Manual Failover on replicated clusters

[Chaunceyan]

Dear Community,

We have clusters A and B. They are replications of each other, subscriptions include.

At first our clients were connecting to cluster A through a hostname, say "pulsar.local". We then changed the DNS pointing pulsar.local to cluster B, nothing happens.

The clients are still connecting to cluster A broker in direct IP connection. We have to manually reboot the clients for the DNS change to take effect.

I wonder if there is better option, like server side operation, to trigger the clients to create new connection.
Will unloading namespace do the trick?