1. Implementation Guide
  2. Implementation
  3. Batch Framework

Batch Framework

RPASCE operations require that the administrative user, who will not have command-line server access, must be able to select, initiate, and schedule RPASCE batch activities.

The RPASCE platform includes an Online Administration Tool (OAT) capability, which allows simple parameterization and scheduling of pre-configured batch tasks. The RPASCE provides an enhancement to the OAT framework that allows a sequence of several batch tasks to be defined. This sequence is built from a list of available batch services, such as Measure (fact) data loading, calculation, segment workspace refresh, and so on. These service tasks run in a defined order, so that you can know, for example, that your daily data updates have been loaded before your workspace refresh tasks are run. The batch tasks are configured to run under the existing OAT framework, so that scheduling them to run once, or on a repeating basis, is the same as for other OAT tasks.

The batch task sequences are defined in a small set of text files, which are specified below, with some examples.

Batch Processes Under the Control of the Implementer

Using the RPASCE batch execution framework, the following are under implementer control:

  • List of batch operations to be run, with available parameterization

  • Order in which batch operations are to be run

  • Scheduling of one or more recurring batch tasks, which can be modified by the administrator, as needed

Batch Processes Not Under the Control of the Implementer

Due to the operational and security constraints of the Cloud Service environment, the following are not under implementer control:

  • Parallelization: The applications automatically parallelize any applicable batch tasks with a number of processes set to match the provisioned server environment.

  • Script file names, file and directory locations: Custom scripting is not supported for this environment, and no knowledge of file system names or locations is necessary in defining and parameterizing the batch task files.

  • Incoming and outgoing file locations: These details are fixed within the RPASCE Cloud Service environment.

POM Jobs and Batch Exec Service

Process Orchestration and Monitoring (POM) is the enterprise batch-scheduling solution for retail applications migrating to the cloud. In POM, the RPASCE schedule contains multiple static jobs for each RPASCE application, namely, three daily jobs (jos_daily_pre, jos_daily, and jos_daily_post) and three weekly jobs (jos_weekly_pre, jos_weekly, and jos_weekly_post).

These jobs must exist in batch_exec_list and can map to other batch control sets for the Batch Exec Service.

For example, the following section in the batch_exec_list defines the mappings for the POM jobs.

# Entries for scheduling in JOS/POM
jos_daily_pre    | exec | *hook_jos_daily_pre
jos_daily        | exec | batch_daily
jos_daily_post   | exec | *hook_jos_daily_post
jos_weekly_pre   | exec | *hook_jos_weekly_pre
jos_weekly_pre   | exec | batch_weekly_pre
jos_weekly       | exec | batch_weekly
jos_weekly_post  | exec | *hook_jos_weekly_post

In this example, the job_weekly is configured to execute the batch_weekly control set, which is a set of tasks within the batch_exec_list that can be configured in the same way as you would if not using POM. The "*" before the control set name indicates that this control set is optional, that is, if it does not exist, the batch execution of the POM job just ignores it without reporting any errors.

For more details about the RPASCE Schedule for POM, refer to the Oracle Retail Predictive Application Server Cloud Edition Administration Guide.