Windows 2000 provides a number of performance reporting objects and counters to help you keep track of disk performance....
Some of the most useful ones involve the disk queues.
For example Disk Queue Length for logical and physical disks gives the number of I/O requests queued for each disk as well as the number of requests in service. Some disk queuing is inevitable, but Microsoft suggests that if the queue never drops below two then there may be a bottleneck developing on the disk. To analyze the queue further use the Avg. Disk Read Queue Length and Avg. Disk Write Queue Length counters.
Another important indicator of disk queue health is the Avg. Disk sec./Transfer counter. This measures the average time of each data transfer, regardless of size. Since this includes the time the request sits in the queue, a high value for this counter can indicate the system is retrying requests. Multiple retries can be due to a disk that is failing, or more commonly, to an overly long queue.
The Disk Transfers/sec. counter gives the number of reads and writes completed per second. According to Microsoft, a value of more than 50 indicates a bottleneck may be developing on the disk.
All these counters can be used on either physical or logical disks. However by default only the physical disk counters are enabled. To apply the counters to logical disks as well, enable the logical disk counters with the diskperf command.
Rick Cook has been writing about mass storage since the days when the term meant an 80K floppy disk. The computers he learned on used ferrite cores and magnetic drums. For the last twenty years he has been a freelance writer specializing in storage and other computer issues.