KBI 310752 Issue Addressed: Tasks Using Shared Monitoring Engine Process Might Be Cancelled Or Delayed

Version

Argent Advanced Technology 3.1A-1310-A or below

Date

Tuesday, 19 Nov 2013

Summary

Argent AT shared Monitoring Engine process recycles every 30 minutes

This is done to overcome errors in the Microsoft PDH (Performance Data Helper) DLL that can cause the PDH DLL to hang

As extra bulletproofing, Argent recycles the Argent AT shared Monitoring Engine every 30 minutes

It will be spawned again when next task is scheduled to run using this shared process

But if a task work order is waiting for the shared process to pick up, it won’t happen until the process is re-spawned

Technical Background

If the shared Monitoring Engine process is recycled, a task work order is waiting it to pick up, and the next task is more than five minutes in the future, the task can be cancelled by Main Engine with log lines similar to following:

18 Nov 2013 13:35:28.529 PANWS A\Administrator About to cancel task (Relator: REL_VM_GUEST_PERF, server: PanA01 – DC – Domain (W2008x32) 70/1). Reason: it has been running for more than 5 minutes

Note:

The issue may happen more frequently when ‘dynamic‘ pool option is used in Relator combined with big pool size, as the next task is more likely not to run on exact same slot in the pool

Resolution

Upgrade to Argent AT 3.1A-1310-B or later