Client original IP address and docker-proxy

[hroost]

Hello,

I found this strange behavior with the client's original IP address and docker-proxy.
Using default docker-compose example with

services:
  teamspeak:
    image: teamspeaksystems/teamspeak_server:58rc12
    expose:
      - "9987:9987/udp"
    ports:
      - "9987:9987/udp"

after a successful connection to the server in server logs and client's Connection Info I'm seeing my original IP.
Then let's do a server restart via docker-compose restart teamspeak and wait till TS client will perform auto reconnecting. Then check IP again - now it's internal docker 172.18.0.1 instead of the real one. Restart TS client and it's again real IP.
Btw, sometimes docker-compose restart goes smoothly (faster?) and I'm seeing real IP. In this case, trying something like that docker-compose stop teamspeak && sleep 10 && docker-compose start teamspeak helps to reproduce

TS client 5.0.0-beta71 macOS

I would appreciate any suggestions how to workaround this (ideally without network_mode: host)

[qknight]

ATM we don't have a solution to this problem. Once we go from beta to release we probably will address this issue. Thanks for reporting!

[qknight]

@hroost please have a look at #19 (comment)

Maybe by directly binding it it will not fallback to the internal address. If this helps you already, I'd love to learn so and add it to the documentation.