Remove an NSX-T Manager Node via the API

NSX
  • You deployed the 2nd and 3rd NSX-T Manager Nodes via CLI since the NSX-T computer manager added to the 1st node is NOT in the same vCenter.
    Ex. NSX-T Manager VMs are in a MGMT-VC use traditional VLANs only for the backed infrastructure, but the NSX-T added Compute Manager is connected to a Compute-VC for client consumption of NSX-T ESXi hosts.
  • You want to decommission 1 of the 3 NSX-T Manager nodes. Another reason could be the VM will not boot up and/or is corrupt.
  • Find the UUID of the NSX-T Manager in scope.   In my case:
    fee71742-3a42-5ad5-5b29-1c1da71ca7d0
    nsxt-3.my.lab
    192.168.0.253
  • Setup you favorite API Client. Insomnia https://portapps.io/app/insomnia-portable/ with Basic Auth for the NSX-T admin account.
  • API – GET https://nsxt-vip.my.lab/api/v1/cluster
  • SSH into a working NSX-T Manager and run
    get cluster status | last 4
  • Per the NSX-T API Manual
  • POST – /api/v1/cluster/<node-id>?action=remove_node
  • Edit for my lab:  https://nsxt-1.my.lab/api/v1/cluster/fee71742-3a42-5ad5-5b29-1c1da71ca7d0?action=remove_node
  • Observation – the output ends with node 02, not node 03
  • Wait apx 3 minutes as the change will stop the NSX-T webpage from loading. The output shows nodes 1+2.
  • The NSX-T GUI shows the 3rd node is gone
    nsxt-3.my.lab 192.168.0.253
  • get cluster status | last 4
    run the command on node 1 to note it sees itself and node 2.
    run the command on the decommed node 3 and it is not part of the cluster.