DSXchange

chandra,

You MUST use the DataStage Director tool to stop a job. If you kill the processes, the internal status in the DataStage engine is not updated.

You can reset the job status from DataStage director, if you have "Enable Job Administration in Director" checked in the project properties within DataStage Administrator.

To reset the job status from DataStage Director, 1) select the job, 2) goto the Job menu and select the "Clear Status File" option. NOTE: this is NOT something you want to do if the job really is running...

Please read the help files or books that you have on DataStage Director before you decide to do this.

Good luck,
Tony

Chandra,

The proper way to stop a job is to press the Stop button (the little square) or Select Stop from the Job menu within DataStage Director.

Tony

Using a unix kill to stop anything is the LAST option you should ever think of employing. If you have a job that is running long because of a query, then stop the query. The job will gracefully top.

If you have to stop a job because it's crunching too long and you want to abort the run, use ANY of the MANY ways DataStage supplies stopping a job. Kill is the last option, because of the nasty things that you have found out about.

Kenneth Bland

What you see in Director as the status of the job is a column from the RT_STATUSnn table for that job. This is updated periodically by the running job. If you kill the job, the "status" column can not be updated by the running job. This is why the job appears still to be running.

This is also why everyone has correctly advised that the correct way to halt a running job is not to kill the process.

There is another reason. Even server jobs run as more than one process; for example active stages (such as Transformer stages) run as child processes of the job itself.

If you must stop the job from the server, use the command line interface (dsjob) with the -stop option rather than the -run option.


Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518