vCenter performance degradation due to excessive Storage DRS migrations for thin-provisioned disks

Products

• VMware vCenter Server

Issue/Introduction

This article addresses performance issues in vCenter Server caused by a high frequency of Storage DRS (SDRS) migrations. This typically occurs when datastores contain a large number of thin-provisioned disks.

Symptoms:

• vCenter Server UI response is slow or delayed.
• Continuous Storage vMotion tasks are observed, even when datastores space utilization is below the configured threshold.
• SDRS logs indicate datastores are "overloaded" due to "Entitled Space" calculations.

Environment

VMware vCenter Server 8.x

Cause

• SDRS calculates space demand for thin-provisioned disks by adding a growth buffer to the current usage. This is controlled by the advanced parameter PercentIdleMBInSpaceDemand, which defaults to 25%. 
• SDRS takes 25% of the "uncommitted" (unused) capacity of every thin disk and adds it to the "used" space to determine the Total Entitled Space.
• In environments with large over-provisioned disks, this calculation frequently pushes the "Entitled Space" causing SDRS to flag the datastore as overloaded and trigger continuous migrations.

Resolution

To reduce the frequency of migrations and restore vCenter performance, adjust the growth buffer calculation:

1. Select the Datastore Cluster > Configure > Storage DRS > Edit.
2. Select Advanced Options.
3. Add or Edit the parameter: PercentIdleMBInSpaceDemand.
4. Lower the value from 25 to 10 or 5.
5. Click OK.

Note: Lowering this value reduces the projected growth buffer, which prevents SDRS from prematurely flagging datastores as full while still allowing it to manage actual physical space constraints.

Additional Information

Storage DRS FAQ