During an NSX upgrade, the upgrade pre-check task fails, preventing the upgrade process from proceeding. The pre-check logs or NSX UI display errors related to ESX hosts, such as "Manager connectivity check" failures, "unmanaged host" alarms, or "MPA connectivity down" alerts.
FAILURE for specific ESXi hosts.The upgrade pre-check process attempts to validate all transport nodes (TNs) associated with the NSX fabric. If an ESXi host has been decommissioned or removed from the vCenter environment without first being explicitly removed from the NSX Fabric, the NSX Manager retains stale configuration for that node. The pre-check validation fails because the Manager cannot establish connectivity to the removed or unreachable host.
To resolve these pre-check errors and proceed with the upgrade, remove the stale NSX configuration for the decommissioned hosts:
Identify the Stale Host(s):
FAILURE messages.Remove from NSX Fabric:
Verify and Retry:
If you are contacting Broadcom support about this issue, please provide the following: Reference KB "Creating and managing Broadcom support cases"
Handling Log Bundles for offline review with Broadcom support