Federation standby global manager not syncing with active global manager

Products

• VMware NSX

Issue/Introduction

In a VMware NSX Federation environment, the Standby Global Manager reports a synchronization status of "not started" when viewed from the Active Global Manager UI. Standard replication prerequisites, including bidirectional communication on port 1236, active async_replicator services, jumbo frames MTU configuration, valid certificates, and matching API thumbprints, are confirmed to be operational.

Example screenshot:

Cause

The synchronization failure is caused by an epoch mismatch within the Corfu DB. The epoch counter in the Standby Global Manager exceeds the epoch number of the updates being sent by the Active Global Manager (e.g., Standby epoch is at 3, while the Active manager is sending updates with epoch 1 and 2). Consequently, the Standby Global Manager rejects the synchronization updates because it evaluates its local database state as newer and more up-to-date than the incoming replication data.

Resolution

Verify that all standard replication prerequisites are met (port 1236 bi-directional traffic, jumbo frames MTU size, certificate validity, and async_replicator service status on both managers).

If all of the above have been confirmed then please collect a NSX Manager log bundle and open a Support Request with Broadcom.