waterfall-help

[craggy trout]

I don't think Waterfall fixes viaversion with bungeecord

I think it is because viaversion modifies the order of the packets compared to vanilla and that is why bungeecord is broken

[reef fulcrum]

ViaVersion was generally just a common example of a plugin in which screws with stuff, and it being stupidly popular

I really do not heave the health or too much of a care to deal with testing stupidly low level changes inside of waterfall

[craggy trout]

The PR that my friend made has been in testing for several months on a private proxy and never gave problems with many players. However, it only applies if the entity rewrite is not activated.

[reef fulcrum]

I need more than 1 server saying that it works for me to be willing to bypass my own desire to actually test stuff before its pulled

[craggy trout]

Ok

I've never had reports of that patch, but I'm going to test it with conflicting things to see what might happen.

[shy lagoon]

hey how can I make it so players are sent back to the last server they played on when they join?

• I have force_default_server set to false
• my modules.yml is attached (a random spigot thread from 6 years ago said it was relevant)

[craggy trout]

You need to download modules that you want in this case Reconnect and move it into a folder called "Modules"

[reef fulcrum]

the thing will automatically download the modules and slap them into the folder, unless you're using custom build jars

[craggy trout]

Yeah

[tidal musk]

Hi! Can someone help my add the domain to my waterfall proxy? The issue arises when I attempt to connect to my server using "sharkpix.pl" – it doesn't work as expected. Instead, it tries to connect me directly to the lobby server, bypassing the proxy.

[craggy trout]

Probably you set the wrong port

[tidal musk]

wdym? for Record SRV i set 25577

[shy lagoon]

I do have all the modules within the modules folder

[lost river]

is that the query-port or the port at the end of host?
Beacause you need to put the port of the second one into the SRV record

[tidal musk]

They are the same

But i actually did it nvm

[edgy crow]

Hey am I correct that the latest build of waterfall (548) still doesn’t support 1.20.2?

[craggy trout]

Latest build supports 1.20.2

[edgy crow]

Waterfall console:
https://pastebin.com/E84Y9D12

1.20.2 server console:
https://pastebin.com/drsi2YfK

I made sure to test my client without any mods. The main server I connect to is paper 1.20.1, but this is what happens when I try to switch to the testing server on 1.20.2.

This happens from a 1.20.1 client. 1.20.2 has no issues

[craggy trout]

Are you using ViaVersion on proxy?

[edgy crow]

No it’s on backend

Made that change ~2 weeks ago

I just reproduced these logs right when I posted them

[reef fulcrum]

Waterfall supports 1.20.2

Reproduce without plugins, etc

[edgy crow]

Without plug-ins on the waterfall side or the backend server ?

[craggy trout]

Waterfall

[reef fulcrum]

Maybe even both

[edgy crow]

These are the only plugins we have

On waterfall

[reef fulcrum]

Which are still plugins

[edgy crow]

Give me a moment…have to make sure no one is online.

Damnit. 2 people joined since I posted this. Will try later

[exotic crypt]

any way to connect a fabric server to waterfall?

[marble kelp]

use a mod like FabricProxy (not sure if this is still the best solution)

[shy lagoon]

Hey I'm still having this problem

All modules are in the module folder
Force default server is set to false
I'm using a plugin called bungeeservermanager but I have also set force default to false in that config

People still get sent to hub when they rejoin instead of their last played server

If I haven't missed a step I'll remove server manager and see if it's to do with that

[reef fulcrum]

I mean, reproduce without plugins is all we can advice at this point

the players last server should be stored in the locations.yml file (only saves to the disk so often), but, that is used to determine where people end up unless you have force default turned on (or disabled that module)

[shy lagoon]

Yeah the locations.yml is being updated so I'm assuming that means the module is enabled, I posted here to make sure I didn't miss something obvious (it wouldn't be the first time 😂)

[craggy trout]

This is epic

[daring stirrup]

Is that a "new" message system for Waterfall?

[craggy trout]

Minimessage

[craggy trout]

Hey, would a tcp-fast-open PR be accepted?

It's supported my Minecraft clients

I would set it on initChannel, in the try/catch since isn't supported by windows

Seems good to acelerate the connections, and velocity haves it

[reef fulcrum]

I mean, sure, low effort, etc

[craggy trout]

Should i hardcode the mode to "both"? Or make it configurable

Makes sense to set it "3" always, since Minecraft take advantage of client-server

[reef fulcrum]

I have no idea, my heads spinning right now

[craggy trout]

I will set it into 3 mode, sorry for Disturb

[craggy trout]

Done, check it when you can

[exotic spear]

any idea how to fix this problem?

[nimble ember]

something messed with a packet

its literally impossible to tell what caused it from this error alone. reproduce without plugins

[exotic spear]

is it via version?

okay

[onyx carbon]

did you install protocolize-bungee?

[languid breach]

what was it

[exotic spear]

nvm it's fixed, i should have not used viaversion on proxy

[tulip surge]

message.txt by @slim oxide: https://pastes.dev/KIWUhjVD0J

[slim oxide]

with waterfall-548.jar

[reef fulcrum]

generally down to a client having been disconnected unexpectidely

[slim oxide]

game cilent show this

[reef fulcrum]

so, something mangled a packet

[slim oxide]

I can't even complete the connection. It just drops out.

[reef fulcrum]

reproduce without plugins/try a vanilla client

[slim oxide]

fixed,thanks for support

[craggy trout]

Did you check the PR?

[reef fulcrum]

No, I've spent the majority of the past week in bed

[craggy trout]

I'm sorry for hearing that, i hope you go up 🚀. Check it when you can, the only thing i did is implementing ChannelOption's tfo on the exception handling of ChannelException because it would not work in windows either.

[craggy trout]

This doesn't need tests barely, implemented on most of software like Paper, Velocity, bungee forks etc

[modest river]

been trying to fix this for 4 hours now, am using waterfall 1.18.2, with the build server (that I am trying to connect to) being 1.19.4

flushed my dns, didn't work

restarted my router, didn't work

it's NOT client side or my internet

other people have connected to the server just fine

it's only me for some reason

can I PLEASE get help for this

[modest river]

update: fixed it, it was an issue with via version

[spiral hatch]

you're outdated anyway

[modest river]

the modern versions also had an error

[craggy trout]

How ever, what's the point of using ViaVersion?

Lot of people comes here with same error

[spiral hatch]

you'll never get a good playerbase without it

[craggy trout]

I mean on Waterfall

[spiral hatch]

using via on any proxy is a death sentence

it has literally no idea what it's doing

it's guessing 🤷

[craggy trout]

At least Velocity doesn't use entity ID rewrite

[craggy trout]

How about impl MiniMessage for Waterfall? (I can do it)

[astral hornet]

can anyone tell me if waterfall works with forge? Specifically 1.20.1

[reef fulcrum]

Basically, no, no proxies work with forge

Velocity + the ambassador plugin are basically your best bet

unless you're within certain mod realms there is the odd waterfall fork which tries to add some hack for it which requires a client mod

[astral hornet]

Hmmm ok, Velocity only goes as high as 1.16 unfortunatly

[reef fulcrum]

No

the release build is stupidly outdated due to a barage of stuff happening, the snapshots are recommended and should generally be stable

[astral hornet]

hmmm ok. I was going by what is in their compatibility page

[reef fulcrum]

what compatibility page?

[astral hornet]

https://docs.papermc.io/velocity/server-compatibility#minecraft-forge-113-and-above

i was wrong, it's 1.13 and above

[reef fulcrum]

Yes, exact same as literally every other proxy

the ambassador plugin is effectively a hack which tries to add support for forge to velocity

[astral hornet]

hmmm ok

[reef fulcrum]

Hopefully the changes in 1.20.2 allow us to restore forge support going forward, but, until frameworks rely on that mechanism (and velocity makes the needed changes, waterfall/bungee will probably be a lost cause, but, who knows)

[astral hornet]

Gotcha, from what i have heard from the ATM modpack team. They are hoping to move away from forge (Because in their own words, the owners an ass) To another version that is better or something, they are stuck on forge 47.1.3 for some reason i don't know

[reef fulcrum]

lex is gone

[astral hornet]

I'm guessing that was the owner?

[reef fulcrum]

yup

[astral hornet]

ohh i dunno then, i'm guessing they are committed to making the move

[fossil dagger]

im a moron!

[fossil dagger]

[18:05:46 INFO]: [/IP:61889] <-> InitialHandler has pinged
[18:06:01 INFO]: [/IP:61893] <-> InitialHandler has connected
[18:06:02 INFO]: [wemmbu|/IP:61893] <-> ServerConnector [hub] has connected
[18:06:02 INFO]: [wemmbu|/IP:61893] <-> ServerConnector [hub] has disconnected
Issue connecting to the HUB

[fossil dagger]

Please help

[delicate phoenix]

Is there anything in the HUB console, that would indicate why the player was disconnected?

[fossil dagger]

Nope. Not even that they were trying to connect.

[delicate phoenix]

Sounds like some short of network misconfiguration, the tips given in #velocity-help apply here too. Check the firewall, make sure you are using the correct serbur addresses, etc.

[fossil dagger]

That's all right

[fossil dagger]

It's all hosted on the same machine so it is automatically localhost

[delicate phoenix]

Can you share proxy logs?

