On UNIX, each Adaptive Server engine is a separate process; therefore, network listeners run slightly differently than they do on Windows, where Adaptive Server is a single process.
On UNIX:
Adaptive Server uses one listener task per port. Each listener task functions as multiple logical listeners by switching from engine to engine, attempting to balance the load. For example, a 64-engine Adaptive Server with two master ports has two listener tasks, but these two listener tasks act as 128 logical listener tasks, so the server has two physical and 128 logical listeners. Starting a listener on engine 3 does not result in Adaptive Server spawning a new listener task unless the port does not already have a listener
A listener task accepts connections on engines on which it is enabled. So a single listener task corresponds to many logical listeners.
Stopping a listener on a specific engine terminates the logical listener for this engine since the listener task no longer switches to that engine. Adaptive Server terminates the listener task in case this was the last engine on which it was allowed to operate.
