DSXchange

Too slow.

When you log directly onto the production server with that userid and issue a "crontab -l", what do you see?

Are you sure you are logging into Director with the same userid that scheduled them originally? It reads that user's crontab to show the scheduling information, hence the question.

Because you needed to look for a dat extension rather than csv? Ok. Mine worked fine with both "@RECORD" and '@RECORD' so that doesn't seem to be signficant.

Sure. Enable 'Create table in target database' on the Input link, Update Action 'Insert without clearing'. Then a normal Output link from the stage to do the select.

Odd. The first entry in the log should always be the 'Starting' message, then the 'Environment variable settings one and Auto Purge messages are always posted last in the log. For example: Log prior to last 14 runs cleared by user Auto-purged 105 log entries. How are you starting your jobs? Do you h...

Why guess? You've got a custom after-job routine, something no-one here will know anything about. Find it in the Manager and double-click on it, see exactly how it parses out the argument, what it is doing with the individual components and under what circumstances the error is logged. If you need s...

Doesn't the Oracle Enterprise stage use sqllder under the covers?

Something that may help from Ernie, our XML Guru:

http://dsrealtime.wordpress.com/2007/12 ... -a-source/

All kinds of good content there.

Was your question not answered?

They are already doing a direct path load.

The problem is you can't have (usable) indexes with a direct path load, so you need to remove/disable them first then rebuild them afterwards, hence the message it logged.

The documentation may use Server to illustrate techniques, but the stages work the same, PX to Server. Your first error notes your problem - the lack of an input link, which is required.

Right, you either run the job or get job/log information - one or the other, not both.

Oh, and running "dsjob -jobinfo" is quite different from using it to actually run jobs. And no matter how many times you ask, there isn't a "single command" to do what you want, unless you consider dsjob or the script you write as that single command.

Those are orphaned client connections, not dsjob related and happen when network or pc issues cause the client to not disconnect cleanly. Running too many jobs at the same time causes different problems.

There's no magic command, and "dsjob" isn't expensive. You will need to run it 100+ times for 100+ jobs. Don't worry about magic and just grind thru the calls, it will be fine.