What happens if a non-leader server is contacted?

When a non-leader server in a HashiCorp Vault cluster is contacted, the request is transparently forwarded to the active leader. This mechanism is part of how Vault maintains a highly available and resilient architecture.

In a High-Availability (HA) setup, Vault uses a leader-follower model where only the leader can process certain critical operations like writes or specific administrative tasks. If a client interacts with a non-leader server, the system recognizes that the request cannot be processed there, and instead of dropping the request or generating an error, it automatically forwards it to the current leader. This process ensures that clients can still perform actions without being aware of the cluster's internal state, thereby maintaining a seamless user experience.

In contrast, other options do not accurately reflect the behavior of Vault's architecture. Ignoring a request or generating an error would complicate interactions for clients unnecessarily. Disconnecting a client would also be counterproductive, as it would disrupt the operation instead of facilitating successful communication. By leveraging the forwarding mechanism, Vault allows for smooth operations and effective load balancing among its servers.