Hi,

 

I do have the following problem monitoring my Hive queries in Hadoop.

 

I create a Sever using the Hive library which connects to an Hadoop cluster
(file system, job tracker and Hive metastore are set up on this cluster).
The needed parameters for the Hive server I've set in the configuration.

Commands sent to this server are executed, that works.

The problem now is that the MapReduce jobs created by Hive don't appear in
the job tracker's job list. The list itself works, showing me other commands
I executed not using Hive. (I use the JobClient Java class to get the list)

 

Why are the Hive query jobs not in the list? How can I track them in Hadoop?

 

 

Thanks

 

 

Von: Nitin Pawar [mailto:nitinpawar...@gmail.com] 
Gesendet: Donnerstag, 28. Februar 2013 16:13
An: user@hive.apache.org
Betreff: Re: Get the job id for a hive query

 

you can set this property mapred.job.name and this should set the name for
the job 

 

On Thu, Feb 28, 2013 at 8:26 PM, Tim Bittersohl <t...@innoplexia.com> wrote:

Thanks for the response,

 

I also found no way to access the job id via java thrift client, all I can
get is a query ID by the query planner.

 

How to set the name of a job where a Hive query is fired with, so I can find
it in the job tracker later?

 

Tim Bittersohl 

Software Engineer 


http://www.innoplexia.de/ci/logo/inno_logo_links%20200x80.png

Innoplexia GmbH
Mannheimer Str. 175 

69123 Heidelberg 

Tel.: +49 (0) 6221 7198033 <tel:%2B49%20%280%29%206221%207198033>  
Fax: +49 (0) 6221 7198034 <tel:%2B49%20%280%29%206221%207198034>  
Web: www.innoplexia.com <http://www.innoplexia.com/>  

Sitz: 69123 Heidelberg, Mannheimer Str. 175 - Steuernummer 32494/62606 -
USt. IdNr.: DE 272 871 728 - Geschäftsführer: Christian Schneider, Walery
Strauch, Prof. Dr. Herbert Schuster 

 

Von: Nitin Pawar [mailto:nitinpawar...@gmail.com] 
Gesendet: Donnerstag, 28. Februar 2013 15:46


An: user@hive.apache.org
Betreff: Re: Get the job id for a hive query

 

With thrift Client I don't think you can get the jobID from hadoop (I may
very well be wrong in this)

 

the other way around this was to have a separate name for each job you fire
through hive and then directly query jobtracker for the same 

 

On Thu, Feb 28, 2013 at 5:18 PM, Tim Bittersohl <t...@innoplexia.com> wrote:

I use java and there HiveClient of the hive library (Version 0.10.0).

 

Tim Bittersohl 

Software Engineer 


http://www.innoplexia.de/ci/logo/inno_logo_links%20200x80.png

Innoplexia GmbH
Mannheimer Str. 175 

69123 Heidelberg 

Tel.: +49 (0) 6221 7198033 <tel:%2B49%20%280%29%206221%207198033>  
Fax: +49 (0) 6221 7198034 <tel:%2B49%20%280%29%206221%207198034>  
Web: www.innoplexia.com <http://www.innoplexia.com/>  

Sitz: 69123 Heidelberg, Mannheimer Str. 175 - Steuernummer 32494/62606 -
USt. IdNr.: DE 272 871 728 - Geschäftsführer: Christian Schneider, Walery
Strauch, Prof. Dr. Herbert Schuster 

 

Von: Nitin Pawar [mailto:nitinpawar...@gmail.com] 
Gesendet: Donnerstag, 28. Februar 2013 12:43
An: user@hive.apache.org


Betreff: Re: Get the job id for a hive query

 

how are you running your hive queries? using hive cli or hive jdbc client? 

 

if you are using hive cli then whenever you fire a query and assuming it is
syntactically correct and its not select * from table operation 

hive cli shows  the job ID and an URL which points to the job tracker url
for the job it submitted  

 

On Thu, Feb 28, 2013 at 5:09 PM, Tim Bittersohl <t...@innoplexia.com> wrote:

I’m trying to get the job id of the job created with a Hive query.

At the moment I can get the cluster status from the HiveClient, but I don’
find any job id in there... 

 

Tim Bittersohl 

Software Engineer 


http://www.innoplexia.de/ci/logo/inno_logo_links%20200x80.png

Innoplexia GmbH
Mannheimer Str. 175 

69123 Heidelberg 

Tel.: +49 (0) 6221 7198033 <tel:%2B49%20%280%29%206221%207198033>  
Fax: +49 (0) 6221 7198034 <tel:%2B49%20%280%29%206221%207198034>  
Web: www.innoplexia.com <http://www.innoplexia.com/>  

Sitz: 69123 Heidelberg, Mannheimer Str. 175 - Steuernummer 32494/62606 -
USt. IdNr.: DE 272 871 728 - Geschäftsführer: Christian Schneider, Walery
Strauch, Prof. Dr. Herbert Schuster 

 

Von: Harsh J [mailto:ha...@cloudera.com] 
Gesendet: Donnerstag, 28. Februar 2013 09:08
An: hive request
Betreff: Re: Get the job id for a hive query

 

The client logs print the Job ID of the spawned job and a tracking URL. Is
that what you're looking for? Its printed for each stage.

 

On Wed, Feb 27, 2013 at 11:06 PM, Tim Bittersohl <t...@innoplexia.com> wrote:

Hi,

 

has the Hive client the possibility to give back the job id of the job
created when running a query? I need that for tracking.

 

 

Greetings

Tim Bittersohl 

Software Engineer 


http://www.innoplexia.de/ci/logo/inno_logo_links%20200x80.png

Innoplexia GmbH
Mannheimer Str. 175 

69123 Heidelberg 

Tel.: +49 (0) 6221 7198033 <tel:%2B49%20%280%29%206221%207198033>  
Fax: +49 (0) 6221 7198034 <tel:%2B49%20%280%29%206221%207198034>  
Web: www.innoplexia.com <http://www.innoplexia.com/>  

Sitz: 69123 Heidelberg, Mannheimer Str. 175 - Steuernummer 32494/62606
<tel:32494%2F62606>  - USt. IdNr.: DE 272 871 728 - Geschäftsführer:
Christian Schneider, Walery Strauch, Prof. Dr. Herbert Schuster 

 





 

-- 
Harsh J 





 

-- 
Nitin Pawar





 

-- 
Nitin Pawar





 

-- 
Nitin Pawar

<<image001.png>>

Reply via email to