After a system running in a PR/SM™ partition is removed from the sysplex, SFM allows a remaining system in the sysplex to reconfigure processor storage for use by the remaining systems.

You specify PR/SM reconfiguration statements on the RECONFIG statement of the SFM policy. PR/SM reconfiguration definitions include the name of the failing system (FAILSYS), the name of the system that is to perform the reconfiguration actions (ACTSYS), and whether PR/SM is to deactivate a specific system or all logical partitions in the addressing range of the acting system (TARGETSYS). You can also specify whether the system performing the reconfiguration is to take over the target system's storage and/or expanded storage. To carry out a PR/SM reconfiguration, ACTSYS and TARGETSYS must be on the same processor. FAILSYS can be the same system as TARGETSYS, or it can be another system, on the same processor or on another processor.

When the TARGETSYS is deactivated, its CPU resources are added to the pool of shared CPU resources, thereby increasing the CPU cycles available to all shared logical partitions. The distribution of the additional CPR cycles is proportional to the weights of the active shared logical partitions. When the TARGETSYS is reactivated, its CPU resources are removed from the shared CPU resource pool.

Figure 1 shows a three-system sysplex, including two MVS™ images (SYSA and SYSB) running in logical partitions on a single processor and one MVS (SYSC) on a single processor. Because an SFM policy is active, XCFPOLxx specifications are not active. SYSB can be configured to take over the resources associated with the logical partition of SYSA if either SYSA or SYSC is removed from the sysplex. This would give SYSB the extra resources needed to handle the work of either SYSA or SYSC.

Figure 1. Three-System Sysplex on Two Processors with Active SFM Policy

For information about the hardware PR/SM feature and the assignment of LPARs, central processors (CPs), and storage elements, see PR/SM Planning Guide.