PMS not accessible over public IPv6 address

[Frostelikus]

I am trying to make my server available via a public IPv6 address from a subnet I got with a server running at Hetzner. The setup looks like this

┌────────────┐
│            │       ┌────────────┐             ┌─────────────┐
│            │       │            │  Wireguard  │             │
│  Internet  ├──────►│  Gateway   ├────────────►│ Plex Server │
│            │       │            │ public IPv6 │             │
│            │       └────────────┘             └─────────────┘
└────────────┘

In general, access to the system running PMS using the public IPv6 address is working just fine. I can ping it and login via ssh. However, PMS is only accessible to clients that are either in the same local network or connected via VPN. I have manually configured the custom server access URL to something like https://2a01-<…>-00bb.897<…>.plex.direct:32400/ and it seems to be correct as I can open it when I am in the VPN/local network. Running nc outside of the VPN/local network shows that the port is apparently open

nc -6 -z 2a01:<...>::bb 32400
Connection to 2a01:<...>::bb port 32400 [tcp/*] succeeded!

Also it looks like that the PMS is listing to all IPv6 addresses

sudo lsof -i -P -n | grep plex
Plex\x20M 600022            plex   10u  IPv6  968760      0t0  TCP *:32400 (LISTEN)

But still, opening the remote server URL in a browser results in a timeout. Does anyone has an idea what could be the cause?

Server Version#: 1.30.0.6486

[ChuckPa]

Plex doesn’t support V6.

I’ve asked and there are apparently issues in other areas which remain unresolved.

If you setup your own FQDN with certificate then add to PMS, that will work.

you can’t use Plex.tv

NOTE: Do not use your “plex.direct” URL. That is your certificate-specific address and will change the next time the certificate updates.

ALSO. Be advised. Most of the clients are ok with IPv6

[Frostelikus]

Thanks for the reply.

Actually it turned out not to be a plex problem, but rather the VPN connection was horribly unstable. In case someone ends up here via google with a similar problem: Turning on the debug log on the wireguard server showed that the connection broke down regularly. Here is some information on how the get the log: Four Ways to View WireGuard Logs | Pro Custodibus Essentially it is calling echo module wireguard +p | sudo tee /sys/kernel/debug/dynamic_debug/control and than check dmesg. I got the connection stable by changing the endpoint on the wireguard configuration to use the IPv6 address of the server.