[HADOOP-5233] Reducer not Succeded after 100%

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 0.19.0
• Fix Version/s: 0.19.2, 0.20.0
• Hadoop Flags: Reviewed
• Assignee: Amareshwari Sriramadasu
• Reporter: Amareshwari Sriramadasu
• Created: 12/Feb/09 10:43
• Updated: 08/Jul/09 16:53
• Resolved: 20/Feb/09 13:18

Description

I have seen a reducer in RUNNING state, when it was waiting for the commitResponse.
Task logs show :
2009-02-12 08:35:36,414 INFO org.apache.hadoop.mapred.TaskRunner: Task:attempt_200902120746_0297_r_000033_0 is done. And is in the process of commiting
and no logs after that.
TT logs say:
2009-02-12 08:35:36,417 INFO org.apache.hadoop.mapred.TaskTracker: Task attempt_200902120746_0297_r_000033_0 is in COMMIT_PENDING
2009-02-12 08:35:36,417 INFO org.apache.hadoop.mapred.TaskTracker: attempt_200902120746_0297_r_000033_0 0.33333334% reduce > sort

This looks like, the task progress went from COMMIT_PENDING to RUNNING again.

Issue Links

• incorporates: HADOOP-5234 NPE in TaskTracker reinit action
• incorporates: HADOOP-5235 possible NPE in tip.kill()