DSXchange

You don't necessarily need to 'calculate' this - look at the actual processes the job generates when it runs. There will be at least two pids: DSD.RUN for the job itself DSD.StageRun for each active process This would be checked via a 'ps' and 'grep' on a UNIX Server, not sure on Windows - perhaps v...

Get bigger rollback segments. Take less time with your query. Run the query in 'off hours' when the source data isn't changing. Any or all will help.

For the first, talk to your DBA. Heck, any Oracle error should be walked by your DBA first if you are unfamiliar with it.

talsimad wrote:i already do it, i have partionned with the round robin method but it seem that the problem presist

Then, perhaps, you don't already do it.

nkln@you wrote:But I do get DSR.ADMIN: Failed to enumerate local groups.

I've seen this when there is no .developer.adm file in a Project. As noted, it's not a major issue, more of an annoyance.

That's not a DataStage thing but rather a perfectly normal SQL thing. They are 'bind variables' and are placeholders for the values your job will supply, which will bind two columns - one to each question mark in the order shown. Some stages (like OCI) will use numbered bind variables - :1, :2, etc ...

Rethink your approach - use the published APIs. There are 'DSGetLog' functions you can use from a job to accomplish this and command line equivalents as part of the 'dsjob' command if you prefer a scripted approach.

You also need to uncheck your 'Disable BBCode in this post' option.

I just mean that the import of SP metadata isn't absolutely required for use by the STP stage.

I thought information of that nature would start shipping with the 8.x releases?

For older versions, you could get the UniVerse User Reference or UniVerse Administration manuals from the IBM web site.

It's one of the many APT environment variables, from what I recall.

Ok then.....One solution I can see is that I will put all the Jobs in one sequencer in which I will keep the option as 'Reset if needed then run' and then Schedule it through the director. Even if the Jobs fail the sequencer will show status as "Finished"( so no need to compile ) and since I have s...

Or 'source' from a Transformer with the proper metadata, constrained to @FALSE that does the Clear or delete and recreate that you need.

Nah... I'd wager it's a settings issue as Ken notes. You usually have to tell your process to 'PRESERVE_BLANKS' in some fashion otherwise they typically get stripped. Then your space gets turned into an empty string and blows the NOT NULL constraint.

More information please.

Are you sure you have a file with an EOF as the record delimiter? That doesn't make a whole lot of sense.

ml wrote:the uvconfig file is placed in the bin folder.

No, it is not. It lives in the 'engine' directory, a.k.a the directory pointed to by the $DSHOME environment variable.

And yes, you need to do this with DataStage shutdown.

What exact version of DataStage? What server platform? If you're not on a recent 7.5.x version, you'll need to upgrade. From what I recall, earlier implementations were a little... buggy. If you are - or perhaps in general - you should contact IBM Support and open a case. As you've noted and from wh...