Artemis cluster config vs. standalone broker config

Question

I'm trying to set up a HA cluster of artemis brokers locally on my machine and want to demonstrate load balancing and failover behavior in a simple example. (before that I tried also a standalone broker). For my cluster I am using udp broadcast and discovery. I set up two brokers for the beginning.

In all those examples for cluster, there was always that auto-created list of acceptors (artemis, stomp,...), where we could add wildcard expressions or prefixes.

Why we don't need them in a cluster anymore? And what does netty-acceptor exactly stand for?

Answer 1

Whether or you not you need a particular acceptor configured in your broker.xml is really up to your use-case. Many of the examples use the default broker.xml which has an acceptor configured for every protocol which the broker supports. However, some examples (e.g. many of the clustered ones) only have the exact acceptors which they need to demonstrate the example's functionality.

I searched through all the configuration files of all the examples and the text netty-acceptor is only ever used as the name for an acceptor. The name of an acceptor simply identifies it uniquely among all the acceptors which are configured.

To be clear, any broker, whether it is clustered or not, can accept connections from any supported protocol granted the proper acceptor is configured.