#waterfall-help
1 messages · Page 47 of 1
.paste
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
if by "back doored", you mean, "they got into a backend server and messed with stuff", there would be nothing in the bungee logs?
Meh, but if they have "back-doored" your server they would have needed to install something previously which wouldn't occur in ya bungee logs
rip cat was faster than me on mobile
They were able to op themselves
Sorry not the best wording
I'm wondering if this could be a paper exploit
Stuff like that would occur on the server level.
Nope
AH ok
That would imply that you had a hijacked plugin
Is spigot good at preventing that?
Preventing what?
You'd want to find out what they did in the first place, if it's a malicous/hijacked plugin, most likely no.
Not like that'd be paper's/spigots job anyway, better get a rid of that plugin.
some guy had some plugin on the spigot forums somewhere, maybe in the resource section, which would scan plugins for potential issues
ye
Thank you!
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
something closed the connection in an unexpected manner
But why?
speak to the plugin author
Can't say, something closed the connection in an unexpected manner
that generally boils down to network issues or plugins closing the connection in a not so client friendly manner
Now it won't let anyone connect via 1.16.4 and kicks everyone when warping (using essentialsx)
But lower versions work
I contacted viaversion and they said it's not their plugin
this is where admin boys grow up to become admin men
gotta start debugging urself and find out
So my server was hacked ^^^ using some exploit
Not exactly a typical scenario
And I've contacted plugin devs to narrow things down and disabled plugins
Very un-typical for a restart to cause teleporting of any kind to kick you off servers and to not allow 1.16.4 connections...
For all intents and purposes, my (decently-sized) server is now unusable
Even after multiple hours of troubleshooting
hm
does it sound like the guy in general
or uh
you got hijacked you say?
then nvm that's different
well if you're hacked then someone else has access to your server and you better wipe and redo or something
unless you've got an audit and stopped it fast enough
I think I stopped it
But my issue is now it seems it's broken
Doesn't seem to be plugin related
idk what's going on then
Know anyone that can?
all that says is that something closed the connection in an unexpected manner, check your proxy logs for errors, best advice is gonna be to replicate without plugins
only real other thing is gonna be network related aspects
ya and idk what "hacked" means at this point anymore
being cagey with info doesn't help us cuz we aint mind readers
Effectively someone opped himself through some minecraft exploit
Was able to remove him, and he seemingly only used worldedit according to logs to destroy our spawn
what exploit
But I restart all the servers (including bungee)
forceop has been dead since 1.8 with that sign/command block stuff afaik
A cow spawner that spawned a minecart command block to op himself
creative mode?
Yes
o o f
ah creative mode
I use creative inventory control
you will have endless fun with that
Yea, you need plugins to protect against stupid stuff like that, basically
But here's the thing
good heccin luck
best advice is to isolate a creative server as faaaar way from the network as possible
it wasn't a big issue because he's gone and I'll patch it I'm sure
Like, to the point of disabling all the darned LP commands and op on there type deal
creative mode's hole is bigger than the one cut into the titanic
Here's the issue now, even on servers he didn't go on, literally no one can now do /warp or /spawn without getting 'disconnected' and it literally won't let anyone connect with 1.16.4 to the hub
one lil patch ain't gonna do nothin
hard to say, maybe something was placed in the world? command block hijacking something, I don't think that they can but god knows
My issue isn't the exploit it is that my server is unusable. I can literally just delete the creative server. My issue is that the entire network is now broken
And it even affects servers he didn't go on
Ah Vault seems to be causing the issue what on earth how
vault just provides an API for stuff
that implies something that broke when you removed vault is the issue
When I un-add it I can now join
but when other joins
it then kicks me
I know chat & perm plugins hook onto it
I think it may be a bungee issue then
Honestly, at this point, I will pay anyone that wants to help me
$15 an hour
This is the bungee error
Doesn't seem related to vault now
It's not plugin related
And the host says it isn't them either
all that tells us is that the server closed the connection
that's useless
any useful info would be in the logs
beyond that, replicate without plugins™
I've disabled them all doesn't fix it
let's you join but only for up to a few minutes
Nothing in the server that could crash you cares about time, or would print nothing useful to the console
that more screams network
Well, the server closed the connection
as to why, god knows
command block in the world kicking people 🤷♂️
Check your users clients log and make sure nothing weird is happening client side?
