waterfall-help

[tired veldt]

How to fix this

[tired veldt]

Pls tell anyone how to fix it

[nocturne fog]

Server skyblock has a priority yet is not defined!

[tired veldt]

Why it is not defining

[nocturne fog]

because you didn't define it in the configuration

[tired veldt]

How to define it

[nocturne fog]

see https://www.spigotmc.org/wiki/bungeecord-configuration-guide/ - add it under servers

[tired veldt]

@nocturne fog

I will give you acsess

You pls define skyblock

[nocturne fog]

no, don't give me access to anything

[tired veldt]

Come in dm

[nocturne fog]

no, I don't give support in dms. I've told you want is necessary, you've put a server called skyblock under priorities but there is no server called skyblock defined under servers. To fix this, you must define the skyblock server under servers.

[tired veldt]

Ok

[pliant egret]

I have given myself OP on my backend servers

and how do i give myself OP on the proxy itself

as running op <username>

gives me command not found

[tired veldt]

Now showing sub map is null

@nocturne fog this showing

I added skyblock under servers

And skyblock showing null

[pliant egret]

Interestingly i found a workaround give myself perms via luckpermsbungeecord

did it via server console

[tired veldt]

Pls tell anyone how to fix it

[delicate phoenix]

Yes, there is no such thing as "op" with mc proxies.

[atomic lantern]

so having some issues with bungeecord/waterfall.

I have 5 servers connected to my bungeecord through lobby perfectly fine, but the 6th will not connect. i have no ideas i just get that server does not exist or io.netty.channel.abstractchannel$annotatedconnectexception

[strong mantle]

Does someone have a recommended plug-in which tracks the amount of total joins from each subdomain for Bungeecord?

[tired veldt]

Pls tell anyone

[blazing ocean]

Your config has something wrong in it

Try pasting it into a yaml parser

[warped kestrel]

can i use hex color codes in Waterfall Modt?

[marble kelp]

if so then only in the legacy-compatible format, don't think the proxy motd supports components properly

use a plugin like ServerListPlus if you want a nice config

[stoic charm]

hey

[reef fulcrum]

sending raw commands? no

there is the plugin channel for some stuff, but, er

oh

am not sure of anything for that

yea, am not aware of any REST style APIs

most people just implement all they need there

[marble kelp]

I use https://github.com/Minebench/Minequery/ but only for internal use and I haven't added real encryption yet

[stoic charm]

i am curently running waterfall 1.18 with a 1.17.1 paper server, However, when i log in using a 1.18 launcher is says Outdated server! im still running 1.17.1

[reef fulcrum]

if you're running a 1.17.1 server you need a 1.17.1 client

[stoic charm]

oh.

[tidal musk]

or

viaversion/protocolsupport

[tidal musk]

Hey, anyone have an idea of what may cause server switching to be "slow" especially when switching to my spawn

basically when switching to my spawn the client take up to 3 goods seconds to be alive again

my proxy and the spigot servers are on the same machine and I use an external HAProxy reverse proxy to my main proxy

network_compression_threshold on the bungee is currently on 256 and on all spigot servers on -1

[tough vale]

1.18 with paper backend servers?

[tidal musk]

Right

[tough vale]

Yeah that's most likely a result of the patch to respect client view distance not being in paper 1.18 yet

[tidal musk]

So for the moment reducing my spawn's render distance would likely improve the issue?

Definitely improved it, thanks!

[frail cedar]

[20:28:59 ERROR]: [ItzDepstrCZ|/ip:62544] <-> ServerConnector [lobby] - encountered exception: net.md_5.bungee.util.QuietException: Server is online mode!

all servers on online mode

i own origo

[gritty umbra]

Your spigot servers need to be in offline mode

[prisma whale]

You would only enable compression on the waterfall proxy, not on the backend servers then

[supple cypress]

Yes on localhost use always -1

[urban schooner]

ServerTap works for paper but not sure ab waterfall

[marble kelp]

https://github.com/Phoenix616/connectorPlugin

oh wait, that's the same guy that wanted external commands, not from another server

(but I gueess technically one could use my plugin standalone I should add a program for that at some point lol)

[meager dagger]

I've updated the issue as asked:
#waterfall-github message

The only thing that is different now, the bungeecord also do not work fine.

How did anyone even test it on 1.18.1 if it don't work?

Fuck me!
the damn AuthMeBungee was in waterfall plugins folder

I've deleted that not configured plugin, and all works fine.

I feel so dumb guys, sorry.
Closed

[weak wind]

why are they talking with a webhook

[blazing ocean]

IRC bridge

[tough island]

Magic!

[tidal musk]

what should i do to switch to waterfall from bungee?

[supple cypress]

Nothing

Just replace the jar

[tidal musk]

would there be any issues regarding plugins or all plugins fully optimized for waterfall?

[supple cypress]

It's the same thing normally

Waterfall have the same structure and same api

[tidal musk]

but velocity different right? i see some plugins support bungee but not velocity etc

[supple cypress]

Yes Velocity is totally different

Waterfall ~ Bungeecord
Paper ~ Spigot
Every spigot plugin will work on paper and every bungeecord plugin will work on waterfall

[tidal musk]

thank you bro 🙂

[ionic kelp]

is waterfall 1.18.1 coming out?

[severe badger]

It already is

[ionic kelp]

oh

wheres the link?

[wise kite]

https://papermc.io/downloads

[severe badger]

There's a Waterfall tab

[ionic kelp]

waterfall downloads only go to 1.18 on that page?

[severe badger]

No

[ionic kelp]

#475 is 1.18

tho

see

[severe badger]

The builds are version agnostic

[ionic kelp]

so wheres 1.18.1 waterfall?

[severe badger]

Right there

[ionic kelp]

didnt work mate

says its 1.18

thought it was supposed to be 1.18.1?

[severe badger]

Like I said, the builds are version agnostic

[ionic kelp]

oh

i get you

i think

[severe badger]

1.18 is the newest minor version out, yet that doesn't mean all previous versions have been ripped out from functionality

[ionic kelp]

anyone know where this comes from?

i want to change the colour of it

[supple cypress]

You can't

except if you fork waterfall

[prisma whale]

you could try opening up waterfall.jar and pulling out messages.yml

[rare forum]

[15:18:01 WARN]: [/123.123.123.123:53080] <-> InitialHandler - corrupted frame: A packet could not be decoded because it was too large. For more information, launch Waterfall with -Dwaterfall.packet-decode-logging=true

Keeps happening by the same ip every 5 seconds

[royal hawk]

My guess is that this is a unwanted service trying to connect to your instance. You can safely ignore it

[marble kelp]

acould just block it in your firewall

[royal hawk]

someone still has to explain to me why md_5 considers requests via http protocol valid for this listener

[tidal musk]

Hey there, I'm looking for some help.

Before 1.18 my plugin worked as intended but after updating to latest (no code changes at all) it won't run my scheduled runnables and write to files.

Code: https://pastebin.com/AcW8ciC3

Now it wont print any of the messages or check the file.

[marble kelp]

check that you have acctually called the code which registers the scheduler from your onEnable I guess?

[tidal musk]

other functions are called in the public void run(), so it does run every minute

it's just that specific function that wont work

[royal hawk]

if you don’t get an exception it’s either not run or it is run and you have an issue elsewhere

[winged pilot]

is there any reload command where like if i add another backend server or something the change will take effect without rebooting and kicking everyone?

[marble kelp]

/greload but using that is not suggested and might cause random errors in plugins

*recommended

[tidal musk]

Hello, do you guys know what I could be missing? https://gyazo.com/090e8098a335901ccdcf49c2aa68662f (All my backends are working, I'm just not sure what I could be missing. But I am using latest version of waterfall just now which I feel like I am missing another update or something).

[reef fulcrum]

stuff like that is generally down to network setup, etc

if ptero, read their documentation

[tidal musk]

ah I see, ptero is up for me tho.

Network shouldn't be down either. All backends are open and proxy. I'm not sure what I've done wrong. I've checked my config settings and nothing wrong with it.

[reef fulcrum]

no idea what you mean by ptero is up

you need to configure stuff to connect over the ptero network interface generally

[tidal musk]

ptero panel is up and running. I've updated it to latest versions.

[reef fulcrum]

which is irrelevant

read the documentation

[tidal musk]

ah which part of the documentation?

[reef fulcrum]

idk, they have a set of pages specifically for setting up a minecraft network

[tidal musk]

I feel like it's something wrong with the configuration of mine but can't figure out what.

I mean it was working fine until I updated to latest waterfall and configuration for it. But I didn't really touch anything beside adding a backend server.

[reef fulcrum]

I can't comment on a config I can't see

all that error tells us is that it failed to connect to the server

[tidal musk]

I can show you my configuration probably give me a sec. I just can't show the private info part.

[reef fulcrum]

which if you mean the IPs by private info, then the config is useless to look at

all that tells us is that it failed to connect

unless you have logs showing much more than the generic failure to connect, I can't really say much other than diagnose your network

[tidal musk]

yeah let me try to give you that.

[reef fulcrum]

which, given you're using ptero, i don't provide support for that

[tidal musk]

Here's my log - https://pastebin.com/MfUZjg4k

[reef fulcrum]

yea, nothing useful in there

just the general "failed to connect to the server"

[tidal musk]

hmm

https://pastebin.com/BkwB70x3 have a look at this (I have it set after read it burns). It's the config.

