DSXchange

When you reset the Job Sequence (that's reset, not re-compile) are there any additional diagnostic messages logged, especially "from previous run" messages? You might also look in the &PH& directory in the project on the server for files containing error information. Use grep to look for words like ...

What I normally do in your original case is to have DataStage generate the data into the base file name, then use ExecSH to change the name of the file to add the date stamp. The command executed is mv /dir/dir/basefile basefile_`date +"%Y%m%d"` You can fiddle around with the date formatti...

I use DB2 and ODBC stages all over the place. We usually have the exact same tables set up in both a "Test" area and a "Production" area. Right now, to switch between the 2, we have to go into each job, switch the "Server Name" and recompile. Is there a way to have that "Server Name" spot read a te...

The other way is to rely on the fact that variables declared to be COMMON are automatically initialized to zero, therefore common /BussOrIndv/ a, aNumVals * only initialize if not already initialized If a = 0 Then * The second field in the array denotes where the substring * will occur: l = Left, r ...

UV_USERS is one of the SQL system tables for DataStage; it records users to whom database privileges (CONNECT, RESOURCE or DBA) have been granted. You require DBA privilege to grant database privileges. To find out who has DBA privilege, execute the query SELECT NAME FROM UV_USERS WHERE DBAUTH = 'YE...

Two things you should do 1. Kill all the processe of all the Jobs attached to that sequencer/Applications from unix level and then clear the status files and RT_CONFIGNNN files for each of the Jobs. 2. You should check whether the log as seen from the director log is taking long time to be reflecte...

What is your new MFILES setting? And what is the NFILE setting in the UNIX kernel? MFILES can not be greater than (NFILE - 8).

There are two main aspects; unit testing of components, then whole-of-job testing. I tend to use the Debugger or stage tracing for unit testing of active stages, and whatever's available (usually Get SQL Info and/or View Data) in passive stages. In both cases, the most important thing to have availa...

10. This means that the writer daemon pauses 100 milliseconds after each write. There are no dependencies, unless you need immediate update of the hashed file, in which case you probably want to use private link cache rather than shared cache. Except, of course, that DCWRITEDAEMON must be 1 for this...

That's just not correct. See my earlier post.

DCWRITEDAEMON should not affect the ability to connect; it's the wait time for the daemon that updates the shared hashed file write cache. In my opinion you're taking a great risk setting it to 10 seconds, (I'd leave it at 10 (=100msec) on your multi-CPU machine). Likewise none of the other paramete...

The only way to connect to SQL Server from a UNIX machine is via the ODBC driver supplied with DataStage. You will need to declare a DSN bu editing .odbc.ini in the Engine and uvodbc.config in the project. On a Windows server, again ODBC is the usual method, though you will also be able to use OLEDB...

The problem is in calling your Routine called MySeq.

Is it compiled? Have you tried to use it as a function without a DEFFUN declaration in the controller code?

When you reset the job, are there any additional diagnostic messages in the log, probably under "from previous run"?

Where it looks like DataStage may be purchased, get onto Ascential for a one-month trial. This includes a Tutorial. They may put a price on the trial copy, which could be discounted from the actual purchase price.

... and set read-only permissions on the share.

That is, it's writable locally but read-only if viewed through the share.

Having writable shared folders is simply inviting viruses and other nasties in!