NSX-T 3.1.1 to 3.1.3.1: Unable to make a Global Manager as standby from the UI after an unplanned switchover where this Global Manager was lost but is back online after the switchover.

Products

• VMware NSX Networking

Issue/Introduction

To ensure successful and reliable unplanned switchover process for NSX-T versions 3.1.1 to 3.1.3.1

Symptoms:

Federation recovered global manager is not set to standby, when the switch state via API call is executed, standby mode changes back to none.

Environment

VMware NSX-T Data Center

Cause

Site manager fails to mark the re-added GM node as standby

Resolution

Issue is resolved in NSX-T 3.2.0 and above.

Workaround:

1. User simulates (or runs into) unplanned failover scenario by powering off the current Active Global Managers (GM)

2. User logs into current Standby GM and uses the UI: System -> Location Manager -> Standby GM Card -> Make Active option to make it active.

3. Wait for the UI to prompt the "Force Active" option and use that option to confirm unplanned failover from Standby GM to Active GM

4. Post completion of failover, user brings back up the old active GMs (power on in this case) and waits for the cluster status to be STABLE on the old active GM

5. User offboards the old active GM which are currently in mode "NONE" after the power on from the current active GM UI: System -> Location Manager -> Old Active GM card -> Remove option and wait for the GM to be removed

6. User re-onboards the old active GM as new Standby GM from the current active GM UI: System -> Location Manager -> Add Standby Option

 

Post step 6, the unplanned failover is complete, and user will have a stable Active/Standby GM federation setup.

Additional Information

Impact/Risks:

Using UI or API call, user will not be able to make old active Global manager as standby after unplanned switchover.