The storage for a coupling facility LPAR is defined in the same way as a non-coupling facility partition. However, the storage in a coupling facility LPAR cannot be dynamically reconfigured. If another partition on the same processor fails, its storage cannot be taken over by the coupling facility partition. Or, if the coupling facility partition fails, its storage cannot be taken over by another partition.

Coupling facility storage is used to contain both coupling facility control information and application data. In early processors, storage was logically segregated between these two functions. The installation was required to configure storage as either control storage (which could contain both control information and data) or non-control storage (which could contain only data). This distinction is now obsolete, and in current processors all storage configured to the coupling facility is available for either purpose. Effectively, all storage is control storage. However, some APIs (such as IXLMG) return information about the amount of coupling facility storage required for control information.

Beginning with CFLEVEL 19, you can also configure storage-class (flash) memory to a coupling facility LPAR. Storage-class memory provides an overflow capability to minimize the probability of structure-full conditions. As with coupling facility real storage, you cannot dynamically reconfigure storage-class memory, nor can one partition reclaim storage-class memory that is configured to another partition. Use of storage-class memory increases the amount of coupling facility real storage that is required by the affected structures.

Some processors also support storage-class memory, or flash memory. You can configure storage-class memory in very large amounts that are relative to the actual structure size. It is used to provide an overflow capability to avoid structure-full conditions. The installation assigns storage-class memory to a structure by specifying SCMMAXSIZE in the CFRM policy. Storage-class memory can contain both structure controls and data. It is not directly accessible to the application; the coupling facility migrates controls and data between real storage and storage-class memory as necessary to satisfy requests that are initiated in the normal manner. In contrast to coupling facility real storage, storage-class memory is not allocated to a structure until it is required, and it is returned to a free pool when no longer in use. When used, storage-class memory also requires additional coupling facility real storage (augmented space) to allow the coupling facility to track the location of the stored data and controls. The augmented space is not included in the structure size but is instead allocated from the coupling facility's free space on an on-demand basis. Like storage-class memory itself, it is returned to the free pool when no longer required. Since access to storage-class memory is slower than access to coupling facility real storage, requests that must retrieve from or write to storage-class memory might incur a performance penalty. See z/OS MVS Setting Up a Sysplex for additional discussion about the implications of storage-class memory exploitation.

The size of the control structures is affected by the CFLEVEL of the coupling facility in which the structure is to be allocated. Different CFLEVELs will have different control structure requirements, with the result that the same structure could have significantly different sizes depending on the CFLEVEL of the coupling facility.

The DISPLAY CF command displays information about coupling facility storage including the total amount, total in-use, total free control and non-control storage, and storage-class memory. The DISPLAY XCF,STRUCTURE command displays information about coupling facility real storage and storage-class memory that is currently allocated to a structure. .