[fossil dagger]

Sure

What file lol

[delicate phoenix]

From the previous conversation in #velocity-help I know that you are running puffer using the standalone version, so pretty much just binding all backend serburs to localhost, and the proxy to the public IP should work.

[tulip surge]

latest.log by @fossil dagger: https://pastes.dev/8kARNUcQvz

[fossil dagger]

there u go

[delicate phoenix]

Strange logs, to me this looks like a network issue. Something isn't right with your network, or something isn't set up properly.

Just to make sure, all serburs are bound to localhost, and each has a unique port? And in the waterfall config, each was added separately?

[fossil dagger]

Yeah

Lemme gime you the config

[delicate phoenix]

Hm, are you using the same network by any chance? As in, are you the member of the internal network, where the serbur is running? In that case you could try directly connecting to a backend server, to see if that works in the first place.

[fossil dagger]

You mean is the server running on my internet?

[delicate phoenix]

Yes.

[fossil dagger]

Nope its not

It's ran in Amsterdam

the netherlands

[delicate phoenix]

Hm, I see. Did you change any firewall settings by any chance (on the machine)?

[fossil dagger]

Yeah

Port forwarded them all

Ill give you a list of forwarded ports

[delicate phoenix]

🤔

[fossil dagger]

Some other fact idk if this is going to help but it uses Firewalld

A Linux 22.04 ubuntu system.

[delicate phoenix]

If you had to port forward, then I presume this is some short of colocation?

[fossil dagger]

It's a VM

On Oracle.

idk if this helps but ive been running it for a year

50 GB of RAM and 8 OCPUS

So not free

But paid

[delicate phoenix]

Hm, interesting, last I used Oracle, port forwarding wasn't needed, like usually with a cloud provider, but changing the firewall settings they provide.

[fossil dagger]

Yeah

Idk if u mean it ill send the basic cmd

firewall-cmd --permanent --zone=public --add-port=PORT/tcp

U add it to the fire wall

Thats what i call port forwarding its a lil bit different than on the WIFI router etc but yeah

I might be stupid.

[delicate phoenix]

Oh yeah, that's the firewall running in the VM; On their web interface, they provide access to a different firewall, that is per Virtual Cloud Network, or what they call it.

Try checking if that's correct. https://docs.oracle.com/en-us/iaas/Content/network-firewall/overview.htm

[delicate phoenix]

Oh yes, port forwarding is a different concept. 😄

[fossil dagger]

I have them all firewalled.

[delicate phoenix]

Alrighty, just making sure no additional rules were injected. I mean, according to your latest log, the player was able to reach the proxy, so the firewall shouldn't be an issue.

[fossil dagger]

Indeed.

[delicate phoenix]

From the info given, all I can say is internally, within the VMs internal network, stuff is not getting routed properly. If the backend serbur doesn't even show a sign of a connection, then likely there is something blocking connections to the localhost interface, like SELinux.

[fossil dagger]

This might sound rude

but can you explain this in english

not meant rude at all btw ^

[delicate phoenix]

Basically, imagine the VM as a box. Outside, is the external network, and inside is the internal. An example of an internal network is the localhost interface, (lo1 for as an example). Now, you bound all your backend serburs to localhost, which essentially means you bound your serburs to this interface, therefore they are a part of the internal network.

Connections reach your VM, your "box" from the outside, the external network. Your proxy should be listening for these connections, handling them, and redirecting the player to the correct backend server. From the point the player is connected to the proxy, it reached the internal network, and is "in" the box. (The player isn't exactly "in".) Of course it doesn't work like this, but as a concept, you can think of it like this. Now the issue with your setup, is that at the stage when the player connects, and the proxy redirects the player to the defined backend, the backend does not respond for some reason.

This could be for multiple reasons, one being the OS blocking connections from reaching the loopback interface (lo1). (This is a very barebones anology I thought of on the fly, please don't hurt me if you don't agree.)

[fossil dagger]

hmhm

had a friend who did do this on free oracle on Puffer

[delicate phoenix]

We could try for hours to debug your setup, and eventually find the issue, but with Oracle Linux, and a Puffer Standalone setup, I would just recommend switching to Ptero, or using the Puffer docker version. I for some reason think it'll just work if you follow the docs. :P

[fossil dagger]

Pterodactyl is so painfull to setup 🙏

[delicate phoenix]

As long as you follow the docs, it's not that hard.

[fossil dagger]

Can you setup multiple servers ez?

[fossil dagger]

Idk what i made it change, (Maybe just the firewall updating) but i got it to work rn

Lobby works, as do LifeSteal and HardCore

Tried every waterfall feature and cant find issues.

And i didnt go thru the pain of installing Ptero luckily

[obsidian steeple]

Heyo
When I updated the Paper servers in my waterfall network to 1.20.2 and I want to join a 1.20.2 server with client 1.20.2 I get following error:

I have the latest version 548 of waterfall installed. someone know why I face this problem?

[craggy trout]

Can you show your proxy logs

[obsidian steeple]

sure

[gloomy igloo]

hi, can someone please tell me why my waterfall server is yelling at me?

[gloomy igloo]

solved

[fleet shard]

Hello, i'm using waterfall and when a player try to connect it doesn't connect to the hub, but it connects to the last server he disconnects

[fleet shard]

someone can help me please?

[reef fulcrum]

that's a feature

enable force_default_server in the config

[undone siren]

having the same issue has freaki2010

[tulip surge]

message.txt by @undone siren: https://pastes.dev/6aEvtPryv9

[reef fulcrum]

reproduce without plugins

[fleet shard]

thanks 🙂

[hazy niche]

how i can clone waterfall src?

[marble kelp]

git clone <url>

[hazy niche]

I cannot edit the files

[reef fulcrum]

clone repo, apply patches

[hazy niche]

./waterfall?

[craggy trout]

"sh waterfall build"

[marble kelp]

pretty sure the readme explains it

[reef fulcrum]

yea, you need bash

[hazy niche]

lol

ah

so?

[marble kelp]

yes

[hazy niche]

thanks ❤️

,i haven't had anything to do with minecraft for a long time, sorry haha

Can I delete the bungeecord/patches folder and the script or?

[marble kelp]

you can delete anything you want, it's your PC...

but it of course depends on what your goal is :P

[craggy trout]

🚀

[hazy niche]

lol i asked this question because i thought this folder had something to do with the waterfall proxy folder

but nevermind

[craggy trout]

You should read contributing.md

[hazy niche]

Is there a fallback function? so if a subserver is stopped that you are sent directly to the lobby

[marble kelp]

you mean players that are on the server? not really. But there are aplugins for that.

[empty urchin]

hello, does waterfall support 1.20.2?

i cant find a build for this version

[marble kelp]

yes it does

[empty urchin]

which verison?

[reef fulcrum]

the latest builds support 1.20.2

[empty urchin]

so why it isnt supporting my server?

thats what is hapening

nothing in console

[marble kelp]

show the log

[empty urchin]

https://pastebin.com/

whats wrong

i have waterfall build 1.20-548

[reef fulcrum]

All we can guess is to reproduce without plugins

[empty urchin]

it was working fine until a switched to 1.20.2

bcs of iadder... xd

and those are only plugins i am using on waterfall server

[reef fulcrum]

Which are still plugins which could be causing the issue

if you're going to generally dismiss what we say, glhf

[empty urchin]

sorry

dont be mad..

i will try to remove them

yeah

some plugin is doing it

i will put them one by one

yes

its okey now

some plugin was outdated

[exotic lintel]

Hey, anyone an idea how to fix ?

[craggy trout]

Try on a vanilla enviorment without plugin and make sure to have latest build

[exotic lintel]

Have fix it 😄
It was from "Triton" ^^

[uneven steppe]

https://media.discordapp.net/attachments/906983088737894450/1167765621585485905/image.png?ex=654f517c&is=653cdc7c&hm=f3ba1406448c246e5998a3080e85d4fec4a873af3387439abfcc2cc95d38d741&

are any of these OS preferred over another for a waterfall proxy server? I'm trying to buy a cheap vm for the proxy and don't really know anything about Linux other than a few ssh commands

[exotic crypt]

judt get a dedi off hetzner

ubuntu/debian are the most used id say

[uneven steppe]

hetzner is like 40 euro a month

[exotic crypt]

128gb ram, i9 9900k for 40€

its one of the best deals u can get

but probably u dont need all thay

[uneven steppe]

a proxy can run on less than 500 MB RAM

[bitter lark]

if only they had UK datacenter 😐

also I'm curous... why is Waterfall EOL? 😦

[reef fulcrum]

Because it's generally a platform with no future which isn't just upstream merges

[exotic crypt]

raw words: bungeecord but better

[reef fulcrum]

I don't have the will to deal with the platform anymore as anything sizable just breaks everything and nobody generally wants to interact with the platform in a wider sense to fix some of the headaches

[exotic crypt]

germany/finland be still good

[reef fulcrum]

e.g. theres the buggy ass tablist rewriting shit which has caused stupid desync bugs for years

[exotic crypt]

depends on player amount, mainly yeah u want 1-2gb ram

[bitter lark]

I was looking and FIN dc but was wondering how pings would look like for US ppl for example...

