abnormal termination of transformer stage

mctny · Post by **mctny** » Mon Mar 17, 2008 9:07 am

first of All I searched the forum but it didnot help me resolve my problem.

I am getting the following error in our QA environment, the job is running fine in our dev environment.

The error is

Abnormal termination of stage O2c_Int_Avenue_20_Transform_or_Rejects_Data_Job..Format_Actuals_Like_Forecast_Stage_1 detected.

when I reset and rerun, the logs show this. abort for the same reason as above.

Job name:O2c_Int_Avenue_20_Transform_or_Rejects_Data_Job
.
.
Message:
From previous run
DataStage Job 90 Phantom 20409
jobnotify: Unknown error
DataStage Phantom Finished.
[20414] Done : DSD.StageRun O2c_Int_Avenue_20_Transform_or_Rejects_Data_Job. O2c_Int_Avenue_20_Transform_or_Rejects_Data_Job.Actual_Forecast_Combined.IDENT1 1 0/50
[20415] DSD.StageRun O2c_Int_Avenue_20_Transform_or_Rejects_Data_Job. O2c_Int_Avenue_20_Transform_or_Rejects_Data_Job.Format_Actuals_Like_Forecast_Stage_1 1 0/50 - terminated.

any thoughts?

chulett · Post by **chulett** » Mon Mar 17, 2008 9:18 am

No, not really. How about explaining what your job does, how it is designed? Let's start there. IDENT1 would be a transformer that it added between two passive stages, btw.

mctny · Post by **mctny** » Mon Mar 17, 2008 9:31 am

the name of the transformer is Format_Actuals_Like_Forecast_Stage_1

the job design is as follows

Source( Ora9i) has two output link both of which has user defined SQL, one go to a hash file( Actual_Forecast_Combined which runs fine), the other goes to the transformer stage above, and there are two output of this transformer one is going to another hash file the other is going to an aggregator (SUM) stage.

The problem is in one of the output of the source oracle stage which is going into the transformer stage.

chulett · Post by **chulett** » Mon Mar 17, 2008 9:40 am

Then you need to track down what's different between dev and test. Are they two different DataStage environments or are both projects on the same server? What about the data itself? Same source? Is there a significant delta in the volumes?

If not, first check could be to copy the job, pop off the Agg and whatever comes after it and replace it with a Sequential stage, see what your data actually looks like and if there is anything... unusual or different... about it in test.

mctny · Post by **mctny** » Mon Mar 17, 2008 9:55 am

Thanks for the answer again.

The environment are two different ones, the source data is same though, the source queries are exactly same ( they are both taking data from production database as we don't have data in our dev or test). The amount of data ( the User defined SQLs can be considered significant.) one is 170000 rows the other is around 7000 both has around 20 columns)

chulett · Post by **chulett** » Mon Mar 17, 2008 10:01 am

I was wondering about a difference in volume in the link that goes to the Aggregator, dev versus test. 'Large' volumes can make it fall over dead. However, it sounds like everything would be identical there.

So, are you saying that your dev and test DataStage servers are separate server? If so, what differences are there between them? Database client software version is the first thing I'd check in your shoes.

mctny · Post by **mctny** » Mon Mar 17, 2008 10:17 am

chulett wrote:....
So, are you saying that your dev and test DataStage servers are separate server?...

Yes.

I am not very sure about the difference but I will try to check,

oracle client should be same in both env.

chulett · Post by **chulett** » Mon Mar 17, 2008 11:06 am

I agree - it *should* be.

mctny · Post by **mctny** » Tue Mar 18, 2008 11:34 am

chulett wrote:I agree - it *should* be.

Both of the clients are same Oracle 9.2.0.7.0, the server version is Oracle9i Enterprise Edition Release 9.2.0.5.0 for both environments.

I haven't seen any differences yet.