DSXchange

Please post all settings you have used in your UniData stage.

When you refer to "repository", I trust you don't mean DataStage repository (which is not based on UniData).

Open the stage type's properties in the Stage Types branch of the Repository to determine its active/passive status.

LOCATE, FIND and FINDSTR statements can be used in routines to search for delimited substrings.

INDEX() searches for a particular occurrence (for example the first) of a given substring.

Why could something that's an unchanging fact be urgent?

The answers can be found via search - I have posted them in the past.

For example the answer to "wether (sic) dsapi_slave invokes of both active / passive" is no.

All instances will need to take short-lived locks on the same RT_CONFIG file during startup. See if you can spread the startup requests by a small amount, say five seconds apart. SLEEP 5 will do it for you.

Welcome aboard. :D There's no real differentiation. Most estimation schemes talk about "the ETL processing" rather than isolating each phase. Since you're writing to a persistent target, I guess you'd have to regard the job as performing a load. Presumably it's getting those data from somewhere (ext...

Is this a server job question or a parallel job question?

Check the permissions to every directory in the path /app/dstage/Ascential/DataStage/DSEngine/Projects/edw

The danger of using PerfMon is that it can generate lots of wonderfully detailed statistics that don't really reveal that much, but which PHBs can react to. System Tuning 101 A system is a set of finite resources (CPU, memory, I/O bandwidth, swap space, network bandwidth, semaphores, to name a few)....

Did you notice that the original poster is on Windows?

There is a very simple approach using stage variables. It may be posted on the server forum, but it will work just as well in parallel jobs. You just have to make sure that the data are partitioned on the first column values so that all values of each distinct value are on the one partition.

Not really. Find a starting point. Under reproducible conditions gradually increase till system starts paging, then back off a bit. (Highly technical, I know, but it's an old and honourable technique!)

Not to mention the odd well-intentioned old cynic asking folks who claim that there are performance problems to define exactly what performance means in an ETL environment! There's a story about selling software into a certain country where it is axiomatic that you will be greeted with "Performance ...

The job might've gotten hung up trying to process itself.

What did your search for "vertical pivot" on this site reveal?