Hi all,
Im new to datastage and i was wondering if there is anyone who can offer me a platform on what the best practices of the tool is in terms of DS_Projects management, parameters and or executions. any other information will be greatly appreciated.
Tunde
Tunde
Best Practices
Moderators: chulett, rschirm, roy
-
ray.wurlod
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
The "normal" stream of DataStage training from Ascential is:
DS304 DataStage Essentials
DS305 Advanced DataStage
DS306 DataStage Best Practice
I think you can get a package deal, even if you take these by what they term "e-learning". Check out their website or contact them for more information. (I am not an Ascential employee, nor ever have been.)
For a quick answer, you have an effectively unlimited (999) number of projects. I would suggest at least four; one for development, one for testing/QA, one for version control and one for production. In addition, you can create "sandbox" projects where individual developers can experiment.
Parameters provide insurance against things that might change over time, including data source names, pathnames, passwords and so on. If anything is likely to change, no matter how unlikely, make it a job parameter. I thoroughly recommend Parameter Manager for DataStage (see link at the top of your browser).
There are several schools of thought about execution models for DataStage, and it really depends on what else your system is doing. Some sites like to have a "master control" DataStage job that runs almost all the time and looks after everything pretty much automatically. Others like to incorporate DataStage in a wider scheme of things, running jobs from another tool (such as a third-party scheduler such as Control-M via the DataStage command line interface (dsjob) or the DataStage API). Still others rely upon DataStage's own scheduling capability, which leverages the operating system's scheduling service (cron in UNIX systems, AT in Windows systems).
Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518
DS304 DataStage Essentials
DS305 Advanced DataStage
DS306 DataStage Best Practice
I think you can get a package deal, even if you take these by what they term "e-learning". Check out their website or contact them for more information. (I am not an Ascential employee, nor ever have been.)
For a quick answer, you have an effectively unlimited (999) number of projects. I would suggest at least four; one for development, one for testing/QA, one for version control and one for production. In addition, you can create "sandbox" projects where individual developers can experiment.
Parameters provide insurance against things that might change over time, including data source names, pathnames, passwords and so on. If anything is likely to change, no matter how unlikely, make it a job parameter. I thoroughly recommend Parameter Manager for DataStage (see link at the top of your browser).
There are several schools of thought about execution models for DataStage, and it really depends on what else your system is doing. Some sites like to have a "master control" DataStage job that runs almost all the time and looks after everything pretty much automatically. Others like to incorporate DataStage in a wider scheme of things, running jobs from another tool (such as a third-party scheduler such as Control-M via the DataStage command line interface (dsjob) or the DataStage API). Still others rely upon DataStage's own scheduling capability, which leverages the operating system's scheduling service (cron in UNIX systems, AT in Windows systems).
Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518