Hi Liwei,
thanks for your reply. After digging into the problem and spending a lot of time i figured out that i forgot to set the "CanPreempt" flag while submitting our jobs.
This, combined with our activation filter logic, resulted in multiple filter activations because once a task finished, if a job with higher priority than the current one was queued, the following task first entered the queued state and then the
running one, firing again the filter.
Sorry for the time you spent replying me... i forgot to close this post....but we're really excited because the new release of our product will be soon compatible with Windows HPC Server...
Thanks for your time...
Gian
Gian