IWMWSYSQ – Querying System Information

This service queries information about the systems in the sysplex that are in goal mode. The Query System Information service, IWMWSYSQ, returns a list of systems running in goal mode and information related to available CPU capacity and resource constraints.

The caller of IWMWSYSQ must provide storage to contain all of the system information. This storage area must reside in the caller's primary address space.

It is possible that the storage required by IWMWSYSQ may change such that multiple IWMWSYSQ calls are required to obtain data. IWMWSYSQ users should take this into account when obtaining the amount of storage that the IWMWSYSQ service can use.

If the caller does not provide enough storage to contain all of the system information, this service will return a return/reason code pair indicating that the input SYSINFO_BLOCK is too small. Output data about the amount of storage required (QUERYLEN) will be set to reflect the required SYSINFO_BLOCK size. However, no system capacity information is returned.

Applications that schedule work across multiple systems in an MVS™ sysplex can use this service to

• Locate the "best" (fastest or most idle) system in a sysplex for scheduling specific work
• Avoid scheduling additional work to systems already critically overloaded
• Factor WLM business importance level information into scheduling decisions

The output of this service is a data area mapped by the IWMWSYSI macro, that provides a point-in-time snapshot of each system WLM is managing in goal mode within the sysplex. A scheduling application can interpret and use this information to schedule one or more types of work to systems with specific operating characteristics. Some examples of operating characteristics you can identify with IWMWSYSQ are

• FASTEST CP SPEED - use the IWMWSYSI data area to identify the system having the fastest single CP speed.
• MULTI-PROCESSING CAPABILITY - use the IWMWSYSI data area to identify the number of online CPs on each available system.
• IDLE CAPACITY - use the IWMWSYSI data area to identify the system with the greatest idle capacity.

If a scheduling application can identify the IMPORTANCE LEVEL of the work it schedules the application can use IWMWSYSI to select the most appropriate system. IWMWSYSI provides a vector containing the amount of capacity consumed at each importance level on each system. Thus, if an application is scheduling importance level 3 work, it can use IWMWSYSI to identify the system that has the most capacity that is either idle or is handling importance level 4 or lower work.

An important use of a scheduling application is to avoid placing additional work on systems experiencing contention. IWMWSYSI provides an indicator for each system that, if on, signifies that the system should be avoided for scheduling additional work. This contention indicator is set if a auxilliary storage, fixed storage, or SQA shortage exists. Also, if work to be scheduled may consume large quantities of CSA, you can use IWMWSYSI to determine the amount of CSA and ECSA that is available on each system.

Note that:

• Multiple applications may simultaneously use the same IWMWSYSQ information to make work scheduling decisions. These multiple applications will have no direct cooperation and will compete for the available systems. It is recommended that before an application schedules a large amount of work it activate a small quantity of work, wait for a built-in delay, and then use IWMWSYSQ to determine the effect of the added work before scheduling the additional work.
• Field SYSI_CPU_UP in macro IWMWSYSI returns the speed of an individual CP on the system, adjusted to compensate for MP effects. However, the CP speed of an LPAR is influenced by the CEC LPAR configuration. Variables such as the number of LPARs, the CP mode (shared or dedicated), capping controls, and the logical to physical CP ratio will all infuence the actual CP speed. SYS_CPU_UP is not adjusted for such LPAR configuration effects and therefore the actual performance may differ. See the PR/SM™ manual for more details.