error while running the job

deva · Post by **deva** » Tue Dec 22, 2009 1:27 pm

Hi ,
When I run the job I am getting the following error

"Operator terminated abnormally: received signal SIGSEGV"

my database connections are good. it is running 3 minutes and after that my job is aboring with the above error.

can any one suggest me please. I am hoping its memory error, so I changed scrach desk path as well. but it won't work.

chulett · Post by **chulett** » Tue Dec 22, 2009 1:38 pm

SIGSEGV is a Segmentation Violation signal, more of a coding than a resource issue. Did you try search here for that message? What is your job design?

deva · Post by **deva** » Tue Dec 22, 2009 1:48 pm

No mostly the design is look up for the data and giving input to shared container and inserting or updating to the database.

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Wed Dec 23, 2009 3:07 am

You need to be more specific.

As Craig suggested, do a search on SIGSEGV. It is definitely an issue in the code.

deva · Post by **deva** » Wed Dec 23, 2009 8:04 am

Hi , looks like no issue in the code. here are the things

- I am running my jobs on grid environment.
- getting data from oracle database.
- lookup with db2 database based on the key.
- using transwarmenr loading data into oracle table.

This is the code.

if I run the job it is running 3 min. and aborting . Is there any issue with look up? but lookup throughing data. But from source I am getting 1 record.

chulett · Post by **chulett** » Wed Dec 23, 2009 8:08 am

'Code' doesn't necessarily mean your 'code', could be a driver or parallel framework bug, etc. Involve your official support provider.

deva · Post by **deva** » Wed Dec 23, 2009 8:33 am

Can any one suggest what is the solution for this issue?

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Wed Dec 23, 2009 8:37 am

What is the volume ?

What happens if you run without any rows ?

chulett · Post by **chulett** » Wed Dec 23, 2009 8:53 am

These are not typically simple to figure out. Involve your official support provider.

deva · Post by **deva** » Wed Dec 23, 2009 9:41 am

The volume of data in the lookup is 3 mil.

we don't have any other support team. If any one know the right answer or suggestions please respond.

chulett · Post by **chulett** » Wed Dec 23, 2009 9:48 am

There's no magic answer. I would suggest that you search for SIGSEGV here and concentrate first on any that are marked Resolved, see if anything in them helps. Then go through the allegedly unresolved ones.

Then experiment. Does the grid node the job runs on matter? Does it work if you constrain it to one node? Can you 'work around' the issue by breaking the job up into two logical halves? Are you hash partitioning the data going into the Lookup based on the lookup keys to minimize the amount of data each node processes? Etc, etc, etc. Don't just sit around and wait for The Answer to fall from the sky.