Datastore Expansion Failure on Multi-Extent VMFS-6 Datastores

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms : 

• Datastore expansion fails on multi-extent VMFS-6 datastores with the error: "Operation failed! Cannot change the host configuration."
  
  
  
  
• This issue occurs even after the underlying LUN space is increased on the storage array. The datastore appears to be nearly full (e.g., 99% used).
  
  
• The affected datastores (e.g., DATASTORE_A) is built on multiple LUN extents.

Environment

VMware ESXi 8.0 Update 3

Cause

The issue arises from an incorrect expansion procedure applied to a multi-extent VMFS datastore:

• Instead of extending the tail LUN, additional space (e.g., +2 TB) was added to the head LUN.
  
  
• Because VMFS concatenates extents, unallocated space on the head extent cannot be utilized.

Example 1: The following datastore expansion will not work.

[root@ESXHOST:~] vmkfstools -Pv 10 /vmfs/volumes/DATASTORE_A
VMFS-6.xx (Raw Major Version: xx) file system spanning 2 partitions.
...
Partitions spanned (on "lvm"):
        <naa_ID_for_Extent_1>:1 >> Head Lun expanded with 2 TB space 
        <naa_ID_for_Extent_2>:1 >> Tail  Lun

Example 2: The following datastore expansion will work.

[root@ESXHOST:~] vmkfstools -Pv 10 /vmfs/volumes/DATASTORE_A
VMFS-6.xx (Raw Major Version: xx) file system spanning 2 partitions.
...
Partitions spanned (on "lvm"):
        <naa_ID_for_Extent_1>:1 >> Head Lun
        <naa_ID_for_Extent_2>:1 >> Tail  Lun expanded with 2 TB space 

Resolution

General Recommendation : 

1. 1. Always expand tail LUNs in multi-extent datastores.

   2. Monitor datastore utilization proactively to avoid critical expansion failures.

Workaround : 

Stage 1: Create New Datastores and Migrate Data (Storage vMotion)

1. Request New Storage: Request the storage team provision two new, single LUNs of sufficient size (e.g., ensure the new capacity exceeds current usage plus projected growth).

• • • • Recommendation: Ensure these new LUNs are single-extent to simplify future expansion (which can be done by simply growing the single LUN).

The new datastores must be appropriately sized. Based on the original utilization data:

Datastore Name ( Source Datastore)	Current Used Capacity
DATASTORE_A	10.7T

 

Datastore Name (Destination Datastore for Storage vMotion)	New Capacity
DATASTORE_A_New	> 10.7T (e.g. 20T)

• • • • Present and Format: Present the new LUNs to the ESXi cluster. Format them as VMFS-6 datastores (eg., DATASTORE_A_New).

2. Perform Storage vMotion:

• • • • Initiate a Storage vMotion migration to move all Virtual Machines from the old datastores (DATASTORE_A) onto the corresponding new datastore (eg., DATASTORE_A_New).

3. Verification: After migration, closely monitor the new datastores and verify that all VMs are running correctly and all necessary data has been successfully moved.

Stage 2: Decommission Old Datastores

1. 1. Unmount: After successful migration and a suitable monitoring period, unmount the old datastores (DATASTORE_A) from all ESXi hosts in the cluster.

   2. Delete: Delete the VMFS datastore.

   3. Reclaim Storage space : The storage team can then reclaim the old LUNs from the ESXi hosts and proceed with reallocating or reclaiming the underlying storage space.