NetherNet

[elyy442]

As of this writing, i've updated to the latest build and i still receive "Search help,Minecraft.net for "NetherNet" for more information. The plugin has been working flawless for me until last week. I've been getting this NetherNet issue and some times in a day, players can join but most of the time, they receive NetherNet

[dromzeh]

Same here. Using latest (+ latest Geyser/Floodgate/Paper 26.2) where console players get the NetherNet screen and can't join.

Sometimes it logs SessionCreationException: Unable to start NetherNet channel right after "Connection to websocket lost", followed by "Session is dead and hit exception trying to re-create it".

My first instinct was honestly to just "listen into logs for keywords and then auto-restart on fail" - there's times this error happens and nothing is logged at all. Also checked the session document while it was broken with nothing logged & it looked completely normal.

As mcxboxbroadcast restart does fix it instantly, you could maybe configure something another to rerun it every hour via rcon or something or so for the time being.

[l33tlinuxh4x0r]

I too have seen this error and lots of the following warnings... WARN Geyser-Spigot] [mcxboxbroadcast] [Primary Session] Connection to websocket lost, re-creating session...

[snowboyz0825]

This is probably due to this change in the changelog

When connecting to a dedicated server, the client probes the server for NetherNet support with a 3 second timeout. WinHTTP does not honor this timeout in the following situation:
The client connects to the server by URL instead of IP.
The server hostname's DNS record contains multiple IP addresses.
The server is behind a firewall that drops the client's probe TCP packets instead of responding with an RST.
Instead of a 3 second timeout for the entire connection attempt, WinHTTP enforces a (3 seconds) * (number of IP addresses in DNS record) timeout.
This bug has been fixed by enforcing an overall 3 second timeout.

[rtm516]

This is probably due to this change in the changelog

When connecting to a dedicated server, the client probes the server for NetherNet support with a 3 second timeout. WinHTTP does not honor this timeout in the following situation:
The client connects to the server by URL instead of IP.
The server hostname's DNS record contains multiple IP addresses.
The server is behind a firewall that drops the client's probe TCP packets instead of responding with an RST.
Instead of a 3 second timeout for the entire connection attempt, WinHTTP enforces a (3 seconds) * (number of IP addresses in DNS record) timeout.
This bug has been fixed by enforcing an overall 3 second timeout.

That's for http signalling which isn't what this project uses

[rtm516]

OP could you provide logs, config and a screenshot of the error message along with the client version (bottom right of the main menu)

[snowboyz0825]

I restarted the bot and it fixed the error. It didn't even log that people were trying to connect so idk what it was, or if it was a different error entirely