NFS datastore automounts as NFS3 instead of NFS4.1 when adding a host to an existing cluster using SDDC Manager

Products

• VMware vSphere ESXi
• VMware SDDC Manager

Issue/Introduction

Symptoms:

• An NFS datastore is mounted as NFS v4.1 on all existing hosts in the cluster.
• When a new host is added to the existing cluster through SDDC Manager, the datastore auto-mounts as NFS3 instead of NFS4.1.
• The datastore name changes after the mount.
  • Example: Original datastore name: Test_NFSDatastore changes to New datastore name: Test_NFSDatastore (1)

Environment

VMware vSphere ESXi 6.x
VMware vSphere ESXi 7.x
VMware vSphere ESXi 8.x

Cause

• The host receives a Create NFS Datastore command from a user account with the mount option set to NFS3.
• As a result, the host mounts the NFS datastore using NFS3.
• Since vCenter already has a datastore with the same name but a different UUID, vCenter renames the newly mounted datastore (e.g., appending (1) to the name).

Cause Validation:

• var/run/log/Hostd.log file confirm the create datastore request was initiated by the user "###-################-##############"
  YYYY-MM-DDTHH:MM.SSSZ In(166) Hostd[2099110]: [Originator@6876 sub=Hostsvc.DatastoreSystem opID=237aa658-de-6eb1 sid=526a2e4a user=vpxuser:VSPHERE.LOCAL\###-################-##############] Creating datastore DATASTORE_NAME
  
  
• var/run/log/vmkernel.log file show the NFS mount request was processed as NFS3 and completed successfully with different UUID.
  YYYY-MM-DDTHH:MM.SSSZ In(182) vmkernel: cpu18:2099110 opID=b683821c)NFS: 297: Command: (mount) Server: (#########.########.com) IP: (###.##.##.###) Connections: (1) Vmknic: (None) Path: (/#########_####_#####) Label: (DATASTORE_NAME) Optio$
YYYY-MM-DDTHH:MM.SSSZ In(182) vmkernel: cpu18:2099110 opID=b683821c)NFS: 340: Restored connection to the server #########.########.com mount point /#########_####_#####, mounted as ########-########-####-000000000011 ("DATASTORE_NAME")
YYYY-MM-DDTHH:MM.SSSZ In(182) vmkernel: cpu18:2099110 opID=b683821c)NFS: 317: NFS mount succeeded for #########.########.com:/#########_####_##### volume DATASTORE_NAME.
  
  
• var/run/log/vmkernel.log file confirms that vCenter renamed the datastore due to the UUID mismatch.
  2025-08-11T07:09:08.847Z In(182) vmkernel: cpu74:2099115 opID=11d2983e)World: 12324: VC opID [email protected]:323-3f3e0ab9-701f maps to vmkernel opID 11d2983e
YYYY-MM-DDTHH:MM.SSSZ In(182) vmkernel: cpu74:2099115 opID=11d2983e)NFS: 2027: Changing DATASTORE_NAME to DATASTORE_NAME (1)
  
  
• On the working host, var/run/log/vmkernel.log file confirms that the same datastore was mounted with different UUID.
  YYYY-MM-DDTHH:MM.SSSZ In(182) vmkernel: cpu56:2098533)NFS41: NFS41FSAPDNotify:6391: Restored connection to the server #########.########.com mount point DATASTORE_NAME, mounted as ########-########-####-000000000022 ("/#########_####_#####
  
  UUIDs can also be compared by running the command "esxcli storage filesystem list" on both the hosts. 

Resolution

Open a case with the Broadcom Support team to investigate the source of the incorrect NFS mount requests originating from the user.

Workaround:

• Unmount the duplicated datastores mounted as NFS3.
• Mount the original datastores from vCenter.