Fine points of how SYNCVAL works

What happens if I specify SYNCVAL in a policy statement for a check that's already running? If you specify and activate a policy with a SYNCVAL value for a check that is already running, the SYNCVAL initial start time value is used only as a synchronization point for subsequent check iterations until IBM® Health Checker for z/OS® is restarted or the check is refreshed.

I want to manually run a check that has a SYNCVAL specified for it. What happens to my synchronization? Nothing! Your synchronization is intact! If you issue a MODIFY hzsproc,RUN command against a check that has a SYNCVAL synchronization point, that manual check run will not interfere with the SYNCVAL synchronized check run schedule. The check runs when run manually and then at its next already scheduled run time.

My check missed the initial SYNCVAL time. When will it run? Lets say you specify SYNCVAL={hh:mm|*:mm) value for a check and activate the policy statement, but the check run misses the specified first run time for some reason (for example, if the check is added inactive, gets deactivated, or z/OS UNIX System Services is down). What happens in this case is that the start time is moved out to the next possible start time that matches the SYNCVAL setting. Note that the start time does not simply move to the next SYNCVAL-synchronized interval instead.

Need an example? Okay, lets say CHECKA needs z/OS UNIX System Services to run and so is was added with ADD CHECK ... USS(YES). Now you define and activate a policy for the check, as follows:

ADDREPLACE POLICY(policyname) STMT(GLOBAL)
           UPDATE CHECK(CHECKOWNERA,CHECKA)
           SYNCVAL(12:00) INTERVAL(00:15)

However, z/OS UNIX System Services is down at 12:00 noon, so CHECKA can't run, missing the noon start time. CheckA won't run at all until noon of the next day.

Gotcha - don't make your INTERVAL shorter than the check run time: If you specify an INTERVAL that is shorter than the check run time, note that you might miss a SYNCVAL point. In this case, the system schedules the check on the next possible SYNCVAL point.

If the INTERVAL is too short with respect to the running of the check, you might miss a SYNCVAL point, in which case the system will schedule the check on the next SYNCVAL point (this is simlar tothe "start time is moved out" case. So maybe it doesn't really need saying (it's really a "don't do that").