Running the primary and backup controllers

During normal processing, the primary controller sends all the data and plan updates to the backup controller, which in turn applies them to its databases.

For every event received from the primary controller, the backup controller checks that the information is aligned by browsing the following information:
  • Run number and TOD ID of the CP (panel 6.6 EQQGSCPP) and run number and latest update of the LTP (panel 2.4 EQQLSTAP).
  • Sequence number of the event that was received.
  • Backup checkpoint data set information.
If changes to databases occur while the connection between the primary and backup controllers is lost, warning messages are logged in the EQQMLOG of the primary controller. Because the sending of databases to the backup controller is not automatic, you must manually start the following transfer procedures after replacing EQQ with your controller subsystem name:
  • EQQSAD
  • EQQSOI
  • EQQSRD
  • EQQSSI
  • EQQSWS
At the end of the process, the backup controller is notified of the result and the database restore procedures are automatically started.
Note:
  • Before you transfer the EQQSIDS data set, run a current plan backup (for details about a current-plan backup process, see The current-plan backup process). Do not transfer the EQQSIDS data set during the daily plan activities, because any updates might be lost.
  • The mass update and batch-loader utilities are not propagated to the backup controller through events, because they directly update the AD VSAM file without logging each single update. For this reason, after you run the mass update or batch loader program you are required to submit the job that sends the AD file to the backup controller.

If the primary controller and backup controller run in different time zones, to ensure that the backup controller scheduling respects the time dependencies by applying the same time of the primary controller, customize the time-dependent-operation exit (EQQUX014) appropriately. For details about the exit EQQUX014, see Time-dependent-operation exit (EQQUX014).

If the connection between the controllers is lost at any time, a re-synchronization process is performed as soon as the connection is established again. The backup controller sends a re-synchronization request to the primary controller, and the primary controller sends all the events that were lost or, if the lost events no longer belong to the active JT, sends the current plans.

Because the MODIFY command is not supported on the backup controller, to update any initialization statement on the primary controller, complete the following procedure. For detailed information about the modify commands, see Modifying the scheduler.
  1. On the backup controller change the initialization statements that you want to modify on the primary controller.
  2. Stop and restart the backup controller.
  3. Modify the required initialization statements on the primary controller.
  4. Stop and restart the primary controller.