OVF Provisioning Converts Non-Root Disks to Thin Provisioning on Individual Datastores
search cancel

OVF Provisioning Converts Non-Root Disks to Thin Provisioning on Individual Datastores

book

Article ID: 410724

calendar_today

Updated On:

Products

VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

When deploying multi-disk OVF images via Aria Automation, you may observe that all disks except the root disk are provisioned as thin provisioned, even when thick provisioning is desired.

This behavior occurs specifically when deploying to individual datastores, contradicting the expected provisioning type from the storage profile. This issue impacts business use cases where consistent thick provisioning is required for all virtual machine disks.

Environment

Aria Automation 8.x

vCenter Server 7.x

vCenter Server 8.x

Cause

The functionality to apply a consistent provisioning type (Thick/Thin) across all disks in a multi-disk OVF image is not implemented for individual datastores. This feature was initially implemented for the first disk only. While support for consistent provisioning across all disks was later added for datastore clusters, it was not extended to individual datastores. This limitation is consistent across vCenter versions 7.x and 8.x, and applies to both VMFS and vSAN datastores.

Resolution

The only viable option for consistent provisioning of all disks in a multi-disk OVF image through Aria Automation is to utilize Datastore Clusters.

To implement this resolution:

  1. Transition to Datastore Clusters: Ensure your environment has datastore clusters configured for storage.
  2. Apply Capability Tags: Apply appropriate capability tags to the datastore clusters to define their storage characteristics.
  3. Update Aria Automation Storage Profiles: Modify your storage profiles within Aria Automation to reference these capability tags, directing provisioning requests to the datastore clusters.
  4. Remove Tags from Individual Datastores: To ensure proper placement and avoid unintended behavior, remove any conflicting capability tags from individual datastores that might override the datastore cluster settings.