DSXchange

What is causing the abort in the sequence that tries to reset your failed job? Could you post the error message that is generated?

This approach will, unfortunately, not work due to the way DS jobs are executed at runtime.

What you can do is compute the value of the stage variable in a sequence, then pass it into the job as a parameter and use the parameter value as the initial value for the stage variable.

... But I can't find the hash file that it created. I went to Manager and checked, can't find the hash file... The Manager will not point to the hashed file location, it will contain the metadata for that hashed file if you decide to load it. You can import the metadata from the physical hashed fil...

No, not at the moment.

There have been thoughts about doing this and discussions on the possible structure and content, but this has not been implemented to date.

The FTP stage in DataStage has the advantage that it treats the data on a line-by-line basis, whereas the command-line FTP interface just moves the file as a whole. When using the DS FTP Stage it means that a long-running FTP can be processed row by row as it comes in or goes out instead of having t...

The syntax of the previous posters message isn't quite correct, you can look up the command line in the documentation or just execute dsjob with no options to get the help.
try dsjob -run -mode NORMAL <project> <job>

There are a number of possibilities, but you will need to supply some more information. 1. You posted in the PX but specified a server job. Which type is it? 2. Are you using a NLS DataStage engine and databases? 3. If you read the Sybase source and write to a flat file, is the data correct? (There ...

It looks like you declared 4 columns at 20bytes apiece in your metadata for the fixed width, and this is incorrect and being flagged by DS.

I would be surprised if a sequence that starts 3 jobs concurrently will overload a system - unless there are other things going on that use a lot of resources. In this case we don't know if the system was just less busy after the recompile (which is my guess) or if the object code was corrupted in s...

At first thought I would guess that the job has been corrupted internally. You can make a copy of the job, then remove single components until the error goes away to identify which stage might be causing the problem.

Is the problem that no records are going out of the modify stage or that the output file has records in it when none are written by the job? How many rows are being read by the job? Do you have "append" set to the target sequential file?

Odd, in this case I would have expected to see an entry in the job log for the output of the previous run.

If you replace the stage where you call your job with a dummy stage (i.e. a command stage that just issues an empty command) does the job still fail?

That is a good hypothesis, but would only hold true if it failed with a -14 every time. Usually this error is produced when a machine is overloaded, i.e. when it is so busy that new tasks are not started within DS limits.

Try resetting the sequence and then looking at the entry titled "from previous run..."

p.s. This is a volunteer site, so using the word urgent in a post here usually results in getting a much slower response than normal, and many expert posters here might even decide not to reply.

Let us assume you have a simple job that you declare as multi-instance. It accepts a source file name as a parameter and all it does is read the sequential file, perform some transformations onto the data and load it into a database table. You can start this one job several times in parallel, with e...