Troubleshooter: Evaluating Proper Disk Queue Length

I've seen varying recommendations regarding the appropriate average disk queue length for my physical disks. Which value should I shoot for?

The Performance Monitor Avg. Disk Queue Length counter for the PhysicalDisk object tells you how many disk I/O requests, on average, are pending for your physical disks. An Avg. Disk Queue Length value of 2 means that at any given time two pending I/O requests are waiting for the disk. The higher this number is, the busier your disks are and the slower your Exchange Server performance is likely to be. There are plenty of old documents floating around that say this value should remain less than 1, but achieving that number in a multidisk system isn't likely. In reality, the value should be less than the number of physical disk spindles in the volume. For example, on a five-disk RAID-5 array, you'd want to see an Avg. Disk Queue Length value of less than 5, meaning that on average each disk had less than one I/O operation waiting for disk time. Sustained values that exceed your number of spindles indicate that your disk subsystem might be a bottleneck.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.