[exotic crypt]

200-600 average i guess

[bitter lark]

woah

[exotic crypt]

depends on their isp etc

[reef fulcrum]

France is probably the best place in the EU for the US in terms of DC availiability + transit

[exotic crypt]

where to get france datacenters

sob

[bitter lark]

OVH

[exotic crypt]

bru

overpriced

[bitter lark]

GRA/EIR

yea well... 😐

[exotic crypt]

not worth it for europe

[reef fulcrum]

SE england has a bunch of transit but most of the DCs I know are in financial london

[exotic crypt]

just go w hetzner

[bitter lark]

wondering how long they're gonna keep this up

lol

it's only first month tho

[exotic crypt]

me getting i9 9900k 128gb for 40€ ez :>

[reef fulcrum]

offer generally requires like a 18-24M commitment these days

[exotic crypt]

sob

so, u said waterfall is eol

[reef fulcrum]

I've been too dead to maintain it for the past 4 years

[exotic crypt]

everyone in future will have to use velocity/bungee?

[bitter lark]

but if u don't pay upfront they're gonna charge u full amount on second month

-.-"

[reef fulcrum]

I'm slowly getting better, most of my fatigue issues are gone, but, basically, yes

[exotic crypt]

ah k

[reef fulcrum]

that's the commitment, not upfront

[bitter lark]

well... I don't like it -.-

also...

lol

[exotic crypt]

go hetzner :^

[bitter lark]

all in all it feels like ovh is slowly jacking up pricing on everything they can

I'm considering it

[reef fulcrum]

le 2202

Dear person, We detected that one of the oxygen atoms you breathed in was filtered through our DC, please see the attached bill for that service

[bitter lark]

lmao

actually I think fsn would be better than hel

pinging from OVH BHS DC

[reef fulcrum]

using directions because it shows 2 terminal points

bearing in mind that most transit basically ends up in france and maaaybe I wanna say norway?

or, well, all of the transit basically ends up going through the baltic sea to sweden and the few places around there? either way, more miles of fibre to travel

[bitter lark]

bhs to eri and bhs to gra

so bhs to hetzner fsn is not bad at all

certainly not worth paying double for 10, 20 ms

well... I guess I'll be moving stuff again

hope hetzner has good antiddos

[craggy trout]

OVH have's way better one.

[exotic crypt]

but hetzner be cheap sob

[bitter lark]

does it actually make any difference?

[stiff falcon]

anyone know why 1.19.1/1.19.2 cannot join anymore?

java.lang.NullPointerException: Cannot invoke "net.md_5.bungee.protocol.packet.EncryptionResponse$EncryptionData.getSalt()" because the return value of "net.md_5.bungee.protocol.packet.EncryptionResponse.getEncryptionData()" is null
this is the error ^

[prisma ridge]

Are plugins on the server?

[stiff falcon]

yes. but the error comes from the proxy

[craggy trout]

Could you send full logs

[stiff falcon]

of the proxy?

[craggy trout]

Some plugin is trying to invoke EncryptionResponse data wronly

[craggy trout]

Yes.

[stiff falcon]

just a btw, we have freedom chat installed

[reef fulcrum]

nothing to do with freedom chat, or, shouldn't be

[stiff falcon]

https://mclo.gs/pOXbXtF

plugins on the proxy

[craggy trout]

Try without plugins

[wintry cobalt]

Hoping someone may know why this is happening on Waterfall. This happens on Velocity too, except they are errors instead of warnings. I don't think I have seen this before until recently. I thought maybe it was because I had servers offline, but when I remove them for troubleshooting it still happens. I am running Waterfall Build 548.

[delicate phoenix]

Could just be a scanner, some IPs in there are from known hosting providers.

[wintry cobalt]

That's strange, I am self hosted.

[delicate phoenix]

Yup, these are external IPs, connecting to your proxy, but "not fully completing" the MC connection.

[wintry cobalt]

That is kind of what I was thinking, but not sure why or where they are coming from or if I should block them in the security appliance.

[delicate phoenix]

Mostly they are harmless, they are just bots or scanners scanning the internet. But, they can also be considered a security risk from a standpoint, so you could try blocking them, if you want.

[wintry cobalt]

Doing some more research on that port and it looks like I should probably block it. I don't want the spam in the console either.

[delicate phoenix]

Make sense. It's interesting that instead of using a range, they use one specific port. They are likely looking for something specific on the internet. 🤷‍♂️

[wintry cobalt]

The good thing is that port is already blocked. I block everything by default and only open what I need. So I just need to block the IP's.

[delicate phoenix]

Consider looking for some sort of software that will do the blocking automatically for you, if this is a common occurrence.

[wintry cobalt]

That is what I am going to do. The servers sit behind Ubiquiti firewall. So I will create a rule that blocks on IP's pinging 61000

[delicate phoenix]

👍

All though, I must mention, that the address you see in the Waterfall console, is the full "inetSocketAdress" where the packet was sent from, not where it was sent to.

So 100.100.100.100:61000 means that 100.100.100.100 sent the packet from, "using" port 61000.

[wintry cobalt]

Correct

I am just going to block the entire IP

[toxic sun]

Having my partner test a connection from latest waterfall build, joining to latest 1.12.2 hub also papermc but he keeps getting this error, any ideas?

[tulip surge]

java.lang.IllegalArgumentException_.txt by @toxic sun: https://pastes.dev/AeNElWRXat

[reef fulcrum]

that would scream a bad config file

[toxic sun]

I think its -Dpaper. disableChannelLimit=true

[reef fulcrum]

no

it's the connection timeout setting

[toxic sun]

Hmm okay

it was -1

now its 10000

does -Dpaper. disableChannelLimit=true do nothing?

[reef fulcrum]

on waterfall? no

[toxic sun]

No on the hub

Im unsure it its bungee or hub throwing the error

[reef fulcrum]

on paper it has an effect, yes

[toxic sun]

I added it to startup parameters and the hub wouldnt start

[dull vortex]

Hello, My Waterfall Bungee Server won’t connect to any Server. I can’t find the solution. I watched a tutorial for connecting all Server to the Bungee and followed all steps, but every time I try to join over the Bungee IP, I become the Error Message: Couldn't connect to a default or fallback server. Incorrectly configured address/port/firewall. But the ports are all correct (I think). You can look into my Waterfall Config and the server properties of my lobby. Please help me.

[tulip surge]

server.properties by @dull vortex: https://pastes.dev/tWdRhmUGlI

[tulip surge]

config.yml by @dull vortex: https://pastes.dev/JpcxpYvdXm

[craggy trout]

Seems a firewall or wrong ip/port config including all backends

[dull vortex]

But all ports are free to use

its all localhost

I don't know if waterfall causes the problem or the managing software pterodactyl

[dry thunder]

`connection timed out` / `connection refused`

unable to connect to server
io.netty.channel.ConnectTimeoutException: connection timed out``` ```
unable to connect to server
io.netty.channel.AbstractChannel$AnnotatedConnectException: finishConnect(..) failed: Connection refused```
**These exceptions indicate that your proxy cannot connect to the address you've specified.**
There can be several reasons for this:
- The server is not running
- Your proxy is being blocked by e.g. a firewall
- Your network is set up in a way that requires you to use a different address.

If you're sure your server is running, you haven't mistyped the address and the proxy's not being blocked, try following the steps below:

If all your servers and the proxy are on the same machine

Try using localhost as the servers' address.

If you're using Pterodactyl or another panel that utilizes Docker

In general Docker containers are network isolated and will require you to use a different address to reach the machine and not the container.

In Pterodactyl's case the default address to reach the machine is 172.18.0.1 (read more here).

If you're using a host

Speak to their support. Your host should know their own network the best and help you set your proxy up.

[dull vortex]

Thank you, that sounds helpful

[dull vortex]

No, still don't work : (

[craggy trout]

Does the connection even ping on backend

[dull vortex]

no

still trying to fix the problem. It seems that the problem is caused by the proxy and not the backend

[dull vortex]

Now it's working and I still don't know what actually caused that problem xd. I tried it with public IP instead of localhost. Maybe that was the problem. I think something in pterodactyl blocks the connection. But thanks for the help!

[craggy trout]

Yes i said you it was a ip/port/firewall issue

[rustic mesa]

Hi! I've got a serious problem with my waterfall...

I have (in theory) 3 servers: my proxy (Proxy_DEV-1), my lobby (Lobby_DEV-1), and my dev server (Game_DEV-1).

To solve this problem, I've tried :

• Deactivate the plugins one by one (I don't have any plugins on the waterfall at the moment)
• Run netstat -ltnp to see if another process was using port 60000 (my proxy is the only one using it)
  - Give my vps a hug

2 more infos:

• I'm running on CloudNET-V4
• The Lobby doesn't receive any connection requests (so it's really the proxy)

Here is the logs: https://mclo.gs/TpkcSsg

[craggy trout]

Try on a vanilla enviorment without plugins

[exotic lintel]

Hey,
anyone an idea how to fix this ?

