VCF 9.0
When TransportNodes were deleted, and the second ComputeManager was deleted, the TransportNode uninstall code failed to remove the VPC's/Subnet's realized vCenter objects (NSX portGroups/folders) from both NSX and vCenter, for the deleted vCenter, due to vCenter connectivity being down. Such stale entries in the system may delay realization of other VPC/Subnets, onto vCenter, due to exception on retrying stale entry.
If you believe you have encountered this issue, open a support case with Broadcom Support and refer to this KB article.
For more information, see Creating and managing Broadcom support cases.