[reef fulcrum]

127.0.0.1

yea, don't do that

that doesn't work, as I've said

[tidal musk]

oh well it worked for my other servers. hmm

ah wait

nvm, failed to figure out the situation.

[tidal musk]

Quick question if I am using latest version of waterfall I need to have java to be updated to java 17 right?

[reef fulcrum]

no

[tidal musk]

ah

I have no luck then. I've tried everything.

[valid kraken]

Do you have more used plugins for waterfall? I never used this system and I'm half lost

I want to put together a good lobby

[reef fulcrum]

as. Isaid, you shouldn't be using localhost when on docker

read pteros documentation

you need to bind and connect to all the servers over the ptero network interface

[tidal musk]

ah so I need to use instead for pteros?

172.18.0.1

[reef fulcrum]

if that's the ip of the ptero interface, yes

[tidal musk]

ah I see, I'll ask ptero discord to see to double check. Ty for the help.

[round pine]

Kinda unrelated, but what are the best ways to reduce latency for players aside from relocating the server? I host on fiber but some overseas users have issues with latency. Are there any services or anything which can help to reduce latency such as VPNs or something?

[reef fulcrum]

outside of private networks like argo, no

[round pine]

Alright

[livid tiger]

Does anyone know how to solve this?, whenever I update a player's group with LuckPerms, either ingame or in the console, every player gets disconnected from the server; both lobby and the server doesn't issue any errors but I get this on the waterfall bungeecord:

[tulip surge]

message.txt by @livid tiger: https://pastes.dev/oLt5v8zHD3

[livid tiger]

Mitrales is another playername

[reef fulcrum]

ou have a plugin across two servers generating duplicate team names

[livid tiger]

what could the plugin be?

I have luckperms on both servers, could that be the issue?

should I put luckperms on the waterfall plugin folder?

[reef fulcrum]

luckperms provides permissions

[royal hawk]

A plugin that’s either displaying something in the tab-list or on the sidebar I’m guessing

[reef fulcrum]

it doesn't create teams

[royal hawk]

I’m guessing a plugin is updating a role name

And then trying to display it

[livid tiger]

maybe TAB?

[royal hawk]

Remove it and see

[livid tiger]

alr

[livid tiger]

So remove the plugin from one of either servers right?

[livid tiger]

yeah its tab

I know its him but idk how to solve it 😩

[frail kayak]

how do I set this up?

trying to run in minecraft in docker containers through docker compose but its failing to connect to other servers

[cosmic fable]

Is forge minecraft version 1.18.1 supported? As every time I try connecting to my forge 1.18.1 server it always comes back saying "unexpected packet received during server login"

[reef fulcrum]

forge 1.13+ is not supported

theres only one proxy which supports modern forge and it requires a mod on the client to work

[cosmic fable]

which proxy is that?

[reef fulcrum]

lightsomething

it's mentioned on the issue tracker in the modern forge ticket

[cosmic fable]

Thanks. I've been looking for something difinitive all day. I'll take a look at that ticket

[hearty shale]

does waterfall support 1.18?

[nocturne fog]

Yes

[edgy silo]

Does waterfall offer a way of seamless server transition? 🤔

[reef fulcrum]

no

[edgy silo]

ah okay

any chance you know of some way this could be done?

[reef fulcrum]

theres been plugins for it before

but, like

is a pita

breaks many things

causes many issues

[tough vale]

I mean technically if you went heavy with packets it could probably be done, but ^

[reef fulcrum]

(and is 100% unsupported

[edgy silo]

I read an old thread regarding it for bungee like 5-6 years ago or something

and I'm working with a lot of armor stands on a survival portion of server which I'd love to off the load by splitting the map into large chunks of lets say 1000x1000

but I'd love to make it seamless if possible

[royal hawk]

See the conversation in #velocity-dev from yesterday to understand why this is hard to do correctly and what you’re gonna face

Start from here #velocity-dev message

[edgy silo]

thanks was just reading through 😄

[magic pagoda]

hi
is there any way that waterfall kill the main proccess automaticly ?

it's just say Killed in the console and the server go off

[reef fulcrum]

No

that's pretty much 100% gonna be on the host machine

i.e. running out of ram

[magic pagoda]

running out of ram(memory) kill the proccess ?

[reef fulcrum]

if your machine runs out of ram it's generally gonna kill something so it doesn't crash

if you're using a linux machine, check dmesg

[magic pagoda]

papers don't work this way

[reef fulcrum]

Literally nothing to do with paper here

its to do with the machine you're hosting on

[magic pagoda]

i ment when paper run out of memory it's crash and don't get killed this fast

[reef fulcrum]

paper doesn't kill anything when it runs out of memory

waterfall doesn't kill anything when it runs out of memory

[magic pagoda]

aow you mean the whole machin

[reef fulcrum]

I only said "your machine" several times

[magic pagoda]

i was want to add more memory to xms and xmx

thanks

[reef fulcrum]

you can't add more Xmx

Xmx is irrelevant here, outside of you giving it too much potentially

[magic pagoda]

👍

it's possible to use some bot attacks to force to kill the server ?

[reef fulcrum]

you're the one with logs here

I'm not going to piss in the wind, glhf

[magic pagoda]

because it was under attack by some dang bot attacks and in dmesg

[2607762.136449] Out of memory: Kill process 11281 (java) score 478 or sacrifice child
[2607762.137011] Killed process 11281 (java) total-vm:9752836kB, anon-rss:5873560kB, file-rss:0kB, shmem-rss:0kB```

my total vm is 5 how the hell it's get to 10

[royal hawk]

How much ram do you have on your machine?

Fair reminder that you can never allocate more ram total than is available or goes over what the system needs. That will lead to this

[magic pagoda]

it's 12gb
i give 3 to waterfall
2 to lobby and 5 to survival

[royal hawk]

My guess is that you allocated more than you have

What does your system use in idle?

You have to account for that

[reef fulcrum]

so

that's 2G of ram

for caches

for the OS

for java itself

[magic pagoda]

need more ?

[reef fulcrum]

literally 0 surprise that especially in a case where native buffers are gonna be thrown around that your OS ran out of ram

[royal hawk]

You have too much allocated to your java applications essentially

The system has too little to work with so it terminates a low priority process to continue working

[magic pagoda]

i don't use gmesg so often so i don't understand it well
but i try to know more about it and what about information

[reef fulcrum]

dmesg is the kernel logs

[magic pagoda]

this is possible ?

[reef fulcrum]

it tells you, the system is out of memory, so it picked a process with a high "score", and said, die, or free up some memory

I literally already said...

I literally explained it...

and I'm out

[royal hawk]

It’s possible because you have too little ram for your system to work

[magic pagoda]

hommm
yeah

[royal hawk]

Always calculate your system overhead under load + jvm assigned Xmx (+15-20%)
And then leave a 10% gap

[magic pagoda]

i owe you one

one last question
linux need restarts to breath like windows or not ?
(i think i know that windows use more memory when it's running all the day and need a restart to free that up)

[magic pagoda]

my htop is this now for memory
this is fine i guess ?
this is ~80% in use
20% is enough

[tough island]

Not necessarily... I only do reboots when required....

Just don't use applications with memory leaks.

[rugged spire]

sorry for the delay in responding, I was without internet so the version here is forge 1.12.2 pixelmon 8.3.3

hence I don't know my warterfall has no plugin or mod so it's pure, the possibility to install or forge inside it ???

[marble kelp]

well 1.12.2 is cearly below 1.13 so it might work

[reef fulcrum]

mods and proxies are generally hit and miss

something that is mangling the connection process

as for what 🤷‍♂️

[rugged spire]

it always gives me this error

the main server is with a sponge and I already allowed it within the config

[reef fulcrum]

as I said

something is screwing up the connection process

I'm not a magician so I can't tell you what

[rugged spire]

😅

[reef fulcrum]

generally you just gotta start removing mods until you find out which one is causing it or something

[rugged spire]

so then I think the best thing to do is to remake the server base, just leave the main mods and go testing?

[marble kelp]

I mean the error smells like a mod issue, not a proxy one but ¯_(ツ)_/¯

make sure you have disabled entity metadata rewriting I guess

(also you might have more luck using Velocity instead of a bungee-based proxy)

[rugged spire]

then the problem arises I tried to initialize the velocity but it doesn't open I already reinstalled java and etc... and none of it opens

[marble kelp]

I mean if you have issues starting velocity then I suggest asking in its help channels

[rugged spire]

I'll try to do this from the server testing the mods first then if it doesn't work I'll try the velocity
I really need the lobby with username and password

[marble kelp]

also depending on the mods you need to make sure that all servers they connect to have all the mods required by the pack

also some mods don't like server switches so connecting directly to the modded server would be prefered

[royal hawk]

This sounds like the handshake relay bug

some mods re-request a handshake

it should work on velocity

well if you have advertise-forge enabled in the config

[cunning spindle]

Hey this is more a general consulting question but, I'm currently using Paper w/ Multiverse running 5 worlds, Overworld, End / Nether, Creative, and Resource. I'm not really hitting the kind of CPU utilization (only 15-20%, on an I7-8809g) I'd like to see, and despite my server TPS being a solid 20 and player latency being a minimum, I'm still getting reports of lag, especially around chunk loading in a pre-genned world. I was wondering if this was just a limitation of Multiverse, some kind of misconfiguration on my end, or whether pursuing a bungeecoord / waterfall setup would be beneficial or warranted. I spoke with them and I should be seeing higher utilization, but it just isn't happening, so I'm wonder if waterfall would alleviate that, and what I should look into should i deicide to migrate. Any advice is greatly appreciated! 😄

[rugged spire]

as well advertforge activated where is this config ??

