DSXchange

1) Could this be a NULL value in a column of type ustring (Varchar) where the column is set to Not Nulls? 2) This could be that the DB2 instance Owner does not have permissions to the DataStage directories that are used during the Load. OR 2) This could be due to User defined SQL being used that has...

Yes Ray, I'm not sure of the answer ... but I'll have another go As I understand DSCAPIOP is used with the DataStage Server. As the sybase oc is a plug-in, I guess it's not a True Parallel component and contains some Server parts. So this indicates that the error is coming from the sybase stage. (I ...

What do you mean by "What authorities" ?

(1,2,3,4) All of these only need to be run once by the user who will be running the job (requiring access to DB2)


5 you do not need to give admin writes to the database for the user to run jobs.

DataStage Server Functions used in the Server Job Transformer (BASIC Transformer) are different from the avaliable functions found in the Parallel Transformer (from a parallel job). You can however place the server transformer on the parallel canvas which works like having a server shared container ...

Does this happen for every project?
Does this happen if you connect to a different DataStage Server?

Are you trying to run the job across multiple machines? (mpp)

Are you using server components (server shared containers/ server transformer) on the parallel job canvas?

T42 wrote:Actually, you should ensure that DataStage starts in KSH

But the PWD message is because Ksh has been used.

If you run DataStage under sh you won't see this message.

Write Method = Load

After the job has failed - if you set the 'write method' property back to 'Load' does the job work again?

If not then what is the property "Write Mode" set to ?

You could use a Server Transformer (including the routine) on the Parallel Canvas?

LD_PRELOAD has nothing to do with running shell scripts from DataStage jobs.

Does the script run OK from the command line? (whats the return code)
Are you using the same user as the DataStage job?
Are you adding the full path to the script name in your DataStage job?

First make sure you 'Force' re-compile the job using the option:
File -> Force Compile

Oh, Are you sure it said *.o and not *.so?

Are you using any routines or custom code within the Transformer?

This sounds like an invironment issue, and the libraries are not being loaded in the correct order etc ... It would help to know what Unix Platform (OS) you are running and the Revision number. (My guess is that you are running on HP-UX) You need to look at the PATH and library environment variables...

Mixing job types by placing server containers within a parallel job will cause a performance hit (when compared to a 'pure' parallel job) and should be avoided where possible.

I think the standard lookup options in PX will read the lookup table once into memory and then perform the lookups (from each node) against the data in memory.

pcmbalaji wrote: Is there any document which lists out pre-requisites for parallel extender ?

Regards,
Balaji

erm ... In the install guide - Look in the unix install software requirements