Setting TIMERSECS

As described, the responsiveness of the self-monitoring and automated recovery actions is governed by the TIMERSECS value that is specified on the SYSPLEXMONITOR parameter of the GLOBALCONFIG statement. The default value is 60 seconds. Evaluate whether this default value is appropriate in your environment and, if not, specify a value that is better suited to your environment. The following considerations are some of the considerations to evaluate in selecting an appropriate value:

• Specifying a smaller TIMERSECS value increases the responsiveness of the self-monitoring functions.

  The monitor periodically performs its checks for problem conditions, every 60 seconds or every quarter of the TIMERSECS interval (TIMERSECS/4), whichever is less. Therefore, when a smaller TIMERSECS value is specified, operator message alerts are issued sooner if a problem condition is detected. Automated recovery actions are also triggered sooner if the RECOVERY option is in effect.

  While this might be desirable, care should be taken to not specify a timer interval value that is small enough to trigger recovery actions for conditions that are transient in nature. For example, the value specified should probably be larger than the spin loop timeout interval for the system (SPINTIME parameter in the EXSPATxx parmlib member). Otherwise, a spin loop timeout condition that might be recoverable could trigger an unnecessary TCP/IP sysplex recovery action, as a result of the TCP/IP sysplex monitor not being able to be dispatched temporarily before any spin loop timeout recovery actions can take place. A good rule of thumb is to specify a TIMERSECS value that is at least two times larger than the spin loop timeout interval in effect for the system.

• Specifying a longer TIMERSECS value has the opposite effect, and the self-monitoring and automated recovery actions are less responsive.

  In addition, a longer TIMERSECS value can cause disruptions to existing TCP connections when problem conditions are detected, even if automated recovery actions are initiated. For example, if the problem condition persists for a long enough interval and it impacts the delivery of data on existing TCP connections, the remote TCP/IP hosts might terminate these connections before a recovery action is initiated.

• Review your current settings for the sysplex failure detection interval (INTERVAL keyword in COUPLExx) and the ISOLATETIME value specified in your SFM policy, if defined. These values indicate how responsive the XCF component should be in removing systems from the sysplex when a status update missing condition occurs. As a result, these settings can provide a reasonable reference point for the setting of the TIMERSECS value. For more information on these intervals, see z/OS MVS Setting Up a Sysplex.