Phantom error - DSP.Close

[gomez]

Hi:
In a random fashion, we get Phantom error like below in about 3 - 4 DataStage jobs.
The following is the message I get when I reset the jobs after abort. The jobs run fine on other days, some days they fail at the DSP.Close step.
I searched the forum for answers to this. There are similar posts but not exactly this. The random behaviour of the jobs seem strange. Without even changing a single thing, they go fine during the next run.
Could this be caused by insufficient memory at the time of execution?

Code: Select all

DataStage Job 15 Phantom 888
Program "DSP.Close": Line 94, Exception raised in GCI subroutine:
Access violation.
Attempting to Cleanup after ABORT raised in stage SVR_ICD0001_PPS_Board_Auth_Budget_to_SAP..TRNS_Board_Auth_PLAN_to_iDoc
Program "DSD.INTClose": Line 141, Unable to open the operating system file "DSD_BP.O/%".
[ENOENT] No such file or directory
Program "DSD.INTClose": Line 141, Unable to load file "".
Program "DSD.INTClose": Line 141, Unable to load subroutine.

Server details :
Windows 2003 Server
Intel(R) Xeon(R) CPU
5150@2.66GHz
2.67GHz, 3.25 GB of RAM

Here are the details about a couple of job designs
Job 1 reads from ODBC, has an aggregator, 2 transformers and writes to two IDoc stages
Job 2 reads from file, has a transformer, 2 ODBC look ups and writes to an Idoc stage

Appreciate any suggestions/recommendations on this.

Thanks!

[gomez]

When I did a search on the server for DSP.Close, there are two files :
One in

Code: Select all

\\tsclient\U\Ascential_7.5.1_backup\DataStage\Engine\catdir

and another in

Code: Select all

\\tsclient\C\Ascential\DataStage\Engine\catdir

Could this be reason for any conflict? The backup folder should have come as part of the install of 8x.

[ag_ram]

Could you enable Server Side tracing and look at the tract file to find which internal BASIC Subroutine calls these programs while resetting?

[ray.wurlod]

DSP.Close can raise an alert if SAP does not respond in a timely fashion.

[gomez]

That seems like the cause of the error. Is there any way to fix this?

Thanks!

[gomez]

I also had to mention, the abort occurs after all the Idocs are written to SAP. Do you say SAP has to send some status back after it received all the Idocs?