Capacity Backup Upgrade (CBU) capability

This orderable feature can be integrated with Geographically Dispersed Parallel Sysplex. This should reduce disaster recovery times by automating the Capacity Backup Upgrade process at a remote site.

Operationally, the planning considerations and limitations for Capacity Backup Upgrade, for mode purposes, are similar to those for Dynamic Capacity Upgrade on Demand. Planning is simple because reserved logical cores can be specified for logical partitions on the backup system. The logical cores can then be brought online, quickly and nondisruptively.

Concurrent Undo CBU is provided to dynamically remove from the configuration processors that had been added with Capacity Backup Upgrade. The server for disaster recovery with Capacity Backup Upgrade activated can now be restored to its base configuration without requiring a system outage.

When a disaster occurs, Capacity Backup Upgrade is intended to provide the extra capacity without disruption. When the disaster is over and normalcy is restored, Concurrent Undo CBU is intended to allow the system to be returned to its previous configuration without disruption.

Annual testing of a Capacity Backup Upgrade system is highly recommended. Invoking the Capacity Backup Upgrade configuration is nondisruptive and now returns to the original configuration, after the test, without disruption as well.

When a typical Capacity Backup Upgrade occurs, typically logical processors are configured online so that the system closely resembles the failing system. To prepare for Concurrent Undo CBU, the most expedient method is to simply configure offline all those logical processors that were configured online in support of the failing system.

In order for Concurrent Undo CBU to proceed, the restored original configuration must have a physical processor for each online dedicated logical processor. Theoretically, the only additional requirement is that at least one non-dedicated physical processor remains that matches each type (general purpose, ICF, IFL, or zIIP) of online shared logical processors. However, it is highly recommended that shared logical processors be configured offline so that the highest number of online shared logical cores for any active LP does not exceed the number of non-dedicated physical cores remaining. For further guidelines, see the rules governing the number of CPs that can be specified for an activating logical partition on page Number of central processors.