DSXchange

HFstage1 HFstage2 | ^ | | V | --------> TransformerStage --------> Both hashed file stages refer to the same hashed file. Use an annotation to warn future developers/maintainers that this is the case. HFstage1 performs lookups (cache disabled, lock for update). This sets a row-level lock in the has...

If you see "#" characters in that style of error message it means that a job parameter of that name has not been found/resolved in the job. This probably means that it has been: - omitted - mis-spelled - mis-cased You must make sure that any job parameter reference in the job design - no matter wher...

We have a ton of DataStage jobs...

May I ask how you weigh them?

I doubt it, since there doesn't appear to be anything in the manuals that describe this. Can you post the reply you get from IBM support so that we can know in future? It's probably one of those undocumented thingies.

The server (SDK) key management routines use persistent storage (a hashed file called SDKSequences) to store the next value. These routines can be thwarted by any non-DataStage process adding new surrogate key values to the target table without updating SDKSequences. If you were to implement somethi...

Did you ever issue the UNIX command make love ?

Yes, server only. That's why I left it out of a question about parallel jobs. However, if the tasks are small, there's no reason not to use a server job; they are part of DataStage EE after all.

The only thing that Chapter doesn't cover well - it's elsewhere in the manual - it how node pools and disk pools are used. A stage can be restricted to operate on fewer nodes than are specified in the configuration file, by specifying that that stage will executed in a node pool. For example, you ma...

Job control routines, such as RK has described, will also work to control parallel jobs.

What version of Windows? There are issues - for example with SP2 for Windows XP - for which you may need to get a patch from IBM. Search the forum for more information - "patch" may be a good search term.

Server jobs have the word "Server" on the design canvas, parallel jobs have the work "Parallel" on the design canvas. By they use the same table definitions (though parallel jobs can represent those table definitions in an extra way, as an Orchestrate schema). They use an almost disjoint set of stag...

Welcome aboard. :D We help those who first help themselves. There is an admirable Search facility on this site, and this question has been asked and answered on more than one occasion. Search for your answer first, then seek clarification on anything that might remain unclear. Also search on the ven...

Welcome aboard. :D There is no C stage. There is no Java stage. You've been misinformed as to the names. There are other ways to execute external commands, but from server jobs you are going to need an interlude of some kind (ExecSH before/after subroutine, or DSExecute() function). DataStage server...

Please report whether this works. I have some doubts. On my systems, not all the environment variables in DSParams are set in the dsenv script.

Disk space. The disk you're writing to. If this is a sequential file, then it's the disk where the file lives. If it's Oracle, there are several possibilities; the logs most likely, since much of the other stuff is on raw disk. Get your Oracle DBA to monitor and to advise. Check the available space ...