waiting infinity for sessions to close was causing the swarm_server to
never fully die when it was terminated.
Why it's own 10s timeout in libp2p_swarm_sup did not come along and
brutally kill the process is still a mystery to me.
The supervisor is a one_for_all, if anything is going down, it's taking all
the sessions with it already. We'll still log to know what is killing
the server.
waiting infinity for sessions to close was causing the swarm_server to never fully die when it was terminated. Why it's own 10s timeout in libp2p_swarm_sup did not come along and brutally kill the process is still a mystery to me.
The supervisor is a one_for_all, if anything is going down, it's taking all the sessions with it already. We'll still log to know what is killing the server.