hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Amogh Vasekar <am...@yahoo-inc.com>
Subject Re: getJobID and job handling
Date Mon, 26 Jul 2010 08:40:03 GMT
I see you are using the new APIs, so this should be relevant for you https://issues.apache.org/jira/browse/MAPREDUCE-118

As you have noticed, in the old APIs the JobClient could be queried using JobID , which was
returned when the job was submitted. There was a thread in hadoop-dev to discuss un-deprecating
the old APIs (not sure where it ultimately went, you may search around the list though)
In any case, the patch on above jira should fix it.


On 7/23/10 9:39 PM, "Michael Sutter" <Michael.Sutter@kit.edu> wrote:

  Hello everybody,

I have a problem with my application and hopefully someone can help me.
I want to submit a large number of jobs, notice the job id and come back
some hours to query the status of the jobs.

The idea was to create the Job (org.apache.hadoop.mapreduce.Job), query
the job id via getJobID() and store it in a database. So far everything
works fine,
except that getJobID() always returns null. I found a workaround with
and cut the JobID from it.

So to my questions:
Is it a bug, that getJobID() always returns null?

And is it possible to query the status when I only have the job id? The
only information I found is to
do it with JobClient.getJob(JobID), whereas JobID is already deprecated.
Or is this the correct
way to do it?

I'm using hadoop 0.20.2.

Thanks in advance

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message