How many megabytes of Coupling Facility (CF) storage are required for the CA TPX list structure?
Determining the required storage involves a number of factors. Use the formulas provided by IBM for a list structure with named entries and adjunct data. (This can be different depending on the CFLEVEL and CF model in use.) For your hardware and software levels, see IBM documentation S/390 PR/SM Planning Guide (GA22-7236-04) or equivalent.
Determine the maximum number of members (instances of CA TPX) that can use this generic name. This number plus one gives you the maximum number of lists in the structure (commonly referred to as the list count). This is also the number of lock entries required.
CA TPX uses named entries without keys and allocates adjunct data for each entry. A single list set entry (with adjunct data) is used for each unique user that is logged on to CA TPX. Data elements are not used for unique users. Mult-users require a single list set entry for all instances of any one user ID with adjunct data and one or more data elements, which are defined with a size of 256 bytes and are allocated in 256 byte increments. The number of data elements associated with a list entry is a function of the number of instances of any one Mult-user and is limited to 16, which limits the maximum number of instances of any one Mult-user to approximately 341.
The amount of storage used in the data element is approximately 12 bytes per Mult-user instance. However, the element and entry ratios are set to one to minimize data element space and maximize list entry space in the structure.
These calculations will give you a close estimation of the storage required. We recommend that you monitor structure utilization, because CA TPX or IBM software maintenance and changes in your environment can affect this.
When CA TPX initializes, it issues message TPXL5101 to the CA TPX log. The message indicates the results of the connection attempt to the coupling facility structure and contains the approximate number of entries the structure will support as reported by the XES subsystem.
In addition, CA TPX provides a sample program (and the necessary JCL to invoke it) that issues the IBM IXLCSP macro. Written in assembler, the sample program asks you to provide input parameters, which are then used to calculate the number of required entries; the program then invokes the IXLCSP macro to obtain feedback from the XES subsystem regarding the approximate size of the structure required. See delivered samples in CB0VSRC with member names starting with CF*.