Superflous error log when ending join session on the initiator

When ending the join session/discovery during a MicroCloud deployment, the initiator looks to be logging the following error message which seems superflous and is most likely caused by some component closing the multicast responder beforehand:

`2025-12-16T11:56:18Z microcloud.daemon[170444]: time="2025-12-16T11:56:18Z" level=error msg="Failed to close network endpoint after context got cancelled" err="close udp4 0.0.0.0:9444: use of closed network connection"`

The discovery already has a ` (d *Discovery) StopResponder()` which is invoked as part of stopping the multicast responder so likely we are just trying to close it a second time which then causes this error as the responder is already closed.

It should be tested if we require the additional `Close()` when the discovery responder's context is cancelled or if we can just skip it or just perform the `Close()` if it is still running.