Fatal Error with succesfull Completion

[Camaj]

Bonjour,

I run the DataStage PX with DB2 in the UNIX system.

I create a job that insert to a DB2 table. When I ran my job twice I got the followed error from DB2 --> DB2 Driver Embedded SQL: -803;
sqlcode = 23505;
sqlstate = SQL0803N One or more values in the INSERT statement, UPDATE statement, or
foreign key update caused by a DELETE statement are not valid because the
primary key, unique constraint or unique index identified by "1" constrains.

When I check in the DIRECTOR log I saw 2 FATAL error. This is normal I tried to insert row already in the table.

My problem is that my job completed with "succesfull".

Wich option I have to set to said to DataStage to stop the job when DB2 got a FATAL error.

Thank !

[kumar_s]

Hi,
if you are calling from command prompt, you can get the return code and can use DsStopJob() to stop the flow.

regard
kumar

[Camaj]

Thank !

But When I call from command prompt I recieive a code 2, this mean a warning !

How I can know, if my job got this error if the return code is a warning ?

[kumar_s]

Not understanding you requirment fully .......

This is again a warning from db2.
if you try to load some duplicate data, the first set of data will be loaded and rest will be rejected.
And it will try to load further with the rest of records.
so instead of rejecting the flow you can try to collect into a reject file.

regards
kumar

[Camaj]

Yes I understand I can collect error to the reject file but in my stream if this condition will happen, I want to stop the processing with a ABORT.

Thank fro your quick reply !!!

[HSBCdev]

I'm also getting this error. What causes this problem? Is it a bug (will it be fixed)? What's the best workaround for it?

Thanks

[ray.wurlod]

One workaround would be to post-process the job log. Use DSGetNewestLogID to find the most recent Fatal event; this will return 0 if there is none.

[HSBCdev]

Thanks - I'll try that

[ray.wurlod]

If you're running version 7.5 or later, and if you're controlling the job from a job sequence, then you can include a Terminator activity to issue stop requests to all jobs running under control of that sequence.

[HSBCdev]

I can recreate this error in 7.1 but not in 7.5 (i.e. the job correctly aborts when one partition fails because of a deadlock). Is it possible to get it confirmed that this bug has been fixed?

[kumar_s]

Nothing mention in release notes about these, upto to my knowledge. Best place to get it confiremed, contact your service provider and rasie a case with IBM tech people.