, 
> {{org.apache.hadoop.mapred.pipes.PipesMapRunner.run(RecordReader<K1, V1>, 
> OutputCollector<K2, V2>, Reporter)}} we do the following:
> {code}
>         while (input.next(key, value)) {
>           downlink.mapItem(key, value);
>           if(skipping) {
>             downlink.flush();
>           }
>         }
> {code}
> This would result in consumption of all the records for current task and 
> taking task progress to 100% whereas the actual pipes application would be 
> trailing behind. 

-- 
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