I changed the world
I've done everything
Testing for multiple people across multiple countries, accounts, and ranks
All have the same issue
I just may have to start from scratch and see if that works
thanks for your help guys
im tryna run another server separate from my bungeecord servers. ( i have the bungee server, and 2 other servers on that proxy running )
But trying to run a server standalone without any connection to bungee works, but im not able to connect, with no reason about it. Do i need to stop all instances of bungeecord and it's servers running in order to run the standalone server i need?
No, you sould be able to connect to the standalone server, even if there's a network runing on the same machine.
Make sure that the port for the standalone server is open and it's different than all other servers/proxys, bungeecord mode isn't on, and the standalone server isn't added to the bungee server config.
"[21:24:14 WARN]: Event PostLoginEvent(player=omitted) took 530ms to process!" this always happens when logging in or logging out any ideas on why?
you have a slow plugin listening to the postloginevent
more specifically, one plugin or plugins is taking more than 500ms
use spark or a proper sampler
I thought that postLoginEvent did not effect the time the players takes to connect
It does
you're holding up the connection
Not to mention, the warning is purely "this event took X amount of time for all plugins to do whatever with it"
I think it might be a plugin I made that sends a webhook when a player connects/disconnects any ideas on how I can make it not affect the time players take to connect? I am using PostLoginEvent listener
Fire off a new thread or a completeable future from the event?
Well a simple future suffices in this case unless you need the result to do something
I am not very experencied with java/ programming what should I search for?
Search and read up on Completeable futures and threads in java. Alternatively you could use the scheduler to start a delayed task with 0ms delay
careful: once youre detached youre unable to edit or interact with the event
so if you need to return data to the event then you cant not block it
Thanks I will look into it
anyway to sort out specific pings?
networking issue
@humble thistle Are you hosting it or is it on a hosting provider?
Make sure that the servers have the same IP. If they do, you should be able to use localhost. It's possible that the hosting provider doesn't use 172.18.0.1 as the docker interface or something, so 172.18.0.1 isn't working.
I'd suggest just using the public IPs in the config.yml. If they're all on the same IP, additionally bind the spigot servers to 127.0.0.1 by specifying bind-address in server.properties to prevent some security flaws. If they're not all on the same IP, leave bind-address as 0.0.0.0
Was thinking it was hosted but I forgot that 172.18.x.x. is usually docker... You could do a quick test to see if that port is available by opening up powershell and running test-netconnection 172.18.0.1 -port 25566
oh it's definitely hosted externally, it's in the filezilla temporary file directory
Nice pickup
In that case powershell will be useless... run nc -vz 172.18.0.1 25566 which should tell you if that port is open
on the server with your waterfall ofc.
help
i have the permission
permissions:
admin:
- bungeecord.command.alert
- bungeecord.command.end
- bungeecord.command.ip
- bungeecord.command.reload
- jpremium.command.forceviewuserprofile
- bungeecord.command.server
- bungeecord.command.send
./server name
and i cant put the command
groups:
Alfre_DGR:
- admin
"can't put the command"
Define, what aspect of it is not working, what does it show?
Do you have a permission plugin installed on the proxy?
what kind of plugin that when u enter portal it teleport you to another server?
i cant use the command /server
i have luckperms in the bungee and all servers
Please don't ping
that's your issue, having a perm plugin disables the perm stuff in the config file
so can i add the permission in the perm plugin?
Yes
bungeecord.command.server i added in luckperms and dont work
are all your LP instances networked properly? did you try with /lpb?
What's the diffrence between Travertine and Waterfall except from 1.7 support?
nothing
And can I move from WF to TT just like so?
99.99999% of the time, yes
There're no api changes?
ABI wise, no
potential issues due to changes in the game dating back way too many years, potentially, but that's purely on plugins
and can I use waterfall as the library in plugin's lib to travertine?
yes, there are no ABI changes
What's the setting to change the server change cooldown please?
Or well server join.
cooldown?
Connect to lobby, then back to survival they need to wait 54 seconds?
