One line error

ramana · Post by **ramana** » Thu Apr 27, 2006 5:45 pm

Hi,

All these days, my search has returned answers for my questions in dsxchange. This time I found many questions of this type, but not proper solution.

My server job is Production and been running for quite some time. Today it started giving me this one liner "Attempting to Cleanup after ABORT raised in stage PLA_Src_Lod" error. I tried resetting, recompiling. Nothing is working out.

Your help will be highly appreciated as it has been.

jreddy · Post by **jreddy** » Thu Apr 27, 2006 6:18 pm

ramana,Can you post the error messages starting from the first fatal error msg in the log for this job.. it would better help us understand the source of this abort.

kcbland · Post by **kcbland** » Thu Apr 27, 2006 7:23 pm

Hey, 2 years and this is your first post. Sounds like the forum Search facility works well!!

Can you describe the job design, what kinds of stages, how they're used? If you're getting an abnormal termination, it could be the result of a math operation on a NULL value. We'll need to try to help pinpoint the issue.

ramana · Post by **ramana** » Thu Apr 27, 2006 8:09 pm

kcbland wrote:Hey, 2 years and this is your first post. Sounds like the forum Search facility works well!!

Can you describe the job design, what kinds of stages, how they're used? If you're getting an abnormal termination, it could be the result of a math operation on a NULL value. We'll need to try to help pinpoint the issue.

Thrilled to see quick responses.

Here is my Job design.

Sequencial file -----> xfm----------> Teradata Mload Stage

This job is used to load 1 mil rows into a teradata table.

I am not seeing any problem from Teradata side as it is not locking the target table.

This Job has been running fine for the past six months. All the nulls are handled in the xfm properly.

kcbland · Post by **kcbland** » Thu Apr 27, 2006 8:38 pm

Without any error mesasges, even when resetting, you're left with having to hunt down the bug.

Check to see if any before/after job/stage routines are being called. If so, make sure they are still compiled. Opening a Routine and changing it without compiling it removes the compiled code. Jobs mysteriously die because the CALL for the Routine can't be handled.

Can you copy the job and replace the Teradata stage with a Sequential stage? If you run that job and it works, then it means that the data is operable by the functions and computations within the transformer. That points the finger to the Teradata stage.

If the job copy fails, then you're going to have to figure out which computation is causing the job to abort. The easiest way to figure that one out is to look at any columns that aren't straight mappings. If they're being mathematically computed, that's where NULL values get into the mix. If NULLs aren't your problem, it would be probably something in a custom DS function that maybe wrong.

Again, a DS function just like a routine is subject to losing its compiled code if saved but not compiled. Mysterious things happen because the job already understands there's supposed to be a function out there, but not finding it causes an abnormal termination.

Good luck.

ramana · Post by **ramana** » Sat Apr 29, 2006 7:32 am

kcbland wrote:Without any error mesasges, even when resetting, you're left with having to hunt down the bug.

Check to see if any before/after job/stage routines are being called. If so, make sure they are still compiled. Opening a Routine and changing it without compiling it removes the compiled code. Jobs mysteriously die because the CALL for the Routine can't be handled.

Can you copy the job and replace the Teradata stage with a Sequential stage? If you run that job and it works, then it means that the data is operable by the functions and computations within the transformer. That points the finger to the Teradata stage.

If the job copy fails, then you're going to have to figure out which computation is causing the job to abort. The easiest way to figure that one out is to look at any columns that aren't straight mappings. If they're being mathematically computed, that's where NULL values get into the mix. If NULLs aren't your problem, it would be probably something in a custom DS function that maybe wrong.

Again, a DS function just like a routine is subject to losing its compiled code if saved but not compiled. Mysterious things happen because the job already understands there's supposed to be a function out there, but not finding it causes an abnormal termination.

Good luck.

Thank you guys, We found the root cause. It was the space problem. We cleared some space and every thing started working fine.

If DataStage log could give more specific error description, it would be much more helpful.

Sreenivasulu · Post by **Sreenivasulu** » Sun Apr 30, 2006 5:34 am

Hi

Space problem in mounts other than the mount where datastage is installed causes problems in the functioning of the tool.

The operating system becomes ustable if any mount has less free space.

Regards
Sreeni