Automatic Space reclamation fails on VMFS-6 datastore due to unsupported VAAI Delete status.

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• Storage space does not reclaim from the backend storage array after deleting data or migrating virtual machines from a VMFS datastore. This behavior is primarily observed on IBM FlashSystem 7300 arrays using ESXi 8.x/7.x.
  
  
• Running the verification command shows the Delete primitive as unsupported:

esxcli storage core device vaai status get -d <naa_id>

Output Example:

VAAI Plugin Name: ATS
Status: supported
Delete Status: unsupported

• Automatic Space Reclamation is enabled on the datastore. Refer: Space Reclamation on vSphere VMFS Datastores

Environment

• VMware ESXi 8.x
• VMware ESXi 7.x
• IBM FlashSystem 7300/9000 series
• VMFS-6 datastores

Cause

The SCSI UNMAP thin-provisioning block reclamation primitive is disabled within the device presentation policy or volume configuration on the physical storage array management console.

When the array signals an unsupported state, the ESXi host suppresses all reclamation tasks to prevent invalid SCSI commands.

Resolution

1. Enable SCSI UNMAP on the Array:
   
   
   • Log into the IBM FlashSystem management console and ensure that space reclamation (SCSI UNMAP) is enabled for the specific volumes/LUNs presented to the ESXi cluster.
     
     
2. Verify VAAI Status on Host:
   
   
   • Run the following command on the ESXi host to confirm the status has changed to supported:
     
     esxcli storage core device vaai status get -d <naa_id>

Output Example:
     
VAAI Plugin Name: ATS
Status: supported
Delete Status: supported


3. Confirm VMFS-6 Automatic Reclamation:
   
   
   • Verify the datastore reclamation priority is set to "Low" (default background) in the vSphere Client (UI) or via CLI.
     
     
     • UI Method:
       
       
       • Select Datastore > Configure > General > Space Reclamation > Edit.
         
         
     • CLI Method:
       
       
       • Run the following command to check the space reclamation parameters.
         
         ## Query using the datastore name.
         
         esxcli storage vmfs reclaim config get --volume-label=<DATASTORE_NAME>
         
         ## Query using the datastore UUID.
         
         esxcli storage vmfs reclaim config get --volume-uuid=<DATASTORE_UUID>
         
         
4. Manual Reclamation (Optional for VMFS-5):
   
   
   • If space must be reclaimed immediately or if using VMFS-5, run:
     
     esxcli storage vmfs unmap --volume-label=<datastore_name> --reclaim-unit=200

Additional Information

• To speak with a customer representative or a Support Engineer, see Contact Support.
• Subscribe to this article to receive updates: How to subscribe to Knowledge Management articles.