CTMPLEX

Multiple Control‑M monitor configuration under an MVS Parallel Sysplex with a JES MAS (multi-access spool) environment.

Control‑M Global Sysplex Manager (GSM)

The GSM is the one (and only one) monitor that uniquely serves the entire Sysplex. Its main activities include: Select and post-processing phases (not including SYSOUT and DO SYSOUT operations); handling dummy jobs and SMART Table Entities; I/O against the main databases (AJF, CND and RES); processing external events (Hold, Free, Delete, and so on) and executing CMEM/Control‑O requests.

Control‑M's Local Sysplex Monitor (LSM)

The LSM is a Local monitor, started on any Sysplex member (but only one Control‑M monitor on each Sysplex member). This monitor performs the SUB (submitting) and SPY (follow-up on jobs, reading and analyzing of job SYSOUT) phases of each job, including the SYSOUT and DO SYSOUT functions of the post-processing phase.

The first Control‑M monitor that is started in the Sysplex becomes the GSM. Every Control-M monitor started after that acts as an LSM, unless it has a higher PRIORITY, in which case it takes over the role of the GSM, while the first Control-M monitor becomes an LSM. This global monitor selection process continues, so that ultimately the Control-M monitor with the highest PRIORITY value functions as the GSM. Only one Control‑M monitor may run on each Sysplex member.

Coupling Facility (CF) Usage

The Coupling Facility list structure is used as common storage for the jobs, and as the method of communication between the GSM and LSMs. The Coupling Facility is mostly used to store active jobs ( jobs that passed the Select phase – and are Eligible for Run ‑ but have not ended yet).

GSM

See Control‑M Global Sysplex Manager.

Global

See Control‑M Global Sysplex Manager.

LSM

See Control‑M Local Sysplex Manager.

Local

See Control‑M Local Sysplex Manager.

RAS

CTMPLEX is designed to achieve continuous operation (24x7), enable the maintenance of the Sysplex without stopping Control‑M, and minimize the delay caused in cases where a monitor or CPU failure occurs. When the GSM unexpectedly terminates (either abended or its CPU terminates), one LSM (the first to detect GSM failure) switches itself to function as the GSM. When an LSM fails, the GSM takes responsibility over the work that was handled by the failing LSM, and passes it to another available LSM. When the Coupling Facility fails (or is not used), the GSM itself can continue working without the Coupling Facility and without the LSMs (as a regular Control‑M, not in CTMPLEX mode).

Workload Balancing

CTMPLEX is designed to work with or without Workload Balancing between Control‑M monitors. In a Workload Balancing mode, the GSM assigns work to LSMs according to a predefined capacity and the current utilization of each LSM (and the GSM itself), trying to make the utilization percentage equal on all Control‑M monitors. In non-Workload Balancing mode, each LSM "picks up" jobs for processing according to its availability. Initialization parameters and real-time Modify commands may enable or disable Workload Balancing mode and change the capacity for the Control‑M monitors on each Sysplex member.

For each Sysplex member, a relative capacity (parameter RELCAP) and a maximum capacity (parameter MAXCAP) can be specified. RELCAP is used for calculating the utilization of members (for workload balancing). MAXCAP specifies the maximum number of jobs that can be processed by a Control‑M monitor (GSM or LSM) running on the corresponding member.

Example: The administrator can give low priority to batch processing during the day shift by utilizing CPUA with no more than 50 jobs and CPUB with no more than 60 jobs. During the night shift, the administrator can give batch processing a higher priority by changing the number of jobs in both CPUs to 150 concurrent jobs.