This can be controlled by the premptive / non-preemptive failover mode on the gateway
| Preemptive | If the preferred node fails and recovers, it will preempt its peer and become the active node. The peer will change its state to standby. |
| Non-preemptive | If the preferred node fails and recovers, it will check if its peer is the active node. If so, the preferred node will not preempt its peer and will be the standby node. |
https://docs.vmware.com/en/VMware-NSX-T-Data-Center/3.2/administration/GUID-7B0CD287-C5EB-493C-A57F-...