[marble kelp]

Nathan: I suggest finding the actual source of those "lags" first

best to look at timings and if you need help with them post them in #paper-help

but yes, if you run any kind of creative/multi-inventory setup then you should really move to a multiserver setup anyways in order to secure your server

(splitting up survival worlds into multiple servers can be more difficult if you want the same inventories and even stuff like map items/nether portals/other plugins to work cross server)

also if you want to look into building a proxy network then it's usually suggested to go with Velocity unless Bungee-only plugins are necessary (and you can't get them to work with a plugin like Snap 👀 )

[cunning spindle]

Thanks phoenix! I'll look into that! Thanks for the advice!

[royal hawk]

Bungee has a forge setting, but that one doesn’t exactly have a lot of merit. I was talking about velocity

[rugged spire]

tended so i'll try the velocity here thanks

[craggy solar]

Whats the differnace between
player_limit: 50

and

max_players: 50 In the listener? Is the player_limit a global hard limit?

I have server full messages and cant seem to bypass them as OP

[reef fulcrum]

proxy doesn't know about op

[royal hawk]

player limit is globally, max players is per listener

[craggy solar]

Right, so if I were to have a server limit, do I set that in my server.proeprties and then just have higher values for bungee to let staff bypass the server.prperties limit via a permission?

I think what ive done, is set everything at 50

player limit, max players and in server properties

[royal hawk]

I’m not sure bungee has bypass permissions

[craggy solar]

yeah, so im guessing ill need to bump the values and restart bungee

unless greload works

best to just reload

[final trail]

ok so is it possible to disable tab completer using waterfall, so only waterfall side commands (bungee commands) will NOT be shown if player does /[TAB-BUTTON]

i'm asking this specifically if player is using MC client higher than 1.13

[reef fulcrum]

there are plugins for that, yes

[final trail]

from what i saw, only viaversion can do this

[reef fulcrum]

there are plugins on the forums, etc, which let you disable tab completion of stuff

[final trail]

when you disable 1.13+ auto complete setting in config, it won't show neither bungee and spigot tab completion commands

[final trail]

even chatcontrol can't do this

and it costs 30eur (lol)

[reef fulcrum]

waterfall itself has an event that lets plugins remove suggestions when they'r ebeing defined

idk what to tell you

I added an event in waterfall to do it like, years ago

[final trail]

and it only works for 1.12 and lower?

[reef fulcrum]

there are plugins, I even wrote one posted on the old forums to let you do it

[final trail]

ok may i ask you to send me a link if i don't find it?

[reef fulcrum]

the event was literally added for the 1.13 command structure logic.

No

I don't have a link to it, it's on the old forums or something

[reef fulcrum]

wrong channel

[exotic remnant]

How do I make it so users can only join through the waterfall server, and not directly to a Paper server?

[reef fulcrum]

firewall

I mean, like

[exotic remnant]

I swear it used to be somewhere in the config

I just forgot where

[reef fulcrum]

by default, you can't connect to a proper configured server for the proxy

if you mean bungeecord mode, that's in spigot.yml, but, is not a security mechanism

[exotic remnant]

Ah okay

I haven't used waterfall in a while so I forgot

[royal hawk]

If you can’t firewall it off then use bungeeguard

That’s a plugin which adds a kind of password to the system

The firewall/whitelist plugins are no the way to go

[exotic remnant]

Thank you

[dawn lark]

What does "this current session mean?" Does it mean until the proxy stops? Does it mean until the player logs off?

[reef fulcrum]

until they log off

[dawn lark]

Got it, thank you!

[flint shell]

hi, I tried to delete all the entries from the location.yml in waterfall. But it always regenerates the all the entries back.

is there anyway to remove location.yml and let it regenerate a brand new file?

[reef fulcrum]

stop the proxy then delete it?

[flint shell]

oh

let me try

oh my, how simple

thank you

[weak wind]

Hi why does it take ages to download waterfall from official paper mc website

[nocturne fog]

It doesn't here, try restarting it I guess

[thin minnow]

Hey i have an issue

I'm getting this error: could not connet to a default or fallback server, please try again later: server not reachable (timeout). Offline? Incorrectly configured address/port/firewall?

Like I can join my lobby and survival server but not my earth and skyblock

[tough vale]

Read the error.

[thin minnow]

look man ive tried everything

over a day of stress trying to fix this

i have tried litterly everything

[marble kelp]

.paste your full log file as well as your config

[thin minnow]

bungeecord log file?

[marble kelp]

yes

[thin minnow]

ok

not going to give my ip from the logs

my bungeecord log

https://paste.dinonode.com/yayeyoduma.apache

[tough vale]

skinrestorer
No support for pirates 🏴‍☠️

[thin minnow]

https://paste.dinonode.com/naciferura.yaml

[thin minnow]

what?

i dont use cracked lmao

check my config

i use it so players can use any skin

[tough vale]

fair enough I suppose

[reef fulcrum]

server not reachable (timeout). Offline? Incorrectly configured address/port/firewall?

basically, those

or, network misconfig

if using a hosting provider, speak to them

[thin minnow]

i did

no one knows the cause

[reef fulcrum]

it's a generic connection failure

pretty much always down to network config

best advice would be use a host who knows what they're doing 🤷‍♂️

this is basically out of the bounds of what we can offer support with if you're using a hosting provider

[thin minnow]

oh well

thanks anyway

[hot wave]

https://bin.birdflop.com/yocebavora

¿?

sessionserver.mojang.com not exist

i dont understand

It is what I said once that the "dns resolver" does not work on my server I do not know why and here is the error that I do not know what it is due to.

[reef fulcrum]

try disabling the netty dns resolver in waterfall.yml

beyond that, if using a hosting provider, speak to them

[hot wave]

I disabled it some time ago and since then it works fine, but I don't know why and I have OVH.

I have tried to enable it before but the error persists.

[reef fulcrum]

does DNS work on your server?

are you using ptero/docker?

[royal hawk]

Linux DNS setups can be a pain

That’s especially the case if you have a complex container network

[hot wave]

ye, i use ptero and docker

sorry for the ping

But I never had problems of this kind, I guess that when upgrading to a new version this started to happen, because the configuration is basically default.

[frail kayak]

is there a way I can make waterfall use online mode for backend? or how would I go about htat?

[tough vale]

Nope. When using waterfall, the only way to connect to the backends should be through the proxy.

[frail kayak]

okay... anything I should do if porting a previously public facing server to waterfall?

backend server is purpur

new to all of this and I have googled and am finding jack shit about any of this stuff, and the stuff I have found, is obscure and makes no sense...

like how do I handle OP? how do I handle individual backend server plugins? is there a plugin to make Xearos minimap/worldmap and bobby function properly?

[tough vale]

OP: assuming you have a permissions plugin, no one should ever have op. If you're at the point of a network, you should have a permissions plugin. If you don't, LuckPerms.
Backend plugins: Varies by plugin. Some plugins have explicit bungee support, in which case follow their instructions (usually consists of either connecting all instances of the plugin to a DB or installing another plugin on the proxy). Other plugins don't support bungee explicitly, but work fine on the backends only i.e. Essentials.
Client mods: easy answer is no. Hard answer is it depends on the mod, some might support it but require something on the server, but often that something is a mod, not a plugin, and it quickly becomes a mess. Talk to the mods' support, they'll be able to give you a better answer

[weak wind]

even after restart
I get 5 mbps speed on downloading waterfall jar from other unofficial websites but only 5-6 kbps from official paper mc website

[reef fulcrum]

all downloads are piped through cloudflare

there ain't much we can say or do

[fluid hearth]

Is there any plugin that can allow players to switch servers by clicking on signs?

[marble kelp]

@fluid hearth those plugins are usually named something along the lines of "server signs"

[fluid hearth]

My current one only works with bungeecord and doesn't work with waterfall

The others I can find don't work on 1.17

[marble kelp]

that's not really possible. all bungee-compatibile plugins should work with waterfall unless they use some very strange and hacky internal bungee stuff

[pale crag]

Should my waterfall instance be on the same Java version as the paper server? My players are getting the java.net.SocketException: Connection reset error

[marble kelp]

Jakubk15: the java version has nothign to do with that but you should be always at least be on latest LTS

-one of the "be" s

[pale crag]

Everything is on Java 17, including the 1.17.1 paper, because JRE 16 is not included anymore in Linux uprising repo

[marble kelp]

well 16 isn't LTS so yeah...

[pale crag]

But should upgrading to 1.18.1 help?

[marble kelp]

I doubt it, that error is due to networking issues

[pale crag]

Is it client side fault or server? I have a hetzner dedi with tcpshield

I doubt it is TCPShield fault, this was happening also before it was installed

[marble kelp]

depends on what disconnects first, the connection of the player to tcpshield? the connection between the proxy and tcpshield? the connection between the proxy and the server?

