Restarting system resource recovery (ARMBSRR) jobs

Overview of recovery for a set of ARMBSRR jobs

The job cards for the jobs in the set must contain a symbolic variable that allows BMC AMI Recovery Manager to number the jobs. BMC AMI Recovery Manager uses numbers 0 through n, where n is the maximum number of jobs into which BMC AMI Recovery Manager is able to split the recovery. BMC AMI Recovery Manager also imbeds synchronization steps within the JCL. These steps execute the ARMBSYN program, which updates and monitors the job synchronization file. The first job of the job set is Job 0, which allocates the synchronization file and then submits recover Jobs 1 - n to the internal reader. The first recover job (Job 1) submits an additional cleanup job that waits on all of the recover jobs to complete. If all jobs complete successfully, the synchronization file is deleted by this cleanup job.

Rerun or restart?

If any of the generated jobs fail, you must first determine what caused the failure and correct the situation. Then you should decide whether to rerun the entire job stream (by resubmitting the generated JCL) or restart the jobs at the point of failure. BMC AMI Recovery Manager provides an EDIT macro called ARMSBGEN to assist in restarting the failed jobs (see Restarting-synchronized-jobs). Sometimes it is quicker to resubmit the generated JCL than to identify step restarts for each recovery job.