Reducers continue to run even if a job failed.
----------------------------------------------
Key: HADOOP-5609
URL: https://issues.apache.org/jira/browse/HADOOP-5609
Project: Hadoop Core
Issue Type: Bug
Components: mapred
Affects Versions: 0.18.3
Reporter: Qi Liu
When a job failed, in rare case, some reducer slots are not freed (continue to
try collecting data even if the job is not running). These occupied reducer
slots are not usable by following mapred jobs unless the task trackers are
restarted.
These reducers continue to report the progress to job tracker, which results
the following logs:
2009-03-31 23:59:59,992 INFO org.apache.hadoop.mapred.JobTracker: Serious
problem. While updating status, cannot find taskid
attempt_200903241312_0074_r_000001_0
2009-03-31 23:59:59,992 INFO org.apache.hadoop.mapred.JobTracker: Serious
problem. While updating status, cannot find taskid
attempt_200903241312_0074_r_000958_0
2009-03-31 23:59:59,997 INFO org.apache.hadoop.mapred.JobTracker: Serious
problem. While updating status, cannot find taskid
attempt_200903241312_0074_r_000440_0
2009-03-31 23:59:59,998 INFO org.apache.hadoop.mapred.JobTracker: Serious
problem. While updating status, cannot find taskid
attempt_200903241312_0074_r_000463_0
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
