Getting started with grid project design

When you design grid jobs, you edit the master template using the IBM® InfoSphere® DataStage® Administrator client. The master template defines the static resources that are available in your grid. You can then further define resources for individual projects by generating a project template.

The following list defines terms used in this section:

Master template
The master template defines the static resources that are available on your grid. There is only one master template for each grid (master_config.apt). Only the administrator can edit the master template.
Project template
A project template represents a subset of the static grid resources that have been assigned to a particular project. The administrator defines which static resources to use for the project, and then uses the InfoSphere DataStage Administrator client to generate a project template. The project template contents are populated to the Grid job properties page. Immediately before job runtime, the Administrator client generates a job configuration file. For examples and more detail about templates, see Configuration file templates.
Static resources
Static resources consist of fixed-name servers such as database servers, SAN servers, SAS servers, and remote storage disks.
Dynamic resources
Dynamic resources consist of compute nodes that are in the default compute node pool and that are assigned dynamically at job runtime. A default compute node entry ($HOSTNAME) is included in the master template to represent a generic compute node. If selected by the administrator for a project, the default compute node entry is included in the project template. After the requested resources are allocated, the engine uses the default compute node entry and any user-defined job requirements to dynamically generate the configuration file.

Grid resources for a job are resolved as described in the following overview procedure. As indicated in the steps, some actions require the administrator-level permission, some require only the user-level permission.

  1. The administrator defines static resources in the grid in a master template.
  2. The administrator defines which grid resources are enabled for a particular project by selecting which resources to include in the project template.
  3. The user defines resource requirements for particular jobs within that project.
  4. The user defines resource constraints for particular stages within that job.
  5. After the job is invoked but immediately before it runs, the resource manager calls InfoSphere DataStage to dynamically generate a configuration file.

By this method, InfoSphere DataStage can assure that jobs run immediately when enough resources of the correct types are available, instead of waiting for a fixed resource that might be tied up on another long-running job.