VCF Installer pre-validation fails at "NFS Datastore Configuration" in VMware Cloud Foundation 9.1

Products

VMware vSphere ESXi

Issue/Introduction

Introduction:

VMware Cloud Foundation (VCF) 9.1 deployments may fail during the Cloud Builder pre-validation phase at the NFS Datastore Configuration check. This failure prevents the automated bring-up process from proceeding and typically indicates a mismatch between the VCF deployment specification and the underlying network or storage infrastructure. Key indicators include MTU discrepancies on vSwitch0, VMkernel port group nomenclature errors, and NFS mount rejections due to Access Control List (ACL) restrictions on the storage array. This article provides the necessary steps to identify these configuration gaps in the domainmanager.log and vmkernel.log files and details the remediation required to standardize the environment for a successful VCF 9.1 management domain instantiation.

Symptoms:

During a VMware Cloud Foundation (VCF) 9.1 deployment, the Cloud Builder pre-validation fails at the "NFS Datastore Configuration" phase.

Environment

VCF 9.1

Cause

VCF automated bring-up requires strict adherence to network and storage prerequisites. This validation failure typically occurs due to common infrastructure misconfigurations:

Network Mismatch: The VMkernel port group name for NFS traffic on the ESXi hosts does not exactly match the name specified in the VCF deployment parameters. VCF deployment spec:
"portGroupKey": "example-dvs-pg-nfs", <- The port group name should be the same on primary host vswitch.
"standbyUplinks": [],
"teamingPolicy": "loadbalance_loadbased"
MTU Fragmentation: Jumbo Frames (MTU 9000) are specified in the VCF workbook, but the physical network path (ToR switches) or logical path (vSwitch/pNIC) is still set to the default MTU of 1500, causing packet drops.

/var/log/vmware/vcf/domainmanager/domainmanager.log

MTU provided in NFS network spec is higher than the MTU configured on switch vSwitch0. During the installation of switch vSwitch0 MTU 1500 will be updated to NFS MTU value 9000 provided in input specification

/var/run/logs/vmkernel.log

NFS mount failed for ####.####.####.####:/<nfs_share> volume <nef_volume>. Status: Unable to connect to NFS server.
Storage Security: The designated IP range for the VCF NFS VMkernel adapters has not been added to the Access Control List (ACL) or export policy on the remote NFS storage array, causing the array to deny the connection.

/var/run/logs/vmkernel.log
NFS mount failed for ####.####.####.####:/<nfs_share> volume <nef_volume>. Status: The mount request was denied by the NFS server. Check that the export ex$

Note: While the causes outlined above were observed in the environment, the issue is not exclusively limited to them.

Resolution

To resolve this issue and successfully pass pre-validation:

Verify Port Group Naming: Ensure the NFS VMkernel port group configured on the primary ESXi hosts exactly matches the spelling and casing defined in VCF deployment spec.
Configure End-to-End MTU: Work with your network administrator to ensure MTU 9000 is enabled end-to-end. This includes the physical switch ports, ESXi physical NICs, vSwitches, and VMkernel adapters.

Test the MTU via the ESXi shell using:

vmkping -I <nfs-vmk-name> -s 8972 -d <NFS-Server-IP>.
Update Storage ACLs: Work with storage administrator to ensure the entire subnet or IP range allocated for VCF NFS network is explicitly permitted to access and mount the target NFS export path on the storage array
Once the infrastructure is corrected, re-run the VCF pre-validation workflow.

Additional Information

VCF 9.1 Planning and Preparation Guide