[
https://issues.apache.org/jira/browse/HADOOP-3150?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12602435#action_12602435
]
Owen O'Malley commented on HADOOP-3150:
---------------------------------------
Which leads back to the fact that I'm pretty uncomfortable with such a big
change to the APIs this close the freeze date. The current OutputFormat method
names are very file system specific. I'd prefer something like:
{code}
public class JobContext {
public JobConf getJobConf();
}
public class TaskAttemptContext {
public TaskAttemptID getTaskAttemptID();
public JobConf getJobConf();
public Progressable getProgressable();
}
public abstract class OutputFormat {
abstract public RecordWriter<K, V> getRecordWriter(TaskAttemptContext context
) throws IOException;
abstract public void checkOutputSpecs(JobContext context) throws IOException;
public abstract setupJob(JobContext context) throws IOException;
public abstract cleanupJob(JobContext context, boolean promote) throws
IOException;
public abstract setupTask(TaskAttemptContext context) throws IOException;
public abstract cleanupTask(TaskAttemptContext context, boolean promote)
throws IOException;
}
{code}
At that point, you might as well move OutputFormat over to
org.apache.hadoop.mapreduce and deprecate the old interface.
I don't think having the TaskTracker rpc handler (handling a call from the
task) block on a call to the JobTracker is workable.
> Move task file promotion into the task
> --------------------------------------
>
> Key: HADOOP-3150
> URL: https://issues.apache.org/jira/browse/HADOOP-3150
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Devaraj Das
> Fix For: 0.18.0
>
> Attachments: 3150.patch
>
>
> We need to move the task file promotion from the JobTracker to the Task and
> move it down into the output format.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
