The production process
IBM Workload Scheduler production is based on a plan that runs in a production period. The production period is defined by the user when creating or extending the production plan. It can span from a few hours to some days (by default it is a 24 hours period). Before the start of each production period, IBM Workload Scheduler executes a program that creates the production plan starting from the modeling data stored in the database and from an intermediate plan called preproduction plan. Then another program includes the uncompleted schedules from the previous production period into the current plan and another different program logs all the statistics of the previous production into an archive.
All of the required information for that production period is placed into a production control file named Symphony™. During the production period, the production control file is continually being updated to reflect the work that needs to be done, the work in progress, and the work that has been completed. A copy of the Symphony file is sent to all subordinate domain managers and to all the fault-tolerant agents in the same domain. The subordinate domain managers distribute their copy to all the fault-tolerant agents in their domain and to all the domain managers that are subordinate to them, and so on down the line. This causes fault-tolerant agents throughout the network to continue processing even if the network connection to their domain manager is down. From the graphical interfaces or the command line interface, the operator can view and make changes in the current production by making changes in the Symphony file.
IBM Workload Scheduler processes monitor the production control file and make calls to the operating system to launch jobs as required. The operating system runs the job, and in return informs IBM Workload Scheduler if the job completed successfully or not. This information is entered into the production control file to indicate the status of the job.