Attempted to remove an NSX Edge Cluster from a VCF environment manually via vCenter and NSX Manager instead of using the mandatory SDDC Manager automation or removal scripts. This resulted in a stale configuration state where SDDC Manager retained database entries for an Edge Cluster that no longer existed in the inventory, preventing proper lifecycle management and causing a mismatch in workload domain associations.
VCF 5.x
VCF 9.x
By bypassing the automated removal process or the official edge_cluster_removal.py script Broadcom KB 316110, the SDDC Manager "inventory" became out of sync with the actual "infrastructure" state.
A case should be raised with Broadcom support so that the resolution can be worked through.