Preproduction plan

The preproduction plan is used to identify in advance the job stream instances and the job stream dependencies involved in a specified time period.

This improves performance when generating the production plan by preparing in advance a high-level schedule of the anticipated production workload.

The preproduction plan contains:
  • The job stream instances to be run during the covered time interval.
  • The external follows dependencies that exist between the job streams and jobs included in different job streams.
A job or job stream that cannot start before another specific external job or job stream is successfully completed is named successor. An external job or job stream that must complete successfully before the successor job or job stream can start is named predecessor.
IBM Workload Scheduler automatically generates, expands, and updates, if necessary, the preproduction plan by performing the following steps:
  • Removes the job stream instances in COMPLETE and CANCEL states.
  • Selects all the job streams scheduled after the end of the current production plan and generates their instances.
  • Resolves all job and job stream dependencies, including external follows dependencies, according to the defined matching criteria.
To avoid any conflicts the database is locked during the generation of the preproduction plan and unlocked when the generation completes or if an error condition occurs.

At this stage only the job streams with the time they are scheduled to start and their dependencies are highlighted. All the remaining information about the job streams and the other scheduling objects (calendars, prompts, domains, workstations, resources, files, and users) that will be involved in the production plan for that time period are not included, but are retrieved from the database as soon as the production plan is generated.

When the production plan is extended, old job stream instances are automatically removed. The criteria used in removing these instances takes into account this information:
  • The first job stream instance that is not in COMPLETE state at the time the new plan is generated (FNCJSI). This job stream instance can be both a planned instance, that is an instance added to the plan when the production plan is generated, and a job stream instance submitted from the command line during production using the conman sbs command.
  • The time period between the time FNCJSI is planned to start and the end time of the old production plan.
Assuming T is this time period, the algorithm used to calculate which job stream instances are removed from the preproduction plan is the following:
if T < 7
All job stream instances older than 7 days from the start time of the new production plan are removed from the preproduction plan; all job stream instances closer than 7 days to the start time of the new production plan are kept regardless of their states.
if T > 7
All job stream instances older than FNCJSI are removed from the preproduction plan; all job stream instances younger than FNCJSI are kept.
This algorithm is used to ensure that the preproduction plan size does not increase continuously and, at the same time, to ensure that no job stream instance that is a potential predecessor of a job stream newly added to the new preproduction plan is deleted.

For more information about how you can open the preproduction plan in view mode from the Dynamic Workload Console, see Display a graphical preproduction plan.

Note: In the IBM Z Workload Scheduler terminology the concept that corresponds to the preproduction plan is long term plan (LTP).