SDDC MANAGER: vCenter RDU Post-Upgrade Task fails to Delete Original VCSA VM

Products

VMware vCenter Server

Issue/Introduction

At the end of a vCenter Server upgrade (using Reduced Downtime Upgrade [RDU] or a major upgrade), SDDC Manager reports that the upgrade finished with warnings
The SDDC Manager UI reports the following message:

Failed to delete VM "vcsa_vm_name" Default Message: "After the upgrade, manually change the VM names of the source and the new vCenter to the desired values"

The sddc manager's /var/log/vmware/vcf/lcm/lcm-debug.log reports the following:

Failed to delete VM 'vcsa_vm_name'.), resolution=LocalizableMessage(id=vcenter.deployment.migration_upgrade.failed_to_delete_vm.resolution, defaultMessage=Manually delete VM 'vcsa_vm_name'., args=[], params={0=LocalizationParam(s=vcsa_vm_name, dt=null, i=null, d=null, l=null, format=null)}, localized=Manually delete VM 'vcsa_vm_name'.))], errors=null), upgradeInfo=UpgradeInfo(startSwitchover=null, endSwitchover=null, error=null, scheduledDowntime=null, remainingReplicationData=0, pausedAt=null), lastUpdateTime=YYYY-MM-DDTHH:MM:SS, description=LocalizableMessage(id=com.vmware.vcenter.deployment.migration_upgrade.desc, defaultMessage=Migration Upgrade of vCenter Server appliance., args=[], params=null, localized=Migration Upgrade of vCenter Server appliance.), service=, operation=MIGRATION_UPGRADE, parent=null, status=SUCCEEDED, cancelable=true, error=null, startTime=YYYY-MM-DDTHH:MM:SS, endTime=YYYY-MM-DDTHH:MM:SS, user=null),[content-type:"application/json", date:"Thu, 15 Jan 2026 03:58:09 GMT", x-envoy-upstream-service-time:"67", vary:"Accept-Encoding", transfer-encoding:"chunked"]

Reviewing the vCenter Server logs /var/log/vmware/vpxd/vpxd.log reveals that the ESXi host containing the vCenter VM stopped responding when Fault Domain Manager (FDM) reconfiguration occurred:

YYYY-MM-DDTHH:MM:SS info vpxd[49847] [Originator@6876 sub=Vmomi opID=FdmMonitor-domain-c8-4ca3ec19] Creating SOAP stub adapter for /fdm on HOST_FQDN:443
YYYY-MM-DDTHH:MM:SS info vpxd[49049] [Originator@6876 sub=DAS opID=dasConfig.cpp:407-40c9b065] Triggering Das host config LRO to install and configure fdm on host [vim.HostSystem:host-###,HOST_FQDN]
YYYY-MM-DDTHH:MM:SS error vpxd[49792] [Originator@6876 sub=DAS opID=dasConfig.cpp:407-40c9b065-28c983cd] [VpxdDas::StartFdmService] Failed to start FDM service on host HOST_FQDN:N5Vmomi5Fault16HostNotConnected9ExceptionE(Fault cause: vmodl.fault.HostNotConnected
YYYY-MM-DDTHH:MM:SS info vpxd[49847] [Originator@6876 sub=Vmomi opID=FdmMonitor-domain-c8-4ca3ec19] Creating SOAP stub adapter for /fdm on HOST_FQDN:443
YYYY-MM-DDTHH:MM:SS warning vpxd[49847] [Originator@6876 sub=Vmomi opID=FdmMonitor-domain-c8-4ca3ec19] [FdmClientAdapter] Login request throttled on host HOST_FQDN
YYYY-MM-DDTHH:MM:SS error vpxd[49847] [Originator@6876 sub=Vmomi opID=FdmMonitor-domain-c8-4ca3ec19] Cannot contact the specified host (HOST_FQDN)
[vim.ClusterComputeResource:domain-c8,CLUSTER_NAME])

Historical analysis of vpxd.log may show intermittent Null response from host errors for the affected ESXi host prior to the upgrade.

YYYY-MM-DDTHH:MM:SS error vpxd[07250] [Originator@6876 sub=MoHost opID=db2c172] Null response from host [vim.HostSystem: host-###,HOST_FQDN]; no error

Environment

VCF 5.x

vCenter 8.x

Cause

ESXi host management connectivity failure during FDM reconfiguration prevents vCenter from executing post-upgrade inventory tasks.

During an RDU or major vCenter upgrade, SDDC Manager deploys a new vCenter VM with a temporary IP address. Upon successful data migration, the automated workflow powers off the original vCenter VM, deletes it, and renames the new vCenter VM to match the original inventory name. If the ESXi host residing the vCenter VMs experiences management connectivity issues, the host drops into a "Not Responding" state, preventing the automated cleanup tasks.

Resolution

To resolve the issue.

Restore connectivity to the affected ESXi host. Refer, the kb 344682 to identify and fix the host not responding issue.
Log in to the vSphere Client using an administrator account.
Locate the newly deployed vCenter Server VM in the vSphere inventory (this VM may currently have a temporary name associated with the upgrade process).
Locate the original (old) vCenter Server VM. Verify that its power state is Powered Off.
Right-click the original (old) vCenter Server VM and select Rename. Append a suffix such as -OLD to its name to free up the original inventory name.
Right-click the newly deployed vCenter Server VM and select Rename. Change its name to exactly match the original vCenter Server VM's inventory name.
Right-click the original (old) vCenter Server VM (-OLD) and select Delete from Disk to permanently remove it and free up underlying datastore capacity.