Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hadoop >> mail # user >> hive task fails when left semi join


Copy link to this message
-
Re: 答复: 答复: hive task fails when left semi join
Kira,

What version of hadoop are you using? The error exit with status code 126
is very rare condition. you may refer  to
(MAPREDUCE-4857<https://issues.apache.org/jira/browse/MAPREDUCE-4857>
,  MAPREDUCE-2374 <https://issues.apache.org/jira/browse/MAPREDUCE-2374>)

There are multiple possibilities for which this error comes but in most of
them if an attempt has failed hadoop tries to schedule the attempt on the
next node.
There is very little info available to make any sense out of this for me.
May be experts will be able to tell more on the detailed error

Didn't the failed attempt get launched again for few more times?
Sorry couldn't be of much help with this as I do not have enough log for my
understanding levels.
On Tue, Jul 16, 2013 at 2:35 PM, <[EMAIL PROTECTED]> wrote:

> Nitin,****
>
> ** **
>
> I check the log of failed task in corresponding machine, the stderr are
> like this,****
>
> ** **
>
> 2013-07-16 16:19:00,057 INFO org.apache.hadoop.mapred.TaskTracker:
> LaunchTaskAction (registerTask): attempt_201307041810_0142_m_000015_0
> task's state:UNASSIGNED****
>
> 2013-07-16 16:19:00,058 INFO org.apache.hadoop.mapred.TaskTracker: Trying
> to launch : attempt_201307041810_0142_m_000015_0 which needs 1 slots****
>
> 2013-07-16 16:19:00,058 INFO org.apache.hadoop.mapred.TaskTracker: In
> TaskLauncher, current free slots : 2 and trying to launch
> attempt_201307041810_0142_m_000015_0 which needs 1 slots****
>
> 2013-07-16 16:19:01,082 INFO org.apache.hadoop.mapred.TaskController:
> Writing commands to
> /hadoop/tmp/mapred/local/ttprivate/taskTracker/root/jobcache/job_201307041810_0142/attempt_201307041810_0142_m_000015_0/taskjvm.sh
> ****
>
> 2013-07-16 16:19:02,011 WARN org.apache.hadoop.mapred.TaskRunner:
> attempt_201307041810_0142_m_000015_0 : Child Error****
>
> 2013-07-16 16:19:06,061 INFO org.apache.hadoop.mapred.TaskTracker:
> LaunchTaskAction (registerTask): attempt_201307041810_0142_m_000015_0
> task's state:FAILED_UNCLEAN****
>
> 2013-07-16 16:19:06,061 INFO org.apache.hadoop.mapred.TaskTracker: Trying
> to launch : attempt_201307041810_0142_m_000015_0 which needs 1 slots****
>
> 2013-07-16 16:19:06,061 INFO org.apache.hadoop.mapred.TaskTracker: In
> TaskLauncher, current free slots : 1 and trying to launch
> attempt_201307041810_0142_m_000015_0 which needs 1 slots****
>
> 2013-07-16 16:19:06,124 INFO org.apache.hadoop.mapred.TaskController:
> Writing commands to
> /hadoop/tmp/mapred/local/ttprivate/taskTracker/root/jobcache/job_201307041810_0142/attempt_201307041810_0142_m_000015_0.cleanup/taskjvm.sh
> ****
>
> 2013-07-16 16:19:09,845 INFO org.apache.hadoop.mapred.TaskTracker: JVM
> with ID: jvm_201307041810_0142_m_-1660811086 given task:
> attempt_201307041810_0142_m_000015_0****
>
> 2013-07-16 16:19:13,456 INFO org.apache.hadoop.mapred.TaskTracker:
> attempt_201307041810_0142_m_000015_0 0.0%****
>
> 2013-07-16 16:19:16,052 INFO org.apache.hadoop.mapred.TaskTracker:
> attempt_201307041810_0142_m_000015_0 0.0% cleanup****
>
> 2013-07-16 16:19:16,053 INFO org.apache.hadoop.mapred.TaskTracker: Task
> attempt_201307041810_0142_m_000015_0 is done.****
>
> 2013-07-16 16:19:16,053 INFO org.apache.hadoop.mapred.TaskTracker:
> reported output size for attempt_201307041810_0142_m_000015_0  was -1****
>
> ** **
>
> From the Web UI:****
>
> ****
>
> ** **
>
> **1.       **Am I make the stderr clear?****
>
> **2.       **If so, how do your regard the error?****
>
> ** **
>
> ** **
>
> *发件人:* Nitin Pawar [mailto:[EMAIL PROTECTED]]
> *发送时间:* 2013年7月16日 16:44
> *收件人:* [EMAIL PROTECTED]
> *主题:* Re: 答复: hive task fails when left semi join****
>
> ** **
>
> Kira,****
>
> ** **
>
> I think the job got completed successfully. If a task has failed on one
> tasktracker hadoop takes care of rescheduling it to another for # number of
> retries.****
>
> I see the job status as 243/243 completed. ****
>
> ** **
>
> can you confirm once if your job has failed and if it has failed can you
> please share the stderr log for that particular task only ****

Nitin Pawar