[reef fulcrum]

connection reset just means that the connection was closed improperly

[pale crag]

That can be also because my server is using a modified version (fork) of waterfall

[reef fulcrum]

do note that that is how the tcpshield plugin will cause connections to close if connections aren't going through their services, iirc

but, basically, that's network level stuff, and, especially when you're using a 3rd party service in the middle, you probs wanna check their support, etc

[marble kelp]

also of course if you don't use waterfall you should contact support of whatever it is you use lol

[pale crag]

idk if it has something to do with it, but the developer requires to add starting argument --add-opens java.base/java.lang=ALL-UNNAMED to launch it on Java 17

[reef fulcrum]

So

you're not even using our software

glhf

[pale crag]

Ok, I will contact the dev, thanks for help! ❤️

[spiral scarab]

Would it be possible to have the server ask for a different texture pack in different servers under a proxy?

[marble kelp]

packs set in server.properties will work just like they normally would: as soon as a player joins they get the pack sent

[thin minnow]

Is that actually a bot

[tough vale]

Nope, IRC

[tidal musk]

Hello, quick question. Is there a way to see if IP forwarding is functioning properly with the paper servers?

[royal hawk]

Easiest way is to check the player UUID on the server usually. That or disable tab-list-rewriting in the waterfall.yml and see if skins still work. If yes then it’s working correctly

[tidal musk]

awesome! thank you

How does waterfall do being run in a docker container? should I run it in host mode to prevent passing the docker server's actual internal IP?

[tidal musk]

OK, so IP forwarding looks to be working just fine but for some reason venturechat config is not populating with any options other than the ipforwarding error 😐

[queen stratus]

How do I setup fallback? Because I don't see any options for it in the config.

[reef fulcrum]

it's the priorities list now

[queen stratus]

Yeah I have it like:

• lobby
• survival
• test
  and if test goes down I don't get sent to lobby

It says "can't connect to fallback"

[reef fulcrum]

yea, you'd need a plugin for that

that only deals with new connections

[queen stratus]

oh ok

is there any good plugins you recommend for that?

[primal wagon]

Hello I have a little problem with waterfall I barely have 20 players but after a while it consumes a lot of CPU. Do you have any idea it's binding.

Thank you

[reef fulcrum]

make sure natives worked, see startup logs for epoll

[royal hawk]

are you running any heavy plugins like ViaVersion or Geyser+Floodgate on that proxy?

[reef fulcrum]

beyond that, attach profiler, see GC stats, etc

[primal wagon]

no..

[royal hawk]

well then, go with what Cat said

[primal wagon]

sorry if we have a viaversion

here are the plugins on the proxy

[lunar dragon]

[15:27:22 ERROR]: Exception in thread "main" java.lang.IllegalArgumentException: Invalid host/address: 172.18.0.1::25573
[15:27:22 ERROR]: at net.md_5.bungee.Util.getAddr(Util.java:58)
[15:27:22 ERROR]: at net.md_5.bungee.conf.YamlConfig.getServers(YamlConfig.java:232)
[15:27:22 ERROR]: at net.md_5.bungee.conf.Configuration.load(Configuration.java:113)
[15:27:22 ERROR]: at io.github.waterfallmc.waterfall.conf.WaterfallConfiguration.load(WaterfallConfiguration.java:67)
[15:27:22 ERROR]: at net.md_5.bungee.BungeeCord.start(BungeeCord.java:279)
[15:27:22 ERROR]: at net.md_5.bungee.BungeeCordLauncher.main(BungeeCordLauncher.java:67)
[15:27:22 ERROR]: at net.md_5.bungee.Bootstrap.main(Bootstrap.java:15)

can you help me?

[supple cypress]

BungeeSK 😳

[supple cypress]

There is two ":" ? In the address ?

[lunar dragon]

oh...

[mystic smelt]

Im not sure exactly what is causing it but I am having trouble with commands on the proxy end, All normal server commands works fine but any commands going to the proxy come back with Unknown or incomplete command see below for error, It was working yesterday and I don't know what I changed, I have bungeecord.* in luckperms and also the permission setting in the config of that. So im a bit confused.

[marble kelp]

do you have any plugins installed?

[mystic smelt]

I have Citizens, Clearlag, CommandNPC, DeluxeMenus, EssentialsX(And Chat), Holographic Displays, HubBasics, Litebans, Luckperms, PlaceholderAPI, plugmanx, punishmentGUI, simplescore, supervanish, TAB, vault, all 3 via version plugins, worldedit, worldguard.

[marble kelp]

on the proxy

[mystic smelt]

Oh my bad.

[marble kelp]

also wtf are you using clearlag for, remove that lol

[mystic smelt]

I only have Luckperms Litebans and HubBasics on the proxy

[marble kelp]

(and I hope you don't enable/disable/reload plugins with plugman...)

[mystic smelt]

Yea I probably will

I also have the latest version of waterfall.

[marble kelp]

make sure you have the same luckperms database on the proxy and game servers or that you are granting permissions on the proxy (with lpb)

also depending on your LuckPerms config you might use the bungeecord server context or enable global permissions in your bungee luckperms config (or something along that line, contact LP support for more help on that)

[mystic smelt]

Ok ill will try that, thank you

[formal citrus]

Hey there are rcon support on bungeecord?

[reef fulcrum]

natively, no, but there was a plugin, I thiiiink

[naive sky]

A lot of players on my server join and they fall through the world into the void, and then afterward, get disconnected, this is what the proxy console says:

[WARNING] [(player)] disconnected with: [{"text":"ReadTimeoutExecption:null"}]

[WARNING] [(ip)] <-> DownstreamBridge - read timed out

[WARNING] [(Username)] <-> InitialHandler - IOException: disconnected with: The server you were previously on went down, you have been connected to a fallback server.

The server is online and has been online for a while now, and there are players online, but some players have this issue and can't join because of this.

[hazy compass]

I'm superdumb, cant find the command that lets me forcemove players from one backend to another, anyone know what it is?

[royal hawk]

where are you trying to do that from?

[hazy compass]

bungeeconsole?

or waterfall in this case i guess

[royal hawk]

then it would be /send
that command is a module and if it doesn’t work then you probably disabled or removed that module

or you’re running a custom build

[hazy compass]

oh i knew i was dumb, i kept trying /server servername playername

[frail kayak]

can waterfall handle a modded client and server?

[marble kelp]

depends on what kind of mods/version

[wintry cobalt]

I probably know the answer to this but want to get clarification. Is it possible to access a server locally that is connected through the proxy? IE connect to it via LAN IP instead of URL.

Reason I am asking is the server is hosted locally and I would like to access from its LAN IP instead of through the WAN IP to bypass some occasional minor network lag. But when I try I am presented with this message, so not sure if I am actually able to.

[marble kelp]

host a proxy locally

[wintry cobalt]

It is hosted locally

[marble kelp]

also I really hope that port is not accessible publically from anything but your proxy machine

[wintry cobalt]

It's not, they only port exposed is 25565 and that is going to the proxy.

[reef fulcrum]

you can't connect to the backend servers without connecting through the proxy or having a client mod which pretends that's happening

[wintry cobalt]

Ok thanks for confirming that. That is what I figured but wanted to make sure there wasn't something I was missing. If anything I can just connect to the LAN IP of the proxy and go from there.

[wintry cobalt]

I do have a legit issue that has me scratching my head. I have two domains. For example purposes we will call it domain1 and domain2

Domain1 is our mini game server and points to the Waterfall proxy.

Domain2 points to the Waterfall proxy as well, but it is a forced host and setup as such in the config

Both of these use to share they same public WAN IP but now point to different public IP that is forwarded to the proxy. The issue is now domain1 which is the mini game server and the proxy is pointing all the traffic to domain2 and essentially ignoring the priorities list. All I did was update IP and everything is broke. Any reason this is happening?

[wintry cobalt]

For some reason Waterfall thinks my forced hosts is the primary server and it is not. It is not even in the priorities list. In fact I have it set as restricted because it's a private whitelisted server. If I turn that server completely off then it connects as it should to the hubs. What would cause this to happen?

[mystic smelt]

Im still having the same issue from yesterday, I have done some tests and it doesn't seem luckperms is causing the issue as I removed it on all servers and still have the same issue. So Im not sure what is causing it.

[serene shadow]

I have the following problem when the player switches from the lobby to the other server, he falls into the mix, the graphics are not loaded then?

1.18.1

.#169

[reef fulcrum]

falling through the world is a long standing vanilla client issues

connecting through the proxy doesn't help and with some people can add non-negligable amounts of latency which god knows how it does that, maybe look towards velocity

forced hosts are fun, it should rely on whatever the hostname that the client sent, but, that's not always perfect, nor is it sane to debug either as theres no logger output for that stuff

[wintry cobalt]

If you're on build 169 I would start by upgrading your build if you're running 1.18.1. You are many many builds behind as the recent is 475

[tough vale]

169's the latest paper version, I bet they mean that

[wintry cobalt]

Ah, yeah, I just realized that after I sent it.

[serene shadow]

Bungeecord is up to date I meant the lobby which has version 1.18.1#169

Can I do a live stream so you can take a look at the problem? I'm bad at explaining^^

[reef fulcrum]

We know what the issue is

it's mostly a vanilla client issue, as I literally already said

as well as apparently bungee in general can make it worse, as I already said

[thin minnow]

I recently moved one of my servers to bungeecord or waterfall network and now all player data was reset, how can I get the player data bacK/

?

