Problem

A cluster becomes unstable after two NSX Advanced Load Balancer controllers go down.

Cause

When a three node cluster loses two controllers, the cluster quorum is lost. It becomes unstable.

Solution

  1. Open a SSH session and log in to the controller that is up and running.
  2. At the terminal, run the /opt/avi/scripts/recover_cluster.py script.
  3. Verify the VIP of cluster is back up and the cluster status is stable.