By default, transport nodes access NSX Manager based on their IP addresses. However, this can also be based on the DNS names of the NSX Manager, providing the ability to change IP addresses for disaster recovery.
After upgrading from NSX-T 3.2 to NSX 4.x, the following symptoms may be observed:
GET https://<nsx-manager>/api/v1/configs/managementshows"publish_fqdns": true
VMware NSX
During the database migration phase of the NSX upgrade, the internal ManagementConfigModel database table fails to properly migrate its existing configurations to the newer schema version.
Because the custom configuration is dropped during migration, the system reverts the publish_fqdns parameter back to its default system state.
There is no pre-upgrade fix to prevent this behavior. You must manually set FQDN publishing to false via the NSX REST API immediately after the upgrade is complete.
You must fetch the current configuration to obtain the active _revision number. Use an API client (like Postman or curl) with NSX Admin credentials.
Request:
GET https://<nsx-manager-ip-or-vip>/api/v1/configs/managementResponse payload snippet:
{ "publish_fqdns": true, "_revision": 12}Note the integer value of _revision. You will need it for the next step.
Send a PUT request to the same endpoint. You must include the _revision number gathered in Step 1 and set publish_fqdns to false (or the desired state).
Request:
PUT https://<nsx-manager-ip-or-vip>/api/v1/configs/managementBody:
{ "publish_fqdns": false, "_revision": 12}Allow a few minutes for the NSX Management plane to push the updated configuration down to the transport nodes.