[reef fulcrum]

for the general wider issue, that's what leafs experimental builds are for on that PR

for the bungee specific one, idk, try velocity

[wintry cobalt]

It has worked fine until I moved one to a different domain and then updated the public IP.

play.domain1.com is the primary URL for our mini game server. That is connecting to our private whitelisted SMP server by default. private.domain1.com is the forced host that is pointed at the same public IP for the proxy. It's very strange.

[reef fulcrum]

don't mention

I can't see your setup so best I can do is make educated guesses towards what the issue is

[serene shadow]

what do you mean by speed am not a pro

[wintry cobalt]

I didn't mention, I replied.

[reef fulcrum]

which mentions by default

[thin minnow]

anyone can help me?

[reef fulcrum]

data loss when migrating towards bungee is generally misconfiguration

make sure that the proxies online mode matches what the servers was and that you setup ip forwarding properly, make sure to read bungees install guide

[thin minnow]

do i send my config file for bungee?

because im lost

[wintry cobalt]

Replying and mentioning are not the same.

[tidal musk]

it gives the same ping highlight

[wintry cobalt]

I replied because it brings context to the topic as it was shuffled

[tidal musk]

(hit the @ in the bottom right corner to disable the ping)

[wintry cobalt]

play.domain1.com points to XXX.XXX.154.209 which points to the proxy.

private.domain2.com points to XXX.XXX.154.209 which also points to the proxy, but it is setup as a force host in Waterfall configs.

Only thing that changed is I moved the proxy to a new public static IP and off the WAN IP. Also private.domain2.com isn't sharing the same domain as play.domain1.com as it was before. private has a new domain name, but the A record points to XXX.XXX.154.209 which is the proxy.

The priorities list is being completely ignored and causing the forced host to be the default server when you connect with play.domain1.com which shouldn't be happening. It should be taking people to the hubs.

[warped kestrel]

[14:16:56 WARN] [LuckPerms]: LuckPerms already has data for player 'BoomblasterDani' - but this data is stored under a different UUID. [14:16:56 WARN] [LuckPerms]: 'BoomblasterDani' has previously used the unique ids [e25a85fa-07b3-3722-b2c4-fa1eaa120035] but is now connecting with '66023ea2-d434-4d71-a6d4-090fc63ca2c2' [14:16:56 WARN] [LuckPerms]: The UUID the player is connecting with now is Mojang-assigned (type 4). This implies that BungeeCord's IP-Forwarding has not been setup correctly on one (or more) of the backend servers. [14:16:56 WARN] [LuckPerms]: The UUID the player is connecting with now is Mojang-assigned (type 4). This implies that one of the other servers in your network is not authenticating correctly. [14:16:56 WARN] [LuckPerms]: If you're using BungeeCord/Velocity, please ensure that IP-Forwarding is setup correctly on all of your backend servers! [14:16:56 WARN] [LuckPerms]: See here for more info: https://luckperms.net/wiki/Network-Installation#pre-setup
I have a small isue. I cant use any bungecord commands like /server etc. I get this message in console when i start the server. Is there a way to fix this?

[reef fulcrum]

either you switch online mode or your ip forwarding is misconfigured or something (probs the latter, that screams that you didn't set ip forwarding up properly)

[mystic smelt]

I am having an issue, when I run any commands that go to the proxy it returns with unknown or incomplete command, see below for error, there are no plugins on the server and don't know what is causing it.

[reef fulcrum]

"incomplete command"

[wintry cobalt]

online_mode: true and the only thing that was updated in the Untangled FW was the IP. The MOTD show correctly, but not connecting to the correct servers.

[mystic smelt]

It isnt though, I do /server <Server Name> and it still comes up with the error

[reef fulcrum]

logs? can you reproduce without plugins on the proxy?

but, idk what you mean by it's saying that

[wintry cobalt]

IP forwarding was already setup and it has worked fine for years. Updating the IP shouldn't break that. If that was an issue it wouldn't connect at all or show the correct MOTDs

[warped kestrel]

so if i put online mode to false it works?

[reef fulcrum]

No, don't put the online mode to false, fix your ip forwarding

[mystic smelt]

Here is the entire log of the proxy and this is what showing in chat when doing /server Lobby https://pastebin.com/RsfDTdiV

[wintry cobalt]

I guess that wasn't directed towards me. I can't tell since there is other conversations going on in here.

[reef fulcrum]

I've literally never seen an error like that before from /server

the proxy doesn't even implement that stuff on commands

[mystic smelt]

Same, it only recently happened

[reef fulcrum]

so, that screams like maybe the servers are the issue? o.O

[mystic smelt]

Ok well that helps, thanks

At least I know the proxy is fine

[serene shadow]

I made a stream to my problem

[reef fulcrum]

I've already explained your problem

I can't explain much more and there is no magical fix

[wintry cobalt]

What would cause Waterfall to ignore the priorities list?

[marble kelp]

force-default-server: false or a plugin

[reef fulcrum]

outside of plugins or ^ or forced hosts matching

[wintry cobalt]

Nothing changed as far as plugins go. Literally all I did was update the IP and moved one of the forced hosts to a different domain.

But that is part of the problem is it's causing the forced host to be the default. If I turn it off then it uses the priorities list. I am really lost right now.

[reef fulcrum]

I mean,s o that means that whatever people ar econnecting with, the thing is apparently sending a hostname which matches that forced host

[wintry cobalt]

Yes, but it shouldn't be connecting to the forced host when connecting to play.domain1.com. That should be using the priorities list. The forced host goes to private.domain2.com Which works fine when you access it directly. But when you access play.domain1.com it takes you to the forced host server and ignores the priorities list all together.

[reef fulcrum]

the forced host system just looks up the hostname of what it has in the map

[wintry cobalt]

They are two completely separate domain names. They are not even the same

[reef fulcrum]

i once again cannot tell you why its going sideways there and there ain't much logging there from the server to provide help

[marble kelp]

you could try providing your config as well as the log with some additional notes regarding which host was used

[wintry cobalt]

It's self hosted, our severs are in a colcation DC.

[reef fulcrum]

    public static ServerInfo getForcedHost(PendingConnection con)
    {
        if ( con.getVirtualHost() == null )
        {
            return null;
        }

        String forced = con.getListener().getForcedHosts().get( con.getVirtualHost().getHostString() );

        if ( forced == null && con.getListener().isForceDefault() )
        {
            forced = con.getListener().getDefaultServer();
        }
        return ProxyServer.getInstance().getServerInfo( forced );
    }

that's the logic used to work out where people go, basically

[wintry cobalt]

I've removed the forced host all together from Waterfall and it still thinks this primary is a fallback.

[marble kelp]

then cencor them I guess

[reef fulcrum]

the priorities list comes into play later on but is kinda shoddy in that it might not always work depending on how the backend works

could not connect to target server

[marble kelp]

completely different issue lol

[reef fulcrum]

it failed to connect to a server

[wintry cobalt]

Target server is HUB01 that I am on

[reef fulcrum]

from there it should start off the priorities list

[wintry cobalt]

HUB01 is the only one on

[marble kelp]

wait, you only have one server and expect it to connect to anotehr one?

[wintry cobalt]

No, I am doing troubleshooting

[reef fulcrum]

that message means that it failed to connect to the initial server it was told to conenct to

behavior there is that it will then just use the prioritist list to connect

[wintry cobalt]

HUB01 is the only one in that list turned on and its the primary. It should be the target.

[reef fulcrum]

that error is specifically from an attempt to connect to something else in which it fell back to the prioritist list

[wintry cobalt]

That's the probelem

[reef fulcrum]

be that the reconnection handler thing or something else

[wintry cobalt]

It shouldn't be connecting to anything else. play.domain1.com is the primary URL for HUB01

I removed the forced hosts completely from the config and shut it off.

[reef fulcrum]

did you set force default?

[thin minnow]

I recently moved one of my servers to bungeecord or waterfall network and now all player data was reset, how can I get the player data back? Like a guide because I asked before but no one showed me how to fix

[wintry cobalt]

No, it is set as false, but that won't fix it, because I will need it to work with the forced host and setting that to true will disable that once I add it back.

Well, will fix it for now

[reef fulcrum]

cool, did you read the bungee installg uide?

[thin minnow]

i dont know where to find it

[reef fulcrum]

stop proxy, delete/rename the locations file, then start again, nick

cool, I told youto go read it

it's on spigots site

[thin minnow]

ok

this: https://www.spigotmc.org/wiki/bungeecord-installation/?

[reef fulcrum]

yes.

[thin minnow]

i have all of that correct exept for vanilla cord

whats that do?

[reef fulcrum]

if you're using spigot or paper, then that's irrelevant

if UUIDs changed, you changed omething

[thin minnow]

im using paper

and i have all of it correct

[reef fulcrum]

Okay, and, before you added waterfall

was paper in offline mode?

[thin minnow]

it was in online mode true

so it was premium

but since it is on bungeecord it had to be changed

[reef fulcrum]

well, is waterfall in onlone mode?

[thin minnow]

yes

premium only

[reef fulcrum]

did you enable bungeecord mode?

