As this is a greenfield VCF 9.1.1 deployment and the failure always takes place when creating the NSX Transport Node collection, first look at the actual failure that occurs during NSX host preparation and transport node setup rather than assuming that the problem is with DNS or NTP.
Broadcom has recorded instances where VCF 9.1 deployments fail at this particular stage since NSX is unable to prepare the ESXi hosts due to the use of an unsupported or incompatible ESXi OS version; there are also similar cases with VCF 9.x caused by unsupported hardware or third-party VIBs or by incompatibility of the vLCM image.
Before rebuilding again, check:
- NSX Manager → System → Fabric → Hosts/Transport Nodes and identify which host shows Install Failed, Orphaned, or another preparation error.
- In vCenter → Cluster → Updates, check whether the desired vLCM image reports incompatible components, drivers, VIBs, CPU, or PCI devices.
- Verify that the exact ESXi build, server hardware/firmware, NICs and drivers are supported for the VCF/NSX 9.1.1 combination.
- Check /var/log/vmware/vcf/domainmanager/domainmanager.log on the VCF Installer/SDDC Manager around the failed task. The UI's null message may be hiding the actual NSX realization error.
It's also a good idea to check the NSX Manager logs before resetting the environment; if all hosts become inaccessible right after this step fails, it may indicate that host preparation or networking changes are being partially applied before the workflow is aborted.
Don't proceed with clean deployments until you identify the underlying error in the transport-node implementation, because the failure occurring at the same stage points to something deterministic in the host/image/network setup rather than a random deployment issue.
Broadcom KB 443705 is particularly relevant here, as shown by: "VCF 9.1 Deployment Fails at 'Create NSX Transport Node Collection' Due to Unsupported OS Version."