ERROR14:20:52
Couldn't register custom payload
Console14:20:52
java.lang.IllegalArgumentException: Channel must contain : separator (attempted to use )
at org.bukkit.plugin.messaging.StandardMessenger.validateAndCorrectChannel(StandardMessenger.java:503) ~[purpur-api-1.20.2-R0.1-SNAPSHOT.jar:?]
at org.bukkit.craftbukkit.v1_20_R2.entity.CraftPlayer.addChannel(CraftPlayer.java:2364) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.server.network.ServerCommonPacketListenerImpl.handleCustomPayload(ServerCommonPacketListenerImpl.java:146) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.network.protocol.common.ServerboundCustomPayloadPacket.handle(ServerboundCustomPayloadPacket.java:46) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.network.protocol.common.ServerboundCustomPayloadPacket.a(ServerboundCustomPayloadPacket.java:12) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.network.protocol.PacketUtils.lambda$ensureRunningOnSameThread$0(PacketUtils.java:53) ~[?:?]
at net.minecraft.server.TickTask.run(TickTask.java:18) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.util.thread.BlockableEventLoop.doRunTask(BlockableEventLoop.java:153) ~[?:?]
at net.minecraft.util.thread.ReentrantBlockableEventLoop.doRunTask(ReentrantBlockableEventLoop.java:24) ~[?:?]
at net.minecraft.server.MinecraftServer.doRunTask(MinecraftServer.java:1351) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.server.MinecraftServer.d(MinecraftServer.java:193) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.util.thread.BlockableEventLoop.pollTask(BlockableEventLoop.java:126) ~[?:?]
at net.minecraft.server.MinecraftServer.pollTaskInternal(MinecraftServer.java:1328) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.server.MinecraftServer.pollTask(MinecraftServer.java:1321) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.util.thread.BlockableEventLoop.managedBlock(BlockableEventLoop.java:136) ~[?:?]
at net.minecraft.server.MinecraftServer.waitUntilNextTick(MinecraftServer.java:1299) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.server.MinecraftServer.runServer(MinecraftServer.java:1187) ~[purpur-1.20.2.jar:git-Purpur-2086]
at net.minecraft.server.MinecraftServer.lambda$spin$0(MinecraftServer.java:320) ~[purpur-1.20.2.jar:git-Purpur-2086]
at java.lang.Thread.run(Thread.java:833) ~[?:?]

Mc Version 1.20.2

[reef fulcrum]

something tried to send a bad plugin message, generally a janky client;

[exotic lintel]

So thats not from the Server ?
Its from an Client ?

Users have this if they use Forge (latest)

[reef fulcrum]

The server was processing a packet sent by the client

[exotic lintel]

ok thx

[tidal musk]

I have 3 servers, the waterfall proxy, the limbo and the lobby. I have set in the configuration two servers, the limbo on port 10 and the lobby on port 11. By default when entering via proxy sends you to the limbo, by console in the proxy I do a /send all Lobby, when I try to send players to the lobby I get this error:

[20:35:06 INFO]: Attempting to send 1 players to Lobby
[20:35:06 INFO]: Send Results:
[20:35:06 INFO]: FAIL: 0
[20:35:06 INFO]: ALREADY_CONNECTED: 0
[20:35:06 INFO]: SUCCESS: 1
[20:35:06 INFO]: EVENT_CANCEL: 0
[20:35:06 INFO]: ALREADY_CONNECTING: 0
[20:35:06 INFO]: [ImRex__|/127.0.0.1:51427] <-> ServerConnector [Lobby] has connected
[20:35:06 INFO]: [/127.0.0.1:51427|ImRex__] <-> DownstreamBridge <-> [Limbo] has disconnected
[20:35:06 WARN]: [/127.0.0.1:51427|ImRex__] <-> DownstreamBridge <-> [Lobby] - could not decode packet! io.netty.handler.codec.DecoderException: com.google.gson.JsonSyntaxException: com.google.gson.stream.MalformedJsonException: Use JsonReader.setLenient(true) to accept malformed JSON at line 1 column 5 path $
[20:35:06 INFO]: [ImRex__] disconnected with: The server you were previously on went down, you have been connected to a fallback server
[20:35:06 INFO]: [/127.0.0.1:51427|ImRex__] -> UpstreamBridge has disconnected
[20:35:06 INFO]: [/127.0.0.1:51427|ImRex__] <-> DownstreamBridge <-> [Lobby] has disconnected```

[tough vale]

Check the backend logs for lobby

[tidal musk]

I fixed it using an older version of waterfall, exactly I used build 545.

With recent compilations I get the following errors

[steep prairie]

I need your help. I have a lobby and survival server. When I connect using waterfall, I get this problem

[05:15:41 INFO]: [siArang] disconnected with: Could not connect to a default or fallback server. Incorrectly configured address/port/firewall?

Can I enter using a domain and not an IP? The problem is that the provider I rented has a server for the IP using a domain and not an IP number.

[reef fulcrum]

a domain literally just maps to an ip address

speak to your host for support, assuming that the server is running, any competent host should know what's going on there

[steep prairie]

then why can't I join via bungeecord even though I have done the setup according to what is available in online communities and other online tutorials

[dry thunder]

`connection timed out` / `connection refused`

unable to connect to server
io.netty.channel.ConnectTimeoutException: connection timed out``` ```
unable to connect to server
io.netty.channel.AbstractChannel$AnnotatedConnectException: finishConnect(..) failed: Connection refused```
**These exceptions indicate that your proxy cannot connect to the address you've specified.**
There can be several reasons for this:
- The server is not running
- Your proxy is being blocked by e.g. a firewall
- Your network is set up in a way that requires you to use a different address.

If you're sure your server is running, you haven't mistyped the address and the proxy's not being blocked, try following the steps below:

If all your servers and the proxy are on the same machine

Try using localhost as the servers' address.

If you're using Pterodactyl or another panel that utilizes Docker

In general Docker containers are network isolated and will require you to use a different address to reach the machine and not the container.

In Pterodactyl's case the default address to reach the machine is 172.18.0.1 (read more here).

If you're using a host

Speak to their support. Your host should know their own network the best and help you set your proxy up.

[reef fulcrum]

would be my guess, but, given you've not provided the actual error, I can't say

[tulip rampart]

My server says connection refuses no further information

[reef fulcrum]

see above

[tulip rampart]

Is it possible to setup via version with bungee/waterfall

[reef fulcrum]

generally, yes

apparently there is a bug with installing it on bungee, however

[tulip rampart]

So u just put it in the plugins folder

I*

[reef fulcrum]

if it works, yes

[tulip rampart]

Alright

Should I just use the spigot one and put it on the proxy server

Or should I put it in the hub server

[reef fulcrum]

It's recommended to install it on all of the backend servers

you can generally however install it on the proxy, but, once again, that might not work atm

[tulip rampart]

Why not

[reef fulcrum]

.

[tulip rampart]

Oh the bug

Alright then

[tulip rampart]

Is this the bug?

[reef fulcrum]

yes

[tulip rampart]

Any idea when it will get fixed?

[reef fulcrum]

nothing to do with us

[tulip rampart]

Does bungee cord have a server that I can ask for this bug in?

[reef fulcrum]

Yes, the spigot server, but md has already said it's not his issue

(even if it's down to bungee or not, they don't provide support for protocol hacks)

[tulip rampart]

Oh

Anyways thank you so much for the help

Got it working

[wintry cobalt]

Hello, is there something in Waterfall that alters there render distance? The reason I ask is I had to downgrade to 543 from 548 and the render distance of all my servers have changed.

I know there is something in Paper itself, but it's strange I have not changed anything in paper and there render distance has completely changed just from downgrader to waterfall.

[reef fulcrum]

the only thing that could be an issue in regards to waterfall is the client settings packet

there is some issue in waterfall which causes the client settings packet to go missing, recent builds have a "fix" for that

[tulip rampart]

i added bungeeguard but now this shows up

[spiral hatch]

make sure waterfall has bungeeguard enabled

[pine sinew]

[21:04:31 WARN]: [:|xdmonkeyboi4] <-> DownstreamBridge <-> [hub] - could not decode packet! io.netty.handler.codec.DecoderException: com.google.gson.JsonSyntaxException: com.google.gson.stream.MalformedJsonException: Use JsonReader.setLenient(true) to accept malformed JSON at line 1 column 4 path $

on latest version of waterfall

just says this when trying to join

i just changed waterfall versions

from a semi older build

[21:07:01 WARN]: [] <-> DownstreamBridge <-> [hub] - could not decode packet! io.netty.handler.codec.DecoderException: com.google.gson.JsonSyntaxException: com.google.gson.stream.MalformedJsonException: Use JsonReader.setLenient(true) to accept malformed JSON at line 1 column 4 path $

[21:07:21 WARN]: [/|cristiancraft132] <-> DownstreamBridge <-> [hub] - could not decode packet! io.netty.handler.codec.DecoderException: com.google.gson.JsonSyntaxException: com.google.gson.stream.MalformedJsonException: Use JsonReader.setLenient(true) to accept malformed JSON at line 1 column 4 path $

all of my 1.8 players are having issues

im able to connect on 1.19 tho

[reef fulcrum]

all I can see is that something tried to send bad json

[pine sinew]

i was just on an older build tho

and it worked

:/

i uploaded the latest version from like 4 hours ago

to my server

and people were just on before i did this

these same people

no other plugins were modified

i was on Waterfall-Bootstrap:1.20-R0.2-SNAPSHOT:4e5a9e4:547

[reef fulcrum]

I mean, very little changed in that build

only a few packets would of been implicated by that

so, idk what to say

[pine sinew]

im gonna download build 547 and put it on my server again straight from ur site

see what happens

[reef fulcrum]

Theres no real way of seeing what went sideways without basically adding some debug logging or something

[pine sinew]

how would i debug it?

