Sequential Jobs

Creating Jobs for Sequential Processing Solutions

You can create EVM jobs as part of a sequential processing system. For example, assume that you are implementing EVM to automate backup procedures at a site that has limited resources. You have one StorageWorks subsystem and it has a limited amount of free space. There is only one tape backup unit.

Your solution is to create EVM jobs that will be run sequentially from a script file or scheduler. Perhaps you design two EVM-backup jobs to replicate all of the storage units. Each job replicates units whose total capacity is permitted by the free space on the subsystem.

 

You set up the script file or scheduler to run EVM-backup job1, followed upon completion by EVM undo-job1 (see illustration above). Undo-job1 creates free space for the next job. When undo-job1 has completed, the script file or scheduler runs EVM-backup job2, followed upon completion by undo-job2. In this manner, limited resources can be kept active around the clock.

When the snapshot replication method is available, a sequential backup job series will essentially take only as long as the tape backup procedures require. Unlike the clone method, there is no time required to copy all of the data on the source unit and normalize the members of the mirror set.

The sequential solution is also useful when job outcomes are not predictable. By designing jobs and script files to use return codes, you can establish conditional relationships and dependencies between jobs.