Fatal: Parallel job reports failure (code 139)

smuppidi · Post by **smuppidi** » Wed May 17, 2006 2:08 pm

Hello All,

I am trying to extract data from Teradata Enterprise Satge and i am getting the following error:

info: Contents of phantom output file =>
RT_SC356/OshExecuter.sh [20]: 1404 Memory fault

Fatal: Parallel job reports failure (code 139)

Simple SQLs are running ok but SQLs with a subselect are aborting with that message.There is a post out there but no clear direction.
Any insight on this would be helpful.

Thanks,
Satish.

roy · Post by **roy** » Sun May 21, 2006 1:13 am

Hi,
Have you contacted your support provider?

I don't have a TD to test saginst

Did you verify your configuration file has the correct memory definitions?
Ask your sysadmins to compare memory segments that are defined in the uvconfig file against the ones used in the machine; there is a chance something is wrong in that direction.

Please post the solution.

IHTH,

kumar_s · Post by **kumar_s** » Sun May 21, 2006 11:30 pm

Parallel job reports failure (code 139)... Hum... Got the same once....
What is the length of the query you execute?
Once i got the same with DB2 database. I tried to execute the qure in command line it executes well. When I try to do the same thought Execute command activity it gave out the same error. Later with some trail and error we found, it is due the the length of the query used, We just roughly eliminated the extraspace. It worked out. didnt noticed the exact length.
But I wont say this can be the only reason. Because, due this this, SIGKILL was raised and hend the job was aborted in my case.

smuppidi · Post by **smuppidi** » Mon May 22, 2006 4:42 pm

Thanks all for your suggestions. I got that fixed. It was due to fault with internal datasets that datastage creates while extracting the source data from Teradata. Admin people took care of that. But this code may also come in some cases due to extra spaces in query.

Satish.

kumar_s · Post by **kumar_s** » Mon May 22, 2006 10:56 pm

"It was due to fault with internal datasets that datastage creates "
Do you mean virtual dataset?
May I konw what where the steps taken to overcome this?

smuppidi · Post by **smuppidi** » Thu Jun 01, 2006 6:56 am

Kumar, it seems the Datasets were dropped and recreated.

kumar_s · Post by **kumar_s** » Thu Jun 01, 2006 10:49 am

smuppidi wrote:Kumar, it seems the Datasets were dropped and recreated.

Thanks for the response.

ray.wurlod · Post by **ray.wurlod** » Thu Jun 01, 2006 3:54 pm

Virtual Data Sets must be re-created. They only exist while the job is running. Are we talking persistent Data Sets here? Was the problem therefore one of design, where you were relying upon existence of a Data Set that was being overwritten? There's nothing you can do in a DataStage job that will actually drop a Data Set.