[reef fulcrum]

eerrrr, no idea

[pine sinew]

:c

[pine sinew]

reverting back to build 547 works

[reef fulcrum]

and the only real changes after that are upstream merges

[pine sinew]

im gonna go build by build from 547 and see where it fumbles

[reef fulcrum]

-Dwaterfall.packet-decode-logging=true - you can maybe run with that to get a full stack trace out of the thing

but, at this point, it's 1:15am, and waterfall is EOL

[pine sinew]

wdym EOL?

[reef fulcrum]

end of life

[pine sinew]

why? D:

[reef fulcrum]

Because I really no longer care to maintain the project due to my health and the lack of community engagement over anything that involves cleaning up hanging turds

[pine sinew]

o

are people moving away from waterfall into velocity?

[reef fulcrum]

1/3rd of proxy servers on bstats are velocity

[tulip rampart]

I use bungee cord

How to enable it

[reef fulcrum]

We don't support bungeecord

and you'd generally install the plugin by installing it on the proxy and making sure you copy the secret over

[craggy trout]

You could let people who want to continue to maintain waterfall, maybe you came to your EOL for this project, not the project itself.

At least is how i think

[reef fulcrum]

Paper as a organisation no longer wants to maintain waterfall

[craggy trout]

I see... they could look for someone else they know to maintain the project

As much as they don't want to, it is still the second or first most used proxy (bungeecord above or below)

[reef fulcrum]

it's open source software

theres already people maintaining forks

[craggy trout]

Yes i know, but i mean as base.

Instead of making it only readable from github, leave it to someone (or organization) to maintain it, even I had several ideas for waterfall, imagine a bunch of expert people

[reef fulcrum]

Basically, I will maintain Waterfall as a minimal effort sort of deal with a while, I myself need to look into migrating an entire server off of the thing

[craggy trout]

Practically as it was this year, I think it doesn't cost to Upstream and PR should always be welcome.

[reef fulcrum]

But, we have little interest in just tossing off the project into the ether along with all the headaches that ensues

It costs time

[craggy trout]

However, right now there is no ideal situation, velocity is practically "dead" for now

[reef fulcrum]

and considering any effort into waterfall is generally trivial stuff considering that the ecosystem is stale and refuses to move forward generally, it's kinda like

[craggy trout]

Some people still like to use a somewhat redundant bukkit-based API

[reef fulcrum]

Spigot after like 10 years finally added some componet stuff to bungee, something we tried to do years ago

[craggy trout]

BungeeCord has a couple of nice PRs, such as improving entity rewrite, avoiding packet re-compression and async machine event attempts.

The problem is md_5, I don't understand why it would not accept this for breaking two 10 year old plugins

[reef fulcrum]

Which is literally my entire point

[craggy trout]

I think Waterfall should have or should break things that use old plugins, or at least add the option to upgrade the current system. In fact there is a PR that accumulates all those in a patch to waterfall.

[reef fulcrum]

Except then we get constant cries over breaking stuff

[craggy trout]

https://github.com/PaperMC/Waterfall/pull/812

This for example seems cool

[reef fulcrum]

I've already extended the door open to many devs to open the door for working on stuff in a more coperateive manner

So we can do stuff like get rid of the built in tab list rewriting BS which has done nothing for the past 10 years but caused bugs

and, nothing

[craggy trout]

I think it's proper removing whole entity rewrite

It is redundant in my opinion and makes bungeecord have to copy the entire buff into the decoder as I remember.

[reef fulcrum]

Pretty sure some waterfall fork already tried that and broke an entire bunch of plugins in the process

[craggy trout]

I have modified waterfall a lot

For example, MiniMessage is the main format for me.

[craggy trout]

The thing is that the main one for me, which is ViaVersion, doesn't work by itself.

[reef fulcrum]

basically, theres an entire general gist of, I don't wanna toss off this project to somebody else, it's still something I need myself, for now

[craggy trout]

I'm sure it's because it alternates the packet shape compared to vanilla and it's like "hacky" or something

[reef fulcrum]

It's just I'm apathetic to the entire project at this point, because theres far too little you can do which doesn't risk breaking plugins and then require like hours of work to patch the BS it causes up

entity metadata rewriting removal took literally months of attempts in order to patch up due to plugins doing weird stuff

[craggy trout]

I will open a couple of PRs, such as the TCP-FAST-OPEN fixed for the Downstreambridge only.

Since velocity has the wrong implementation of TFO as well and for it to work that way it would need implementation in the backend.

[reef fulcrum]

I mean, pretty sure I said like 2 years ago or something that TFO is literally useless right now

[craggy trout]

Amm.. stills on velocity

[reef fulcrum]

idk if linux has it enabled by default

Yes

ik, idk why he added it

[craggy trout]

I think so

[craggy trout]

The problem with this feature, upon investigation, is that it would only give attackers an advantage because the connections would come from linux and speed up the process. Since most of the players are windows players.

[reef fulcrum]

it removes a small amount of back and forth if you have a cookie

[craggy trout]

Yeah

[reef fulcrum]

imho, it makes literally 0 sense to support it for the proxies inbound

[craggy trout]

impl stuff like this

[craggy trout]

What ideas would you like to implement? I might be able to help with some of them.

[reef fulcrum]

https://sysctl-explorer.net/net/ipv4/tcp_fastopen/

[craggy trout]

seems you need to enable manually for work

[craggy trout]

I would like to open an implemented libdeflate compressor using the .SO compiled from velocity, but I don't know if I have permission (I already did this in an open source waterfall fork, but now is closed source)

https://github.com/LuciaNishimiya/MangoCord/blob/main/Waterfall-Proxy-Patches/0040-Implement-libdeflate.patch

[reef fulcrum]

We generally have perms to use velocity stuff in the paper org

[craggy trout]

According to some graphs it is somewhat faster than cloudfare's zlib in compressing and decompressing large amounts of data.

[craggy trout]

I will open one, use the compiled .SO and some native folders like this https://github.com/PaperMC/Velocity/tree/dev/3.0.0/native/src/main/java/com/velocitypowered/natives

[reef fulcrum]

velocity natives are already pushed into the paper repo

paper uses them, etc

[craggy trout]

But i'd need to build the so

Or no, i saw paper implementing this

[reef fulcrum]

natives are already in jar from velocity

[craggy trout]

Okay, i don't need that folder stuff

[maiden badger]

does waterfall support mc 1.8.8? @forest halo

[bitter birch]

(6545e263cdf07f25f6aca11d) // @maiden badger (@xioas / 148642095219998720) has been warned by @forest halo (105923848263753728)
Reason: do not ping random staff for no reason.

[maiden badger]

sorry

its urgent

[craggy trout]

Latest support from 1.8 to 1.20.2

It's a multi-protocol proxy

[maiden badger]

yess ty ty ❤️

[craggy trout]

Show Last month

I mean..

[twin solstice]

that'd just be stupid nitpicking

you could do the same to waterfall showing off only upstream updates & a 2 month gap between updates

[craggy trout]

I did not say waterfall is updated... i said there is no good situation for both

Read properly please

[twin solstice]

no you said velocity is dead without elaborating to which i reply to that it doesn't look like that

if you're going to tell me to read properly, atleast make yourself clear

[tulip rampart]

hey my friend is not able to join one of the backend server in the proxy

[craggy trout]

I am, you're not clear

[craggy trout]

I said that neither of the two is in an ideal situation, if you see the updates of velocity and waterfall these two months are practically adding version support and bug fixes.

Whether you are a denier of this software is another matter.

Besides velocity took weeks to support 1.20.2, when bungeecord already released the pre-version the same day.

That's why i said "there is not an ideal situation

[craggy trout]

Ammm, that's on #paper-help

[tulip rampart]

hi i added bungeeguard to my server and this error comes

[reef fulcrum]

I mean, this is the waterfall support channel; all we can say is "make sure you configured it properly"

generally, make sure that you actually installed the plugin on the proxy and copied the secret over, as per the installation instructions

[tulip rampart]

hello i linked my SMP server with the proxy and playerdata is not transfering?

[reef fulcrum]

"playerdata is not transfering" what data?

[tulip rampart]

like its wierd

when i connect to the SMP from proxy i dont have my loot

[reef fulcrum]

That's not weird

it's a proxy

it doesn't sync player data, etc for you

if you want that, you'd need a plugin which adds that to the servers

[tulip rampart]

yes but the data is saved in the SMP server right

why it not load

[reef fulcrum]

I mean, stuff you do on the SMP server will stay on the SMP server

[tulip rampart]

yes

but when i connect to the smp server it was empty

[reef fulcrum]

so, your items won't transfer from the lobby to the SMP server

[tulip rampart]

nono the items that where on the SMP are not there

[reef fulcrum]

then you'd need to work out why they're gone

make sure that you're in the correct place, make sure that you have bungee mode setup properly, etc

[tulip rampart]

they are there for players who dont join the server

is there a way to recover deleted items from a server

[wraith oriole]

Hey, i have geyser, and floodgate on my waterfall proxy, and floodgate on both of my backend servers (lobby, and survival server). whenever a bedrock player joins my server it says "Connection closed" is there anyone that can help me resolve this issue/bug?

@twin solstice can you please help me, ik im pinging you but i saw you earlier today helping others and i posted 5 seconds later and no answer this is something that i really need fixed as me and my friend are trying to play my server if you understand this can you plas help me, my post/bug issue is above.

[weary shadow]

does anyone know about a waterfall plugins that can replay chat messages after this configuration thingy when switching servers?