[thin minnow]

yes

[reef fulcrum]

did you enable ip forwarding?

[thin minnow]

yep

[reef fulcrum]

cos, those are the only real things which can impact that, as everything is based off the UUID

make sure that you stopped everything and that everything is properly saved

otherwise, if you configured it properly, UUIDs will be passed through properly unless you've got some plugin breaking that which I doubt, but, otherwise 🤷‍♂️

[thin minnow]

well i dont know what to do

im lost and sad

[reef fulcrum]

I told you what to check

outside of that 🤷‍♂️

[wise kite]

Please send large files/logs to a pastebin

[thin minnow]

ive did everything correct

[reef fulcrum]

other than asking for a copy of logs/configs, idk what to say

[thin minnow]

u want me to put mu files in that?
https://paste.gg/

[reef fulcrum]

if you want people to look at them and check, yes.

[thin minnow]

cant right now

[wintry cobalt]

Ok that fixed it. I renamed it and let it generate a new one. It didn't see it as a fallback which was positive. So shut it back down and added back the forced host and started it up and tested and now each domain is going to their respected servers again. After looking in the old one that makes sense because it was acting as if things were getting cached.

All servers are connecting as expected too. Thank you for your assistance on this.

[marble kelp]

that's basically how force-default-hosts: false is supposed to behave. forced-hosts via (dub) domains does not override the logout server

[wintry cobalt]

Correct, but in my case they were, but it's been resolved now.

[thin minnow]

https://paste.gg/p/anonymous/f4e6bda1cf334c73b2220bdc5e5fddb9

sorry it took so long

[weak edge]

should you use aikar flags for waterfall?

[marble kelp]

no, see pins

[tidal musk]

^ no

use those flags instaed

although I guess they're still Aikar's flags :P

[marble kelp]

xD

[scarlet urchin]

Hi, i'm wondering, is it actually possible to implement regional servers in minecraft java ?
Something where you can connect to a server that will redirect you to a proxy located in your region, and, obviously that original server doesn't stay in the data routing ?

[royal hawk]

Not really without multi-advertised networks or geographical DNS load balancing (which doesn’t work over half of the time because DNS privacy)
Once a connection is established it cannot be changed until you disconnect for minecraft java
So the first one is really your only option if you absolutely require it

[scarlet urchin]

That looks like a pretty good solution, thanks

The point is just doing that to get the best performances possible on a server network (you match europeans on EU servers and americans on a USA server)

[marble kelp]

tbh just using sub domains and letting the player input the one you need is a lot more reliable

(and then choose the address to display to the player on your website based on ip 👀 )

[scarlet urchin]

I also thought about that yeah

Thanks for that tho !

[wraith owl]

hello, if i have a waterfall server with ranges from 1.8.x to 1.18.x, can 1.8 users still use the shulker boxes and all of the latest features?

[nocturne fog]

waterfall does not provide any translation from newer versions, what that means is that with the latest version of waterfall you can have between a 1.8 and 1.18 server

so if you run a 1.18 server, without something like viaversion/viabackwards/viarewind or protocolsupport, old client versions will not be able to join

[wraith owl]

ah i see thanks

[tidal musk]

No, the client doesn’t know about the newest features that has been added after 1.9

Translation plugins like ViaVersion just translates the stuff, but they can’t add the features since the client does not know the features that added in the client in 1.18.1.

[stark hawk]

Hi. I've got a (hopefullky basic) permissions question. I'm using latest version of Waterfall with BungeeSafeguard installed on the proxy to provide whitelist functionality. Also latest LuckPerms installed on proxy and all backend nodes. I can use 'whitelist add <user>' command on proxy console, but can't for the life of me find out how to give permission to the server admins to use that command in-world. What am I missing, please?

[blazing ocean]

You are using /lpb on the proxy right?

[stark hawk]

That command is not recognised on the proxy console.

[blazing ocean]

But /lp works? didn't know that lol

[stark hawk]

No, that doesn't work on the proxy console either. In fact, I'm not even sure if the plugin loaded on the proxy as no LuckPerms sub-folder was even created.

[blazing ocean]

Does anything related to luckperms exist in the proxy's log?

[stark hawk]

I'm one step closer. I hadn't realised there was a Bungee specific version of LP for the proxy. That's now loaded, so the /lpb command now works from the proxy console.

[blazing ocean]

You can use /lpb editor to set groups and permissions up for Waterfall

[candid kayak]

Hello! Is there an effective way to find lag on a waterfall proxy using the spark profiler? I understand that due to the nature of bungeecord being multi-threaded it shows a bunch of threads, but I'd still like to learn more about what may be consuming too much CPU time.

[mental holly]

should i use aikar flags on waterfall or na?

[delicate phoenix]

See pins.

[neon stratus]

does anyone know what this means on my proxy?
InitialHandler - NativeIoException: readAddress(..) failed: Connection reset by peer

[royal hawk]

Connection got terminated by the other side

[glossy agate]

Hello,
Following the Log4j problem, I would like to know if Waterfall version 1.16 was vulnerable?

[reef fulcrum]

yes

[glossy agate]

I can use the latest version 1.18#475 (by replacing the server file) won't it break the plugins and the connection to paper 1.16.5 servers?

[reef fulcrum]

should be fine

[glossy agate]

Thanks

[sinful laurel]

[/127.0.0.1:25565] <-> InitialHandler - read timed out does this spam means that some of my servers try to ping the proxy directly?

[royal hawk]

Not necessarily
Some hosting software does that to check on status of the running software (which is stupid ofc)

[sinful laurel]

Which is weird because it is a dedicated machine behind a very restrictive firewall and tcpshield

[marble wraith]

it used to work fine.... wtf

restarting... very strange

very weird, it's running, even restarted

oh maybe my version...

it's my version, whooops

[royal hawk]

99% sure then it’s caused by TCPShileld

[sinful laurel]

But why 127.0.0.1, do they send spoofed packets to me 🤨

[ancient aurora]

My Waterfall Don`t startup...

[tulip surge]

Please Use a Pastebin
Using a pastebin allows others to better provide you with support.
Unfortunately, I was unable to automatically upload your file.

[reef fulcrum]

You broke the layout

See where it's telling you that a server doesn't exist, your config is malformed

[rustic talon]

Server has not starting ...... :(

Help :(

[tidal musk]

Please don't randomly ping people

[rustic talon]

;-;

[rustic talon]

.

[tidal musk]

Your server has started correctly. Make sure you're using the bungeecord/waterfall egg for ptero, else it won't report the status correctly (but will still work)

[rustic talon]

How ?

;-;

[tidal musk]

Are you using a hosting provider or running pterodactyl yourself?

[rustic talon]

🤔

[rustic talon]

I am not understanding?

[rustic talon]

My server 😂

[tidal musk]

Are you renting your minecraft server from a hosting provider?

[rustic talon]

No

Falix hosting ;-;

[tidal musk]

See if you can modify the server type in their client panel, otherwise contact their support

[solemn siren]

I am having a problem when doing the compatibility with bedrock because I get this error having the ip forward activated and the bungeecord also activated in the spigot server, then I would like to know how I can fix it, sometimes it happens to bedrock only and other times not.

[errant dew]

do anyone know about multibungee setup can anyone help me with than
@royal hawk can u help me??

[severe badger]

No need to cross post

[errant dew]

this area is for proxy devs ri8

[visual jungle]

that is #waterfall-dev or #velocity-dev

[errant dew]

K

[thin minnow]

u just showed someones ip u know

i would delete that

[thin minnow]

Enable bungeecord in the spigot.yml of all the backend servers

[sinful laurel]

Are you serious?

[thin minnow]

yes

that was someones ip

[sinful laurel]

Use Google to improve competence level what this address means

[thin minnow]

ok

confused but whatever

[sinful laurel]

Just Google the address

[thin minnow]

why

[sinful laurel]

To not post any stupid replies any further

[thin minnow]

ok

[bold sierra]

What causes waterfall/bungeecord to ReadTimeOut?

[marble kelp]

network issues

[reef fulcrum]

read time out means that nothing was recieved in 30 seconds

[bold sierra]

Hmm, weird

Because when I disable bungee and join directly to the server that doesn't happen

Both are on one network (server and waterfall)

[marble kelp]

how do you connect between the proxy and server?

[bold sierra]

Thing is that absurd amount of packets are sent

But it doesn't kick them on spigot server when directly joined, so it doesn't make sense to me

[bold sierra]

https://bin.jamo.tech/oyiyelepumimeyac.yaml

We do have absurd amount of packets lmfaoo

[marble kelp]

I mean, it's pretty obvious why

[bold sierra]

I know that we have absurd amount of packets

But I don't understand why I don't get read timed out when joining the server directly

Disabled particles

https://bin.jamo.tech/nifopuwabemofoji.yaml

Well most

Still getting timed out

With bungeecord (any kind) my ping goes up, without bungeecord it stays the same

[neon stratus]

I’m not sure where the best place to ask this is, but I think my issue is relating to router settings. I’m using bungeecord and self hosting, but even when I join through the local ip I get a lot of read time outs. An ideas? Thanks

[round nexus]

hi, how do I do motd colors? it doesn't seem to work like on a vanilla server

[neon stratus]

Try & or the s color code symbol, forgot the name

I think you can do Alt+0 1 6 7 if you have a num pad

[round nexus]

I have for example \u00A7a

[neon stratus]

oh, yea i have never used that. does §a not work?