OR logging off and on quick
Not a waterfall thing from the sounds of it
o-o hmm ;o not got anything else on there, says "you must wait 54 seconds before logging in again"
that's gonna be something on the servers themselves if it's affecting server switches
Oo I see, I shall take a peak, thought it'd be bungee but that, kinda makes sense being the server? 😄 thank you. :3
Project Waterfall build #389: SUCCESS in 3 min 23 sec: https://papermc.io/ci/job/Waterfall/389/
Project Travertine build #167: SUCCESS in 2 min 4 sec: https://papermc.io/ci/job/Travertine/167/
HELP
how can i resolve
this
UpstreamBridge - NativeIoException: readAddress(..) failed: Connection reset by peer
what is your question?
the error
UpstreamBridge - NativeIoException: readAddress(..) failed: Connection reset by peer
yes i got it, discord lagged
the connection was closed in an unexpected manner between the proxy and the server, iirc
check your server logs, etc
I do not see anything unusual in the server log as it enters and leaves normal. The little error appears in the waterfall
recommend network_compression_threshold=-1 from bungeecord?
no, you want compression, as to how much, that's questionable to your setup and needs
I put -1 on my servers, always when one player hits the other, he is disconnected
Do I go back to 256?
🤷♂️
ok
Is there a way to make a custom waterfall queue system?
Well Waterfall supports bungeecord plugins so I guess there is?
where would I get the Travertine API?
Not the one you install, the one you use for development
@quiet depot you want the internals? Those are not provided by the maven repo, either install an artifact locally or find another way
Yes its a fork of bungeecord
Could anyone find a great youtube video about setting up waterfall? If yes, dm me with the link pls
- Download jar
- java -jar WaterFall.jar
ready
hey question, how do I set a fallback server? So if do a restart on a server all the people on it go back to the lobby
and when players log off in the gamemode and rejoin, they are in the lobby
you'd need a plugin to deal with that
and force_default in the config file will force people to join the default server
the default server can be set to the lobby or hub server right?
force default just makes it not reconnect the player to their last server and uses the priorities list
oh alright
all of a sudden im unable to connect to my hub server. I havent changed anything in the config, im on the latest version, im not sure whats going on. if anyone can help, thank you.
nevermind, i was using the wrong IP address
Hey, I want to make a fork of WaterFall, but, is possible if I work with Eclipse ?
use travertine as a base as it has the update scripts and stuff f you're planning to maintain it, but, yea, works fine in eclipse, you'll need the lombok plugin because upstream is fun, and to learn to use the scripts (works the same as papers)
Okay, thanks and how I can get the source of Travertine please ?
it's on the github, in the same org as paper and waterfall
Okay, I download with that for import to Eclipse ?
No, clone it
in bash
./travertine p, and then open it in eclipse as a maven project
You might wanna read over the paper dev docs, as the scripts all work the same
Okay, I go to check and try that, thanks you 🙂
Hey! I made a plugin that makes a webhook request when postloginevent is triggered that is delaying connections to the server and when the webhook receiver is offline it causes all players to be disconnected how can I make the eventhandler run async? I am changing @eventhandler to @EventHandler(priority = EventPriority.LOWEST) should that help?
Do the request on another thread
This is my first plugin and java experience would that be "new Thread()" ?
use the scheduler, it uses a thread pool
Thread creation is potentially slow as balls
would that be something like Executors.newSingleThreadExecutor();
bungee already gives you access to a thread pool through the scheduler
oh
What should I search for? I am not sure
it was either exposed on Plugin somewhere or on the Proxy thing, you'd need to check the API, my brain is like, gone
might be this https://docs.phoenix616.dev/bungee-api/
idk I am using this, it works
- you don't wanna be calling that all that time, that's redundant
use a chached pool
- i forget, getScheduler on Proxy or something
yes its getScheduler but then idk what to do
runAsync(Plugin owner, Runnable task)
Schedule a task to be executed asynchronously.
?
yes
but I am listening to postLoginEvent
so?
nevermind, I got it working after a lot of tries thanks
It is now
private Main plugin = Main.getInstance();
@EventHandler(priority = EventPriority.LOWEST)
public void onJoin(final PostLoginEvent postLoginEvent) {
plugin.getProxy().getScheduler().runAsync(plugin, new Runnable() {
@Override
public void run() {
mywebhookthing
}
}```
does that look fine? it works
Yes
Nice, I am very happy with that thanks once again
quick question i want to whitelist people but for some reason it wont match their proper UUID but when they are in the server with the whitelist off and i add them it works properly?
how can i fix that issue?
is there a fix for DownstreamBridge <-> [lobby] - read timed out in waterfall?
magical? no
You'd need to work out why the proxy hasn't recieved a packet in so long that it timed out
damn, sounds like a lot of tracing. Im gonna assume its just my upstream as it only happens with people far away from the server. (15mb up, 200 down)
@brittle dirge did you correctly configure forwarding?