[craggy trout]

Only i can say you is something refused the connection

Check your firewall i.e if you have ovh its system always asks for a TCP ping before a UDP ping to accept UDP connections, you must add a rule to the firewall to avoid that.

[tulip rampart]

Level.dat file Isint loading for players who join via the hub but if u join with the normal in then it works

[craggy trout]

For what?

[craggy trout]

No

[normal depot]

Hello!
I'm finding my self in an interesting situation.
I have a Paper server and a Waterfall server. I can join the Paper Server fine but when I go to join through the Proxy, I get the error "Invalid signature for profile public key. Try restarting your game."
So I restarted my game, it does it every time.
I restarted the servers. Still happening.
I changed enforce-secure-profile to false on both the proxy & game server. It still happens.
Here is my server.properties and Waterfall config.yml

[tulip surge]

server.properties by @normal depot: https://pastes.dev/oqqUo5EVSM

[tulip surge]

config.yml by @normal depot: https://pastes.dev/Y7gO0EfCjY

[reef fulcrum]

The settings are irrelevant for that

see client logs, etc

[normal depot]

But it's happening to more than 1 person

[reef fulcrum]

that changes literally nothing of what I've said

[normal depot]

So my client logs just say the same thing as the error.

[15:23:08] [Render thread/WARN]: Client disconnected with reason: Invalid signature for profile public key.
Try restarting your game.

The only other thing is, when first loading the game, it shows this in the logs

[15:12:03] [main/WARN]: Invalid registry value type detected for PerfOS counters. Should be REG_DWORD. Ignoring: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\PerfOS\Performance\Disable Performance Counters.

[reef fulcrum]

No idea

outside of checking your system clock, all we can tell is is what it says; your client sent a bad key

[normal depot]

I wonder if the recent time change has messed with something

hm..

[spiral hatch]

this happens on mine but it doesn't effect anything

[normal depot]

So you get this error and can still connect?

[spiral hatch]

yep

[tulip rampart]

can somone one tell me why gyser not downloading

[tulip rampart]

skin restorer not workingplease help

on bungeecord

[craggy trout]

Hey... what can I tell you this is Waterfall support not external plugins but, still pass the logs?

[vivid vapor]

I updated Waterfall to the latest version yesterday and ever since it seems to be randomly not letting people in saying

io.netty.channel.unix.Errors$NativeIoException: recvAddress(..) failed: Connection reset by peer```online mode is enabled and the accounts (including mine) are all premium

[reef fulcrum]

Mojangs servers are having issues

[vivid vapor]

ah lovely

weird thing is it almost always works on the second time lol

[craggy trout]

Probably we could add to Waterfall some cache session for 24h?

I managed to impl this, and works fine

[craggy trout]

So the connections that have already passed will be able to continue playing until the servers are fixed or the cache expires

[vivid vapor]

i dig that idea

honestly i don't rly see a reason not to cache it, it's not like a player is gonna un-buy the game and play cracked after so nothing's gonna change

[stark sequoia]

We're definitely not putting an arbitrary authentication server just because every service on this world is down for a while once and then

[craggy trout]

I can make a PR, i will after finishing zlib replacement with libdeflate

[vivid vapor]

if it even gets accepted, nacio seems against the idea lol

[craggy trout]

Well he didn't understand the point

[stark sequoia]

How does the flow go then

[craggy trout]

This system uses the LoginResult and caches its results, such as the address, this does not add any kind of separate "arbitrary" server.

[craggy trout]

It could use for example TimeUnit hours for adjustable time, or a hardcoded one like for example 24h

[craggy trout]

This is very useful, especially for large servers that already have something similar to this implemented.

You can't not have something for this type of situation.

And adding this in Waterfall and velocity would be ideal, I can do it in Waterfall

[vivid vapor]

especially since the servers have been having issues like all day

[craggy trout]

Yep, i will pull request this when i finish compressor improvements PR

[stark sequoia]

Again, what do you want to cache and how does that authenticate a player that never joined or left; or make the PR

[vivid vapor]

from what i understand it caches the player when they join while the auth servers are working, then if they go down that player can still join

[craggy trout]

You cache the player's connection itself for next time, not the server.

[craggy trout]

That's why I told you that you didn't get it, this doesn't make sense. In that case it would be much better to implement servers that would also serve to verify premium connections.

[craggy trout]

Really the only thing I see "bad" about this is that it can increase a bit the ram consumption, in exchange for providing excellent support.

[craggy trout]

We can also set expiration times of a couple hours, this will give Mojang time to repair their servers, so don't expect it to take too long.

[stark sequoia]

Yeah, so this won't be happening

You're not truly authenticating the player and have no way to verify the player you assume is the same actually is

If you want to implement something like that, do it in a plugin

[craggy trout]

?

[reef fulcrum]

Theres already plugins which do that

[craggy trout]

I think would be worth impl on proxy natively

[reef fulcrum]

no

[craggy trout]

Why

[reef fulcrum]

1. the project is dead

2. The feature introduces a few dozen headaches such as concerns over security

3. There are already plugins which do it, I would much rather waste my time working on API to resolve plugin limitations than trying to shove in features into the proxy, especially when plugins already exist

[craggy trout]

I think that one doesn't really count

[reef fulcrum]

I mean

dead projects generally don't accept sizable features

[craggy trout]

I guess is dead to you, I still need it

And seems lot of people tho

[reef fulcrum]

That doesn't magically turn it into a project I'm willing to spend hours on.

[craggy trout]

If I remember correctly, you still wanted to keep updating it and add a couple of things to it.

[reef fulcrum]

generally low effort things given that I'm still relying on it myself for now

I've generally given up with anything major given that between my health and the lack of care from the ecosystem, it's not worth any substantial investment from us

[onyx sorrel]

is there a waterfall plugin that detects what version the user is on and automatically puts them in a specific server if they are on that version

?

[craggy trout]

That's too specific

[true crane]

How can I fix this?
https://pastes.dev/U3tBf6WmlP

[delicate phoenix]

Something sent malformed JSON. Do you have any plugins installed on the proxy?

[true crane]

ya

[delicate phoenix]

Could you try reproducing the issue without plugins? You can make a separate test network with pure Waterfall and Paper, and then see if that works.

[true crane]

alr

[craggy trout]

Mostly like "negativity/ultimateantibot" issue

[true crane]

btw this is what the server I tried to connect says https://pastes.dev/6NIGfQczMb

[craggy trout]

Try on an vanilla enviorment i.e Latest waterfall without plugins

[true crane]

Just removing the jar files will do right?

[delicate phoenix]

Yup, make sure to make a backup just in case.

[craggy trout]

Also you can run -Dwaterfall.packet-decode-logging=true for more useful information.

[true crane]

Still the same thing

[true crane]

on the console?

[craggy trout]

No, on startup flags

[craggy trout]

Do you have Waterfall on latest build?

[true crane]

ya

[craggy trout]

Try this on a vanilla enviorment without plugins and vanilla client etc

[true crane]

alr

[delicate phoenix]

Did the paper instance have plugins installed?

[craggy trout]

According to the log, It didn't even decode the proxy to the connection.

So likely it will not have reached the upstreambridge (backend)

[delicate phoenix]

Hm, interesting. By default nor Waterfall, nor Paper should send malformed JSON, so it's likely a plugin somewhere.

[true crane]

Might take time as the host I'm using doesn't let me edit the startup

[craggy trout]

The only plugins that I think can cause this are the "antibot" and the anti-cheat trying to verify or requesting something to the connection

[true crane]

I have to create a ticket to do that (and I have but it will take time)

[craggy trout]

Try without plugins, i can't reproduce on latest Waterfall without plugins

[true crane]

I do want to point out that online-mode=false

Maybe that can be the issue?

[craggy trout]

No

For the proxy to work, you must have it in offline mode on the backend. Otherwise it will throw a QuietException

[delicate phoenix]

Do note, that offline mode isn't exactly supported here, check #welcome for details on why.

[craggy trout]

True, but this error has nothing to do with that. Indirectly maybe because it uses a plugin that he wouldn't need if it was online mode, but I'll be glad to help him.

[true crane]

Wait it turns out the waterfall version is outdated

I'm so sorry let me quickly update the version and see

[craggy trout]

👍

[true crane]

I just need to switch the jar version right with the later one?

[delicate phoenix]

True, however it's still supporting piracy which is against the Discord TOS, and the rules of this Discord server.

[craggy trout]

I see it a bit redundant, since a user can come and before asking for help activate the online mode in the proxy. How ever keeps as opinion

I respect that, but they still have a patch that literally accepts cracked mode

[true crane]

It worked

Updating the version worked!

[craggy trout]

That's good

[true crane]

Thank you so much Simon and Ismael for helping and I'm so sorry for bothering you'll

[delicate phoenix]

I don't want to get into an argument about offline mode, and I see your point. Personally, I don't have anything against offline mode, however different places have different rules.
Here, understandably, as a large guild they enforce online mode to not endorse a potential TOS violation, which btw, in my opinion is debatable, in addition to avoid dealing with headaches that come as a result of not authenticating with Mojang. I'm here to help people, and I don't want to get into drama, so I follow the rules. 🤷‍♂️

[feral monolith]

Hey we are getting random Error authenticating USERNAME with minecraft.net errors. causing players to see Alex / Steve skins, is this something we can fix or is waterfall considered dead and we should switch?