[reef fulcrum]

I wanna say that & works but may need the section character

[round nexus]

what would be the & equivalent of \u00A7a?

[reef fulcrum]

Really? use a plugin if you want anything beyond the basics

well, that unicode escape is the section symbol

[neon stratus]

https://preview.redd.it/yldfhnb4xp411.png?auto=webp&s=747bcfe3d9db9aba2b0bfecacf8299619f8eca38

[round nexus]

colors work, what about newline?

[neon stratus]

just use a pipe

easier to format anyway

[reef fulcrum]

yaml multiline string

[round nexus]

I'll try

[reef fulcrum]

\n should also work I think but can be more janky, sometimes it apparently says nope

[neon stratus]

this is what i use:

motd: |2- &r&f&lMC&r&a&lBlock&r&6&lBuilds&r&f&l.Net [&r&a&l1.8 &r&f&l- &r&6&l1.17&r&f&l]&r&r&r&r&r &r&f&r&r&c&lCreative &f>> &2&lSkyblock &f>> &e&lTowny Survival

[round nexus]

holy shit

[neon stratus]

yea color codes look real nice when they arent parsed

[round nexus]

what does |2-mean? and all the & codes?

[neon stratus]

the |2- is a pipe i think its called, and its just how yaml knows that theres 2 lines

but dont quote me on that lol

the & codes are the color codes

so &aTest would make "Test" green

and &a&l makes it bold and green

the bold &l always has to go after the color

theres motd generators online but i havent had much luck with them actually working

its easrier to just trial and error to get it perfect

[round nexus]

&r&f&lwhat's that?

f is color

l is bold

[neon stratus]

&r is reset, &f&l is white bold

[round nexus]

oh reset

[neon stratus]

&r just resets the formatting back

[round nexus]

nice

[neon stratus]

just come back to this everytime

i bookmark it

[round nexus]

also, my server-icon is like supersmall 😉

[neon stratus]

theres an icon size idk what it is though

[bold sierra]

So I did a test with Velocity, it seems that velocity runs the same as joining the server directly, I'm assuming this is bungeecord issue then?

[reef fulcrum]

I mean, the thing sits in the middle of the connection and does stuff

it's long been an issue that bungee can add substantial amounts of latency for some random ass reason

[bold sierra]

Well it only happens when there's a lot of packets being sent

Like chat spam

Or in our case since it's a prison server we do have quite a lot going on on each player mines

Thing is we tried running Velocity before switching back to Bungee, I think we had random disconnection issues

[reef fulcrum]

well, yea, velocity's network pipeline is much better designed

[bold sierra]

I guess I'm gonna give Velocity another go

Could it be cause of compression tho?

[reef fulcrum]

well, that would add some latency

[bold sierra]

I did try to set it as -1

Since on our server jar we use velocity compression

[reef fulcrum]

I mean

de/compression takes CPU cycles

the primary thing is tryna find the balance between bandwidth and CPU cycles

[bold sierra]

I mean setting the compression to any value really did not help

[reef fulcrum]

for the proxy, it generally makes little sense to compress traffic over the loopback

but, it's not gonna magically solve the fact that bungees network stack is kinda bleh and has rando issues where some people just get screwed by the thing

[bold sierra]

I guess I'll move back to Velocity but I'm very afraid we going to have the same issues. As we had a lot players complaining about randomly being disconnected

[reef fulcrum]

well, velocity is more receptive to stuff than we are at this point

I've generally just given up as theres little I can do these days which isn't mostly snakeoil best solved elsewhere or likely to induce issues, be that mods, plugins, etc, etc

[bold sierra]

Wouldn't it be possible to rework Bungee's pipeline?

[reef fulcrum]

without breaking plugins?

no

[bold sierra]

Right

[strange glade]

I'm using centos 7, my minecraft server is on, but I can't log in because the firewall is active, how can I log in without turning off the firewall?

[sweet mountain]

If your firewall isn't allowing you to login, that just means you need to add a rule to your firewall.
Your servers should be behind your firewall, but your proxy's port should be open.

Make sure you have something like -A INPUT -m tcp --dport 25565 -j ACCEPT

Probably a dumb question, but what causes this?
I don't recall changing anything, and it was working perfectly before.

I tried troubleshooting for a while, but I am probably missing something obvious.

[20:34:44 INFO]: [/127.0.0.1:62121] <-> InitialHandler has pinged
[20:34:49 INFO]: [/127.0.0.1:62124] <-> InitialHandler has connected
[20:34:49 WARN]: [/127.0.0.1:62124] <-> InitialHandler - overflow in packet detected! Cannot receive string longer than 16 (got 17 characters)

[reef fulcrum]

generall offline mode and too long of a name at that stage

[sweet mountain]

Interesting

[visual jungle]

is waterfall bungee fork

and does bungee plugins work on waterfall

[tidal musk]

yes

to both

[visual jungle]

good

then i can use

[umbral crypt]

Yesterday I was asking about velocity's compatibility with modern versions of Forge. I completely forgot to ask about Waterfall. Does Waterfall (since it's a Bungeecord fork) happen to support more modern versions of Forge, such as 1.16.5? Thanks!

[royal hawk]

No proxy has support for it

that’s due to how forge on 1.13 and newer is designed. You can’t achieve reliable compatibility without modifying the client (and server)

[umbral crypt]

That's sad :/ haha. That approach is what I was told yesterday in the Velocity chat too. I asked on the Forge forums yesterday if anyone had any idea how to actually go about doing that. Since I'm developing my own mods, I'm happy to put in the extra work to create that compatibility, just not sure how to actually get there

[royal hawk]

If you use forge 1.13 and newer you can’t achieve this without changing how forge works fundamentally

its not a matter of designing a mod that respects certain things

this is by design in the mod loader

Forge syncs data before the login sequence is complete. Since you can’t reset the login phase you have no way to renegotiate that data. And that’s what breaks all of this

[umbral crypt]

Ohhhh, okay. Do you think Mixins could fix that, or would not even that work?

[royal hawk]

That would in theory but you have to find a way to achieve this that won’t break other mods in the process. Believe me if I tell you that i tried dozens of times with different approaches to this. All of them failed at some point with some mod

[umbral crypt]

Damn. I was developing my project in Paper but then decided I wanted to go to Forge simply for the Conquest Reforged mod (adds like 10,000+ blocks and textures that brings MC to a visual fidelity that I've never seen before - so it'd be a whole new level of immersion). So that would be the only mod that I would be using that isn't developed by me. If you don't mind me asking further, if you did test with Mixins, what did that process actually look like?

[royal hawk]

To put this in perspective let me copy my notes on project „Unity“ and post them here, one sec

[umbral crypt]

Thank you!

[cobalt dawn]

Was stress testing my Waterfall network by making a ton of users join at once, and I noticed that after a few seconds players started getting timed out. This is a bit of an issue since I have a rather large player base and I need the network to be able to handle a lot of players joining at once. Does anyone know how to fix this?

[umbral crypt]

Since I know this is something you've worked on too, here's the current solution I've been offered over on the Forge forums https://forums.minecraftforge.net/topic/108187-forge-compatibility-with-proxies-mixins/

Not really what I was ideally looking for, but it's a solution I suppose

[scenic relic]

Guys

[solemn siren]

?

[modern pawn]

Hello, I encountered this problem when starting waterfall. Can someone help me?
Waterfall-Bootstrap:1.18-R0.1-SNAPSHOT:de46489:477

openjdk version "17.0.2" 2022-01-18
IBM Semeru Runtime Open Edition 17.0.2.0 (build 17.0.2+8)
Eclipse OpenJ9 VM 17.0.2.0 (build openj9-0.30.0, JRE 17 Linux amd64-64-Bit Compressed References 20220128_115 (JIT enabled, AOT enabled)
OpenJ9   - 9dccbe076
OMR      - dac962a28
JCL      - 64cd399ca28 based on jdk-17.0.2+8)

https://cdn.discordapp.com/attachments/289587909051416579/937340397984829460/unknown.png

[royal hawk]

you have a headless JVM build. You need a full JRE/ JDK.

[limpid cape]

What does the restricted option mean and how do i configure it?

[marble kelp]

it sets whether or not the per-server permission (bungeecord.server.<servername>) should be used with the /server command

see https://www.spigotmc.org/wiki/bungeecord-configuration-guide/ for all bungee config.yml options

[limpid cape]

And I can set up the permissions in the permissions: block?

[marble kelp]

yes. ideally you should use a permissions plugin like LuckPerms though

[limpid cape]

Hmmm, how would that be passed through to waterfall then?

[marble kelp]

what do you mean? you install LuckPerms on the proxy

if you want to share the same permissions in your whole network then you use the same database between them

[limpid cape]

Hmm alright...

Sorry I'm not as knowledged in setting up Waterfall in production environments. Let me look into that

Luckperms doesn't seem to be supported in 1.18 forge tho, which one of my servers will run

let me look for alternatives

[marble kelp]

modded servers are most often not supported by anything

[limpid cape]

True

[marble kelp]

also how are you doing permissions on the forge server now?

