Live Optics | Optical Prime | Queue Depth Values

Queue depth is an important metric to understand, along with other metrics, to help identify potential bottlenecks and maximize system performance. Queue depth is the number of outstanding I/O operations waiting to be processed at the disk drive from the Disk Controller. However, queue depth by itself is not a measure of good or bad.

Queue depth is often misunderstood in Optical Prime. It is frequently pointed out to suggest a disk limitation too quickly. To understand queue depth in an Optical Prime project, you need to know the number of disks that make up a particular drive. Unfortunately, Optical Prime does not provide this information.

Optical Prime monitors the number of outstanding I/O operations from the host OS's perspective. Host-level queue depth is reported differently than at the storage array level, and the two may have different values.

To extract value from the queue depth in Optical Prime, you need a few guidelines, the corresponding latency values, and, optionally, knowledge of the system itself.

Obtaining knowledge of the system might not be an option, but this guideline helps get you started:

• Host-based queue depth should be kept between 2 to 4 outstanding I/Os per drive that makes up the drive or LUN. For example, in a RAID set of five drives, a disk queue depth of 10-20 or below in an Optical Prime project would be in a normal range.

However, this is a tedious and largely unnecessary task when trying to get general information. Using the Viewer to show all graphs at once will place the Disk Queue and Latency graphs on top of each other. Each graph represents the same time recording, making it easy to look for corresponding values.

• Low Queue Depth and High Latency: Potential network or server bottleneck
• Low Queue Depth and Low Latency: Under-used or properly designed disk resources
• High Queue Depth and Low Latency: No particular problems
• High Queue Depth and High Latency: Potential disk bottleneck

The last entry is where you can start to look at the disk as the problem. If there are mirroring spike values between latency and queue depth, then inspecting the disk would be a logical activity.

 

If you have any questions, please reach out to Live Optics Support at liveoptics.support@dell.com.