[winged pilot]

that's not a waterfall issue

[craggy trout]

Mojang servers are suffering now

[lime river]

Hey, had a server running last night and the connection was fine. I restarted it and now I'm getting this whenever I try to connect

Error occurred processing connection for GutinGongoozler AnnotatedNoRouteToHostException : null: experimental/172.19.0.14:25565

I didn't change any settings

[reef fulcrum]

your network went derp

[lime river]

Is there anything that can be done to prevent it from happing? I've had it happen before, and a restart usually fixed it, but I don't want to kick everyone to fix a test server.

[reef fulcrum]

I mean, you're using an IP address inside of a private network interface

you;d generally need to work out why that interface is no longer routable

[lime river]

I think that's an artifact from docker

[twin solstice]

waterfall will cache dns

ip addresses are not persistent by default in docker

[lime river]

I see

[twin solstice]

dunno if there's a public plugin to fix it

either make your docker servers use static ip addresses or switch to velocity which as far as i can remember doesn't have the issue

[lime river]

I want to switch to velocity eventually, but the guy before me did and it caused a lot of issues. Idk what he did wrong but there was a significant memory leak.

I'll just set up a custom network with docker, should be able to have a static ip.

[plucky crag]

can anyone help me set up a Waterfall server im having a very hard time because there are no updated tutorials all I need help with is adding the server to my network

[reef fulcrum]

the config just has a list of servers, you would add an entry to that?

[obsidian ginkgo]

what exactly do you need help with specifically

have you tried something that’s not working

[plucky crag]

How would I do that

I dont know anything about this

its my first time making a network

I used to make just regular mc servers

[reef fulcrum]

it's yaml

[obsidian ginkgo]

you should probably look around people don’t usually spoon feed here

[reef fulcrum]

the format is generally self explanitory

[obsidian ginkgo]

it’s pretty basic stuff

if you try something and get errors etc you can ask but i mean generally people don’t wanna just spoon feed

[plucky crag]

so if this is what it says by server

i would do this

servers: {
lobby:
blah blah blah
}

or is that incorrect

[reef fulcrum]

I mean, no {}

[plucky crag]

so that without the brackets?

[obsidian ginkgo]

its not java it’s yaml

[reef fulcrum]

and you'd append them to the existing section, but, yes

[plucky crag]

okay thank you thats all I needed to know

<3

[reef fulcrum]

servers:
  main:
    motd: '&1&kJust another Waterfall - Forced Host'
    address: localhost:25565
    restricted: false
  main2:
    motd: '&1&kJust another Waterfall - Forced Host'
    address: localhost:25565
    restricted: false
...

[obsidian ginkgo]

wait whatre the brackets you put suppost to do

[plucky crag]

idk it was just there originally

[obsidian ginkgo]

i’m so confused

[obsidian ginkgo]

i’m pretty sure it’s not like that in the default config

but idk because i don’t use waterfall

[plucky crag]

it is it says

servers: {}

[obsidian ginkgo]

but it’s basically the same for configing it as velocity

[obsidian ginkgo]

yes that means no servers

[reef fulcrum]

that's because somebody set it to an empty section

[obsidian ginkgo]

so there’s not a null error

[plucky crag]

I was going to use velocity but it said like it doesnt support some plugins so i didnt use it

[obsidian ginkgo]

that’s not the default config

either way

i can assure you that

[plucky crag]

well i downloaded it from the official papermc site

[obsidian ginkgo]

itd show an example server

or something

[reef fulcrum]

the config ending up like that would generally indicate that somebody had a bad configuration file and it sorta like, read it and saved it back like that

[plucky crag]

Uhm

so I did it

and when i started my server

it said no servers defined

even though i saved the edited config

now its just gone?

[reef fulcrum]

I mean, all I can guess is that you misformatted it

[plucky crag]

servers:
lobby:
motd: 'ForgedMC Network Lobby'
address: ip
restricted: false

is that incorrect

im bouta just give up on a network

and stick to making regular mc servers

way to complicated

for me

[delicate phoenix]

Just a small thought, but if you run into formatting issues with .yml, https://www.yamllint.com/ is always a good site to check the file with.

[random oriole]

Is Adventure supported in Waterfall?

[reef fulcrum]

Not natively

however, you can use the adventure-platform library for waterfall/bungee

[mossy lark]

What must the motd look like so that it is in two parts and centered?
" &5Game Town [1.8-1.20] &1Minegames | &2CityBuild | &6OneBlock"

[reef fulcrum]

generally you just have a lot of spaces

and \n for a newline

unless you use a yaml multiline

[mossy lark]

no matter how i build it, it's never centered^^

^__^

[reef fulcrum]

There is no way to centre the text

All you can do is use spaces to try to align it

Noting that it will look ugly for anybody using a resource pack with custom fonts

[fervent apex]

adding space did also not work :/

[reef fulcrum]

Might need to abuse colour codes so that it doesn’t truncate it

&A &a &a &a

Otherwise, no idea

[fervent apex]

okay, i'll try that

[mossy lark]

At most, you could use a plugin that did motd to adjust accordingly.

[craggy schooner]

some random thought, if i used waterfall 1.20.1 but joined on 1.8.9 (this scenario has no viaversion, viarewind, etc.) but the backend is 1.8.9 would it let me join?

[marble quest]

waterfall supports multiple protocols, so yes you can (1.8 to 1.20.2 iirc)

[agile flower]

I am getting this error whenever i join a server

[tidal musk]

..

so it kick me when i joins the server

[hexed zodiac]

is there a 1.8 version of waterfall or do i have to scrap waterfall support for now

[reef fulcrum]

latest waterfall supports 1.8+

[hexed zodiac]

oh aight

[faint jasper]

hi when I restart a lot of servers at once and they load back up, sometimes they go out of "sync" and if I run /server survival it sends me to /server survival2 and vise versa

it only gets resynced when I restart my proxy

[reef fulcrum]

I mean

waterfall literally just has a name > address map

it has literally 0 logic that could screw that up unless you're relying on docker hostnames, in which case, DNS caching will potentially fuck you

[faint jasper]

hmm then dafuq is going onnn lmao

[true crane]

Is there a doc that list all of the perms for waterfall? and what that perm does

[reef fulcrum]

No idea if we have a perms doc page for WF, I doubt it

apparently we don't really add any perms over bungee, however

[marble kelp]

just don't send messages that early, wait untel the user is fully joined

well... the weird thing is that the message does actually get sent to the player xD

[marble kelp]

iirc velocity doesn't have that issue

[humble light]

Can I disable /op in waterfall bungee?

[marble kelp]

waterfall does not have /op

[tidal musk]

Hello I have made a Bungecord server with Waterfall. in the config I have taken the point Tablist Global_Ping only it is now so that it does not work so players on server B are not displayed in the table list on server A. what can I do?

[humble light]

I mean, I've seen in some servers, /op is not allowed in any spigot server connected to the proxy

[marble kelp]

use a permissions plugin?

[humble light]

I already use luckperms

wait, lemme send you an example.

nvm I figured it out.

[upbeat heron]

can I enable ip forwarding while still allowing direct logins to the server?

[lost river]

no. However you could create a second proxy that also uses ip forwarding.

[upbeat heron]

is that possible with velicity? I don't have any plugins setup yet so I'm going over options

I'm kind of in a transitional phase so it'd be great if ip forwarding would work while still allowing direct connections

[lost river]

no, it is the same with velocity

but velocity allows you more secure forwarding.

as is: no one else can simply set up a velocity server and forward players to your server if you use modern forwarding

@upbeat heron why do you want a transitional phase?

[upbeat heron]

I guess I don't

[stiff veldt]

hi, my players getting kicked frequently from all servers, only normal disconnect msg in the paper servers but in the bungee this comes:
[SKYBlock] - could not decode packet! io.netty.handler.codec.DecoderException: java.lang.IndexOutOfBoundsException: index: 3, length: 60 (expected: range(0, 12))

The msg changes abit each time. any ideas for what could cause this?:D

[reef fulcrum]

something is mangling packets

[stiff veldt]

like a resourcepack?

[reef fulcrum]

No, like a plugin

[stiff veldt]

in the waterfall server or in the actual paper server the person is kicked from?:D

[reef fulcrum]

can't say

[lime river]

Is there any way to assess what plugin might be using CPU resources? Its sitting at 40% and slowly increasing.

All the plugins look equal in spark so idk if there's an option I'm missing

[reef fulcrum]

There is no option, all that exists is a sampler

[lime river]

Gotta go the mod way and remove one at a time?

[reef fulcrum]

pretty much

[lime river]

Aight, ty

[iron geyser]

get too much

how can i fix this

[iron geyser]

you can help

[reef fulcrum]

no

[iron geyser]

how can fix this

every time come

[reef fulcrum]

Please stop pinging me

[iron geyser]

sorry

[reef fulcrum]

I have no interest in providing support at the moment

[iron geyser]

plases sir

[marble kelp]

make sure you test without any plugins

[craggy trout]

Kick yourself

[iron geyser]

stil not work

i find my self in all server

i not any server

but in bugeecord still pinge me

when i refresh server list then pinge me

[prisma crypt]

[20:36:14] [Server thread/INFO]: [TAB] §x§f§f§5§5§5§5ViaVersion returned unknown protocol version 764 for player .yeesub9281. This may result in plugin not working correctly for them.
[20:45:42] [Netty Worker IO Thread #0/WARN]: [/176.232.56.34:0|.RANJWOW] <-> DownstreamBridge <-> [Towny] - overflow in packet detected! A string had more data than allowed. For more information, launch Waterfall with -Dwaterfall.packet-decode-logging=true
[20:45:42] [Netty Worker IO Thread #0/INFO]: [.RANJWOW] disconnected with: The server you were previously on went down, you have been connected to a fallback server

My players who log in from their phones are being kicked out for this reason. What should I do to solve the problem?

pls help me

[craggy trout]

try without plugins and enable the debug flag

[spiral hatch]

Don't put ViaVersion on a Proxy

[prisma crypt]

Sorunun nedeninin bu olduğundan emin misiniz?

