Hi Franky,
This value is the same as what you would normally see from the system
perfmon counter "% Disk Time" for the LogicalDisk object. This attribute can
sometimes have a value greater than 100% when you have SCSI disks and use
RAID controllers. This is because % Disk Time counter also includes time for
overlapping IO requests and indiciates the queuing time for servicing that IO
request. You should take a look at "Average Disk Queue Length" and "Average
Disk second / {Read | Write | Transfer}", which are much better indicators of
your unlying storage subsystem performance.
Also refer to this excellent Performance Tuning Guideline for Win2k8,
http://www.microsoft.com/whdc/system...f_tun_srv.mspx (take a look
at the secion on "Performance Tuning for Storage Subsystem").
"Franky" wrote:
Quote:
I've started the system data collection sets, and they've been active for a
few weeks now.
But when I look into the table snapshots.performance_counter_values and
select the records which correspond to the instances
object_name='LogicalDisk' and counter_name='% Disk Time' I retrieve some
records which have values up to 700.
Can someone explain this. |