[limpid cape]

not, i've not set that up yet

I had like 4 servers on 4 different ports and wanted to consolidate them into one

[marble kelp]

so how did you plan to do it? xD

[limpid cape]

Almost didn't plan for that lmao

[marble kelp]

also modern forge can't really be supported by a proxy anyways

[limpid cape]

yeah

Guess I'll go with fabric

it has the main mod i want in it anyway

[marble kelp]

(also technically LuckPerms supports Forge if you use Sponge)

[limpid cape]

(yep, but that's only 1.16.5 as latest ver)

[modern pawn]

I thought the tar.gz file was decompressed directly into the directory and could be used after configuring environment variables. Do you mean that I still need to compile or something?

[royal hawk]

Not really; There are a few things wrong with your statement; The jvm relies on a few system libraries that aren’t present if you don’t install it using your package manager
So to speed this up, what’s your OS and system version?

[flat gull]

Hello there, I encountered a weird error, it speaks for itself.

I tested 3 different locations and clients, official MC client and multi mc, friends that tried to connect to it as well.

ERROR Error authenticating <Insert Username> with minecraft.net
INFO java.lang.Error: Method became inaccessible:
LoginEvent(cancelled=false, cancelReasonComponents=null,

I will provide the full error message upon request. Need to find a new pastebin service.

I also tested that I can connect to other servers, updated the waterfall jar to the latest version, and used java 17

[lost river]

is there a stack trace below the error line?

[flat gull]

There is

There was a string which looked like a token, don't know if it's sensitive data I can share freely

It probably isn't, but better safe than sorry

[lost river]

just replace the thing that looks like a token with XXXXXX

[flat gull]

Sure

Give me a moment

https://pastebin.com/if53MtWy

Replaced data placeholders:

USERNAME
XXXXTOKENXXXX
IPANDPORT

[lost river]

Here is the cause:

30.01 18:19:35 [Server] INFO Caused by: java.lang.IllegalAccessException: class net.md_5.bungee.event.EventHandlerMethod cannot access a member of class net.lucypoulton.squirtgun.bungee.BungeeListenerAdapter with modifiers "public"```
Is `net.lucypoulton.squirtgun` something from you or a downloaded plugin?

[tropic blade]

Is upgrading from bungee to waterfall just a drag n drop same as spigot to paper?

[serene tide]

Yes

[flat gull]

That is...I know the plugin

But I didnt get the squirtgun version

[lost river]

maybe try updating your plugins

[flat gull]

They are updated

[flat gull]

It was the plugin mentioned, the bukkit bridge seemed to have some kind of issue, the dev is unsure themselves.

Thank you for resolving my issue, didn't think this would really interfere with any auth.

[tidal musk]

for a server that gets a lot of connection request from time to time lets say 200 requests across network in a minute only, what are vital configuration settings to change?

[reef fulcrum]

I mean, theres very few config options in the first place

for stuff like that, that's more a tech solution

i.e. load balancing

[tidal musk]

i mean my machines are very good all running on AMD Ryzen 9 5950X, machines connections are in a private network all machines are in the same location and all have ram more than enough

so when people start rushing to server,

they all get stuck kin queue

[reef fulcrum]

paper servers?

[tidal musk]

when i do /send queue lobby, all queue comes to lobby instantly

[tidal musk]

yes, latest versions

[reef fulcrum]

don'tmention

it is worth noting that there is an option in paper for joins per tick or something

I added a buffer as people where upset that joins would flood in especially on startup and kill the server as you'd process every join attempt on the first tick

[tidal musk]

if it is becacuse of paper config, wouldn't that command work like that still?

as an op when i send people to a server does paper ignores join attempt per tick

[reef fulcrum]

well, was a note

waterfall doesn't care about op

waterfall also doesn't have its own "queue" system outside of the CPU speed

that would scream like a plugin or something

[tidal musk]

yes actually i know that is why i find it really weird

[tidal musk]

i have a queue plugin yes also

the other day, i changed network_compression_threshold: from 256 to 64

it worked greatly but still sometimes connections stuck

[tidal musk]

if i lowered it, would that work?

should i look into bungee discord for bungee config file actually 😄

[reef fulcrum]

disabling compression between the proxy and the server is generally the standard thing

proxy has its own config too, some people have said disabling that can help with some stuff but that varies from person to person from network to network

[coarse rose]

Hello.. i have this problem

Is It Java?

I have Java 17

Help pls

[reef fulcrum]

1. In the future, send the actual log, not a picture

[coarse rose]

Sorry

[reef fulcrum]

2. you have the headless jre, you need the full one which includes extra libraries

[coarse rose]

how can i install it?

[reef fulcrum]

depends on your distro

install the package which doesn't have "headless" in the name

[woven bison]

Does anyone know what's going on?

The ports and the firewall are well configured because when I disconnect the server from waterfall, I can join it normally

[dire ermine]

So I did a major worldedit three days ago on my test server, turns out I didn't have fawe downloaded and my server was in a crash loop and I had to force close the session. After trying to reboot it up, it says that the port that the server was on is now binded to another server. So I changed the port to a different one and it boots up just fine but now I'm unable to connect to it and it times out my requests to connect every time now.

Anyone know a work around for this?

[royal hawk]

you probably still have a corpse of the other server running or you didn’t account for the changed ports in firewall/networking or NAT

[woven bison]

maybe you know the answer to my problem?

[royal hawk]

It tried for five seconds to get to the server but never got a response; You have either a network issue or a firewall issue

[woven bison]

but why when I create a regular server without waterfall, I can access it?

on the same port and ip

[royal hawk]

Are you using hosting software like pterodactyl?

[woven bison]

dedicated server

Hetzner is the supplier

[royal hawk]

server and proxy on the same server?

[woven bison]

Yes

[royal hawk]

and your firewall is set up correctly to allow loopback? under linux that network adapter is commonly named „lo“
I assume you’re not using containers; so you are using 127.0.0.1:port for the server?

[woven bison]

we use containers

And a virtual lan

Linux on a firewall

Windows for servers

Specifically opnsense

[royal hawk]

Then I say its on your part to get your network forwarding in order. Doesn’t surprise me now that external > forward > container works but container > forward … > container doesn‘t
This is usually especially tricky and I‘m sorry I really cant help past this point

[smoky zealot]

this seems more like a networking issue, try to get a minimal repro with like raw tcp connections or something

[tidal musk]

is there a way for when the backend server restarts it sends them to the hub server?

[reef fulcrum]

using plugins, yes

[smoky zealot]

i'm curious about this little quirk: when i tell bungeecord that the sever is on 0.0.0.0:25565, it can't find the server and i get the following error:

[18:18:31 INFO]: [SirJosh3917] disconnected with: Could not connect to a default or fallback server, please try again later: io.netty.channel.AbstractChannel$AnnotatedConnectException
however, when i change the IP to 127.0.0.1:25565 or localhost:25565, it then finds the server fine.

this is confusing to me, as when i direct connect within the minecraft client to 0.0.0.0:25565, i'm able to connect just fine. what could be the cause of this?

my waterfall instance is NOT containerized

[marble kelp]

0.0.0.0 is not a real IP

it means that a server should bind to all available interfaces

don't ask my what the client does with that that it magically works with an invalid ip lol

*me

[smoky zealot]

i still find it really interesting how minecraft can connect but waterfall can't
i'm on linux so maybe this is a reasonable explanation? https://superuser.com/a/1307057
i guess the mc client somehow uses more closer-to-the-os apis than whatever waterfall uses?

i'm still trying to learn networking stuff myself lol, thanks for mentioning what you did

[marble kelp]

I assume the mc client just does something dumb and non-standard

[smoky zealot]

i'm using old mc so that's very likely lol

[tidal musk]

What plugin ?

[marble kelp]

e.g. my KickInfo does that

HubKick, the second oldest spigot resource should also still be able to work

[tidal musk]

Ok

[exotic lintel]

Hey, i got an realy strange Problem...
I have set up the Proxy and Players can connect to it.
All Plugins on the Proxy dont work for Players. I can use it and of course without * Permission.
Im using Luckperms and Luckperms Bungee. Every user have the Permission for this but they only get the Error: Unknown command /.... <- [HERE]

Some Forums and Discord Users says thats an Permission problem or an Plugin block this.
But i have removed every Plugin which can do it. And the Problem is still happend.

Does anyone had the same problem in the past and have maybe an fix for this ?

MC Version 1.18.1

[marble kelp]

check LuckPerms verbose mode for what permissions it checks

[exotic lintel]

Have tried this maybe 1/2 days ago.
There is nothing :/ No Error no Informations

[marble kelp]

are you sure that you are syncing the permissions to the bungee or if not that you have set permissions on the bungee?

you also want to make sure to use verbose mode on the bungee

tbh that's more of a LuckPerms support thing than a general waterfall one seeing as the plugins actually work for you

[exotic lintel]

Have tryed with an User and on Bungee via /lpb verbose record and /lpb verbose paste i got an link
and there is nothing ö_Ö

[marble kelp]

well what does it show when you run the commands and it works?

[exotic lintel]

wait a sec im connecting 🙂

There a the permission what i need to do this command

oh and on each Server (and proxy too)
i have set this in the Settings from Luckperms

server: Server

and the same mysql database

[dusk perch]

What's the best way to reload waterfall permissions?

Permissions meaning the stuff in config.yml of the proxy.

[reef fulcrum]

greload