DSXchange

I had DataStage 7.5 running on Windows XP Professional, the version 7.5x2 is specifically for PX on Windows.

Can you post your whole source line, perhaps you have mismatched commas or quotes elsewhere in it.

Do you have a 4-node configuration file?

Then I'm confused with the

uccio2002 wrote:...i catch a fatal error after an update on db2 via odbc...

part. How are you issuing this?

Have you monitored your scratch space while the job was running? Perhaps the broken pipe was caused by space issues.

Have you tried "stop job" from the director? If this hasn't helped, then the job is in a waiting state where it isn't able to respond to stop requests, which would be in the database in this case. Can you have the Oracle DBA check the locks and clear them, that will most likely also cause your hangi...

Specifically - I hadn't heard of "DSSched" so I searched both the UNIX and the Windows clients for any files of that name, then I looked into the documentation PDFs and didn't find any hits, either. So I am at a complete loss as to what "DSSched" could be but remain very interested.

That is a fair statement regarding 8 - there are so many additional things that happen when creating a project at V8 compared to V7 that going through the API is the only way that it can be done.

It might be useful. What is "DSSched"?

How many rows get processed in DataStage before you get a timeout? I still think that you are doing a big SELECT that doesn't return anything and thus times out. The number of rows in the table isn't the factor here, but the time it takes to deliver might be.

DataStage on UNIX defaults to using cron and crontabs. You can schedule through the director interface or directly in cron.

I assume you are referring to the little graphical objects placed on the links in PX jobs. These show whether repartitioning (which includes going between sequential and parallel type stages) or sorting is taking place on that link.

thumati.praveen wrote:...But I am trying to run the job it is again asking to recompile job like this going on every time...

When it asks you to recompile it usually gives a reason, what is that? Also, are you using one designer instance for both the job and shared container?

Have any rows gone down the ODBC link in that time? If not, it is an inactivity timer and perhaps you could make your job write to a file and then write a second job to load that file to the table.

You waited 19 minutes before asking again? Not even premium service contracts will get you a response within that time frame.