Stretched Gateways Do Not Automatically Failover Between Locations in NSX Federation
Article ID: 441764
Updated On:
Products
Issue/Introduction
In an NSX Federation Active-Standby Stretched Tier-0 and Tier-1 architecture, a complete Edge cluster failure at the Primary location results in a data plane outage.
The active forwarding state does not automatically shift to the Secondary location, and North-South traffic is black-holed.
Environment
3.x
Cause
This behavior is strictly by design. NSX Federation intentionally suppresses automatic inter-site location promotion for Stretched Gateways to prevent split-brain routing states and control plane partitions during an outage.
Resolution
Do not architect disaster recovery plans assuming automatic inter-site failovers for Stretched Gateways in NSX Federation.
During a Primary location failure, log into the NSX Global Manager UI.
Manually execute the Network Recovery workflow to change the Stretched Gateway location configuration from Secondary to Primary.
To prevent node-level failures from causing site-level outages, ensure the local Primary Edge Cluster contains N+1 redundancy. This allows automated local High Availability (HA) to function without requiring a manual cross-site failover.
Additional Information
Feedback
