This issue is specific to the SoftLayer data center fabric. If ARP is sent for an IP address, a MAC will be returned back (from the fabric) even though the IP is not in use. This appears to have been the behavior with Softlayer fabric for a long time and is not easily changeable.
In NSX 4.x, NSBU added a check to ensure the LB VIP is not already used. However, in the SoftLayer environment, it sees the IP has an associated MAC and thinks it is in use. NSBU has provided a 'force=true' parameter which can be added to the API call to force setting the IP on the LB.
Workaround:
#curl -k -u 'admin:<nsx-manager-password>' -X POST "https://<nsx-manager-ip-or-fqdn>/api/v1/cluster/api-virtual-ip?action=set_virtual_ip&ip_address=<virtual-ip-address>&force=true"
curl -X GET -u'admin:######' https://<cloud-builder-ip-address>/v1/sddcs -k |json_pp > bringup.jsonFrom the above json, get the id field
ssh admin@<cloud-builder-ip>
#su
# pg_dump -U postgres -h localhost bringup > bringup-db.dump
#psql -U postgres -h localhost
#create table processing_task_bak as select * from processing_task; #create table processing_context_bak as select * from processing_context; #update processing_task set status='POSTVALIDATION_COMPLETED_WITH_SUCCESS' where task_name='SetVipForNsxtManagersAction' and execution_id='87bee553-e332-4282-#####-############'; #update processing_context set status='PENDING' where processing_state_name='ConfigureNsxtSecurityCertificates' and execution_id='87bee553-e332-4282-#####-############'; #update processing_task set status='COMPLETED_WITH_FAILURE' where task_name='IssueVMCACertsForNsxtManagerAction' and execution_id='87bee553-e332-4282-#####-############';Note: In the above commands, the execution Id '87bee553-e332-4282-#####-#############' is an example. Please get the Execution Id from Step-2
NSXT Manager deployment gets stuck.