5.4.3 Estimating processing performance in a disaster recovery environment

(1) Estimating the processing performance for normal operation when disk copy is used

During normal operation, the contents of the shared disk are copied from the main site to the remote site. Therefore, the hardware performance and the method of copying data between disks might affect the processing performance of JP1/AJS3 at the main site.

The effects on the processing performance might also change according to the quantity of jobs and the job output processing. Therefore, execute the maximum number of jobs during the system test, and estimate the processing time.

To Page Top

(2) Estimating the time required for the switchover operation during disaster recovery

The following describes how to estimate the time required to resume JP1/AJS3 operation at the remote site after the main site has stopped because of a large-scale disaster.

The following figure shows the processing from stoppage of operation at the main site to the start of disaster recovery for the JP1/AJS3 service at the remote site for the resumption of operation.

Figure 5‒5: Processing steps from stoppage of main site operation until the resumption of operation at the remote site

As the time required for switching over operation in disaster recovery, estimate the time required for Disk switchover, Startup processing, and Status checking of jobnets and jobs shown in the figure. The following describes how to estimate each time.

Time required for disk switchover

Disk switchover is required to allow the remote volume to be used at the remote site. For details about this operation, see the hardware documentation. You should also estimate the processing time by switching over the disk with the maximum number of jobs running during a system test.

Time required for startup processing

When the JP1/AJS3 service is started at the remote site, the status of jobs running at the main site when the disaster occurred changes to Unknown end status, and the status of a jobnet changes to Interrupted. Recalculation of the schedule is also performed. Accordingly, the time required to start the JP1/AJS3 service is related to the number of jobs that are being executed.

You should estimate the processing time by starting the JP1/AJS3 service with the maximum number of tasks running during a system test.

Time required for checking the status of jobnets and jobs

When the JP1/AJS3 service at the remote site is started in disaster-recovery mode, the status of jobs running at the main site at the time of a disaster changes to Unknown end status, and the status of jobnets changes to Interrupted. Use the log output by JP1/AJS3 or a work application to check the status of jobnets and jobs, and the progress of work tasks.

For details about the statuses of jobnets and jobs when the start mode is disaster recovery, see the description about the disaster-recovery start in 6.2.1(3) Jobnet and job statuses for each start mode in the JP1/Automatic Job Management System 3 Administration Guide.

To Page Top