[spiral hatch]

Usually is

[devout roost]

Hi

My server dont pass the listening in and if i change the host in config it say it cant conect to the server pls help

[craggy trout]

You must ensure have the correct address, port and query stuff

[devout roost]

Its all good and it dont run

[devout roost]

Pls help

Here are some captures

Can someone see the captures and help me pls

[last yoke]

I am using normal BungeeCord and i had setup the config correctly as i saw but still i am grtting this error

enforce_secure_profile: false
server_connect_timeout: 5000
listeners:
- bind_local_address: true
  force_default_server: true
  forced_hosts:
    pvp.md-5.net: pvp
  host: 0.0.0.0:2552
  max_players: 100
  motd: '&1Another Bungee server'
  ping_passthrough: false
  priorities:
  - lobby
  proxy_protocol: false
  query_enabled: false
  query_port: 2552
  tab_list: GLOBAL_PING
  tab_size: 60
remote_ping_cache: -1
network_compression_threshold: 256
permissions:
  admin:
  - bungeecord.command.alert
  - bungeecord.command.end
  - bungeecord.command.ip
  - bungeecord.command.reload
  - bungeecord.command.kick
  default:
  - bungeecord.command.server
  - bungeecord.command.list
  - serverswitcher.switch
log_pings: true
connection_throttle_limit: 3
prevent_proxy_connections: false
timeout: 30000
player_limit: -1
ip_forward: true
groups:
  Sahil:
  - admin
remote_ping_timeout: 5000
connection_throttle: 4000
log_commands: false
stats: 98a36421-19c4-4a1c-b834-b3cb411a7dbe
online_mode: false
forge_support: false
disabled_commands:
- disabledcommandhere
servers:
  lobby:
    address: 88.99.13.119:2796
    motd: '&1Just another BungeeCord - Forced Host'
    restricted: false

[upper stirrup]

@last yoke Wait patiently, don't post other places demanding help.

[last yoke]

This channel is really inactive so i did that!

Can you help please?

[upper stirrup]

Sure. Looking closer, I see you wish to run with online-mode=false. Therefore, this is the end of your support here as we do not provide support for such servers.

Have a nice day. 🙂

[last yoke]

I saw a tutorial by a big youtuber so i did that

[upper stirrup]

Your image does not show turning it off in the waterfall settings. We do not support piracy.

[spare osprey]

I need help

https://mclo.gs/nUNbys8

new log

[upper stirrup]

bind(..) failed: Cannot assign requested address

Means the server can't start on that port. Report this to your apparently ad-filled host.

[devout roost]

pls help

[reef fulcrum]

Speak to your host for support

[devout roost]

okey

[autumn parrot]

i added a skyfactory3(ftb modpack) server to my network but i can connect to it

[reef fulcrum]

Mods and proxies is generally "glhf"

[autumn parrot]

it just says

[12:06:14 INFO]: [<user>|/<ip>:<port>] <-> ServerConnector [skyfactory3] has disconnected```

[autumn parrot]

glhf?

[reef fulcrum]

good luck, have fun

[autumn parrot]

k

ty

[reef fulcrum]

Basically, the modded server needs to be the first server that a modded client joins

the proxy being unable to reset a client means that it cannot support cross server transfers nicely

a bit of a hackjob exists for velocity in the form of a plugin which is generally more reliable and even has a client mod if you can ship your own pack to make server transfers mostly "just work" within the limitations of that

[autumn parrot]

k

[autumn parrot]

i think i found a simular mod to this for most other serverprogramms

[solar mulch]

Does anyone have a script handy that automates the downloads of the latest waterfall jar? The v2 api seems to requires specifically build numbers and does not accept 'latest'.

[soft oriole]

Hey, I've just updated my proxy and paper servers to 1.20.2. When trying to change servers using /server all it says is that the server is restarting even though the server is fully online

[stuck nacelle]

.

[winged pilot]

skinsrestorer

piracy is not supported

[worthy badger]

Hey how do I make a mod pack server that offers multiple different mod packs like complex origin or ftb using waterfall?

[wicked tusk]

that's the neat part: you don't

the only chance you have is velocity

but even then server switching is still not working well with lots of mods

[worthy badger]

How do servers like ftb do it then

[wicked tusk]

idk those servers

but you can set up forced hosts

so instead of connecting through a lobby, you connect straight to the modpack server

that's how i've got my servers set up

[worthy badger]

Can you help me tomorrow?

[wicked tusk]

you can ask your questions in #velocity-help and then others can answer too

[worthy badger]

I’ve been at this for the whole day

Oh

Ok

[wicked tusk]

use the ambassador plugin

[worthy badger]

Can I at least have your server config files

[wicked tusk]

you can, they're all on github

but they're not in a format most people understand

[worthy badger]

Ik but can I see what ones u have

Or

[wicked tusk]

https://github.com/alexstaeding/homelab/tree/master/kubernetes/minecraft

[worthy badger]

Tysm

Can we go in DMs though I have some questions

[worthy badger]

Also what format is that and why isn’t it the default

[wicked tusk]

lets just move to #velocity-help

[worthy badger]

Ok

[craggy trout]

Did you update SkinsRestorer to latest one

[distant juniper]

hello

[dry thunder]

`connection timed out` / `connection refused`

unable to connect to server
io.netty.channel.ConnectTimeoutException: connection timed out``` ```
unable to connect to server
io.netty.channel.AbstractChannel$AnnotatedConnectException: finishConnect(..) failed: Connection refused```
**These exceptions indicate that your proxy cannot connect to the address you've specified.**
There can be several reasons for this:
- The server is not running
- Your proxy is being blocked by e.g. a firewall
- Your network is set up in a way that requires you to use a different address.

If you're sure your server is running, you haven't mistyped the address and the proxy's not being blocked, try following the steps below:

If all your servers and the proxy are on the same machine

Try using localhost as the servers' address.

If you're using Pterodactyl or another panel that utilizes Docker

In general Docker containers are network isolated and will require you to use a different address to reach the machine and not the container.

In Pterodactyl's case the default address to reach the machine is 172.18.0.1 (read more here).

If you're using a host

Speak to their support. Your host should know their own network the best and help you set your proxy up.

[distant juniper]

my server when i connect send me a message

because i uploaded via version but didnt work

[reef fulcrum]

The connection failed

see your logs

[distant juniper]

on?

[reef fulcrum]

the proxy

it failed to connect to the server, so the server is basically irrelevant

[tulip surge]

message.txt by @distant juniper: https://pastes.dev/w3YwJ0bzNq

[tulip surge]

message.txt by @distant juniper: https://pastes.dev/gFn2uJfm2a

[reef fulcrum]

missing the actual connection error

[distant juniper]

how i do to resolve it?

[reef fulcrum]

all I can see i s that you have a plugin that crapped itself, but that's irrelevant

you're missing the actual error about the connection failure

all you've got there is a log from one of your plugins shitting the bed

[distant juniper]

:\

[tulip surge]

message.txt by @distant juniper: https://pastes.dev/CzfHVTNran

[reef fulcrum]

all I can tell you is that the server apparently kicked you

[distant juniper]

mh

is the problem of lobby?

[reef fulcrum]

[distant juniper]

oh i wand etered the lobby but the server kick me

:

is a problem of the lobby

[flat bolt]

Caused by: net.md_5.bungee.protocol.OverflowPacketException: Cannot receive string longer than 765 (got 1403 bytes)

[tulip surge]

message.txt by @flat bolt: https://pastes.dev/7QGePirBIY

[flat bolt]

using a kinda orthodox setup where this is the subproxy instead of the main proxy and this is when I set ip-forwarding to true and it forwards all the player info

[nimble ember]

proxy in proxy isnt really supported

[flat bolt]

well I understand but is there any way around this problem

[tulip surge]

message.txt by @sweet garden: https://pastes.dev/ss1CD6Twee

[sweet garden]

help

[reef fulcrum]

Cancelled ServerConnectEvent with no server or disconnect

your issue is down to plugins

[proud olive]

Can anyone explain to me why Scoreboard, Tag and Tablist plugins are making servers inaccessible in <1.12 using the latest Waterfall builds?

[craggy trout]

Do your plugins support latest protocol changes

[proud olive]

I don't know

Everything was working until at least build 544

So after that it is possible to see various changes to packages, scoreboard, etc. in the changelogs. And then the server stops being accessible.

[queen flower]

I need help port forwarding the waterfall proxy

[slim beacon]

waterfall updated to 1.20.3 that quick .-.

anyone tested it yet?

[viscid thicket]

Works fine for me 🙂

[queen flower]

@tough vale

[tough vale]

Don't ghost ping people.

[queen flower]

im making sure

i accidentally

did that