I/O controls for optimizing performance after the bottleneck analysis

Storage administrators must respond quickly to sudden changes in I/O traffic. Shared infrastructure resources can degrade in performance at unpredictable times. If the bottleneck analysis detects a spike in total IOPS, as shown in the following figure, the root cause is an insufficient amount of resources available.

Because adding resources cannot be performed quickly, or might not be possible, the most efficient solution is to manage the I/O traffic. For a storage administrator, this situation must be treated as an emergency. In the Set IO Control window, set an upper limit for the volumes overusing resources immediately when you detect them.

You might use the upper limit setting as a temporary measure to manage the resources as planned for daily operations. If critical resources require less I/O prominence, remove the upper limit setting. All I/O control tasks are saved to the History tab.

Continue checking the History tab to monitor the upper limit settings by searching for tasks.

Preventing noncritical resources from causing performance degradation

When you are notified of performance degradation through an alert, perform the bottleneck analysis to detect the disruptive resource:

• Review the trend charts through E2E or Sparkline View to compare performance of selected resources.
• Use the Analyze Shared Resources window to identify the noncritical resources that are disrupting I/O traffic.
• When the Resources using the bottleneck window opens, you see a list of volumes that correspond to the trend chart.
• Identify the target volumes issuing many I/O requests.
• Select the target volume and then click More Actions to select Set IO Control. For your reference, give the task an appropriate name in the description field.
• Continue monitoring the History tab. When I/O control is no longer needed, select the target volumes of the task in IO Control Settings and click Off, or modify those limits as needed and resubmit.