DSXchange

Is there any downside in runing load using DB2 API stage in parallel Execution mode? The target DB2 is on mainframe.

I doubt if DB2 API stage can execute in parallel. As far as I know only DB2 Enterprise stage can execute in parallel. Please check the documentation.

mavrick21 wrote:I doubt if DB2 API stage can execute in parallel. As far as I know only DB2 Enterprise stage can execute in parallel. Please check the documentation.

At the very least, nothing prevents from setting it to run in parallel. And according to DataStage framework, the stage that can be set to run in parallel will initiate process for every node. I did not check it because so far I have no target and no data. I am just considering if my jobs should be built that way. Any additional information on the topic will help. I know that similar thing can be done with ODBC stage, but API stage have better controls.

If it runs in parallel then there will be more connections to the database as each instance of the operator will have it's own connection. This can be good or bad depending upon the database server and it's configuration (max sessions per user, max sessions total for instance, number of nodes, partitioning). The more sessions your job uses, the less that are available for other users, so usage patterns also come into play. People may notice if you monopolize the system.

You may wish to discuss the options with your DBAs to work out an equitable usage scenario.

Regards,

At the very least, nothing prevents from setting it to run in parallel. And according to DataStage framework, the stage that can be set to run in parallel will initiate process for every node.

Did you get a chance to look at the Score? As far as I know for few stages even if you set it for parallel execution, DataStage would execute it sequentially. This can been verified by looking at the Score.