Version

Argent Advanced Technology 3.1A-1401-E or below

Date

Monday, 3 Mar 2014

Summary

When Relator using dynamic scheduling has execution interval more than 3 hours, the tasks may be randomly skipped for some monitored machines

For example, a Relator has interval of 4 hours started from midnight, it may be executed randomly with interval 6-8 hours

Technical Background

Argent AT treats the monitored machines as a batch to spread across the time line of specified interval

The batch has a batch start time based on when Argent AT service calculates the schedule

Argent AT also recalculate Relator schedules once a while even Relator is not changed

By default, it is once every 3 hours

When Argent AT recalculates the schedule, it recalculates the batch start time too

As a result, the existing schedule may be pushed off

For example, when Argent AT engine is about to recalculate schedule, a machine of the 4-hour monitoring interval scheduled to be monitored in one hour time, the next runtime becomes 3 hours away after recalculation

Things get even more complicated if there are multiple recalculations during the Relator interval

Resolution

Upgrade to Argent Advanced Technology 3.1A-1401-T2 or later

Customer cannot upgrade immediately, he can edit registry HKLM\Software\{PRODUCT}\MAX_INIT_INTERVAL to 0 so that the automatic schedule recalculation is disabled

Note: this workaround can alleviate the issue but it does not completely correct it

In case that Relator, licensed nodes or CMDB-X is changed, the schedule will be always recalculated