I think so
Does anyone know of any motd generator for waterfall?
no matter how I set there is always only one slot someone could help me with the problem
Show you config.
that's because you're pinging the proxy, you can set the limit in config.yml, iirc
bad paste
max_players: 0
what do you mean with that?
that's in your config
ofc, that will apply a hard limit on the proxy, so, might not work as you'd hope if you have things like full player override perms and such
other option is to use a plugin for dealing with the ping
there should be, for example, 0/100 or 20/100
the listener uses max_players
Do you have an MOTD plugin on the proxy or something?
https://hastebin.com/mowanevoyu.yaml what happened here?
What was it? I have it too and I wanna fix it
likely they misconfigured their motd plugin or something
Oh I don't have 1
your issue is some plugin apparently took too long processing events, use spark or a proper sampler or something
well, max_players in config.yml
the bug came from my server through a plugin @fresh kiln
best resource for install on pebblehost?
my waterfall gets fucky after a while. about an hour after players play on it, it just soft crashes and all players are timed out. i cant even restart the waterfall, i have to kill it. no errors in console except time out. how can i find out why?
Cut and paste all plugins to another folder and try again, if not try updating the jar file (might be corrupted)
If it works with all plugins removed, add them back 1 by 1
aight
So, I only left 4 plugins enabled:
TCPShield,
AuthMeBungee,
XeraBungeeQueue,
BungeeGuard
Issue happened again. Now I replaced every jar. Sigh
Weird
@tender axle run a profiler and see if youre having a memory or gc issue. I would assume thats the case
Yea that would work I guess
Does anyone know a good plugin for removing reacurring pings in the bungee console? i already looked the spigot page through and there doesn't seem to be any..
ConsoleSpamFix could have worked, but it doesn't support bungee so :/
ok so i tried to connect to my server using bungeecord and this happened: https://mclo.gs/OCfVRhw
server tried to connect to it but didn't recieve any form of response from the server
can you ping it from the server you're running the proxy on?
You mean this? https://www.spigotmc.org/resources/console-spam-fix-bungee.65471/
Because the reviews seems to say that it isn't working anymore.
oof, well, you can also just use a custom log4j config for it if needed
is that something built in already or?
well, yea, waterfall uses log4j, so you can just use a custom log4j config file and add some filters
not sure on the specifics there
"just", i looked at some log4j code - looks complicated xD
it's just xml
look inside your waterfall jar for the default one
(xml is a markup language, like yaml)
Is waterfall just like bungee?
waterfall is a fork of bungeecord with some improvements
(but let’s be honest, you should just use Velocity)
Is there anyway we could add the hoverboard where it displays the players in serverlist
without any plugin?
nope
what is the difference between modules and plugins?
technically, nothing, "modules" are just what the built-in stuff
hm ok
Waterfall seems to break VanillaTweaks Villager Workstation Highlight datapack, is there any fix for this?
@round pine disable entity meta rewriting and see if it still happens
That's found in the config for waterfall, I'm guessing?
yup
Any noticable differences with that disabled for other features?
Not really, miiiight break some plugins but most of the popular ones causing issues afaik have updated
it's in waterfall.yml
Ah. Did you set up information forwarding?
perhaps not, where can I find this one?
If you didn’t then that’s likely the cause
Did you enable bungeecord mode in the Spigot config?
ofc
and you enabled ip forwarding in the bungee config yml?
yes

@ electronicboy doesn’t disabling entity meta rewriting also disable tab list rewriting?
@round pine do you need to rely on any bungee specific plugins and if yes which ones?
disable_tab_list_rewrite: false
disable_modern_tab_limiter: true
use_netty_dns_resolver: true
disable_entity_metadata_rewrite: true
only these
I was gonna suggest you try velocity but it should be fixed with tab list rewrite disabled
ill try it
still no villager highlight. the station works fine with the previous change as well.
Well
Remove the plugins and try without them
I have a sneaking suspicion here
The bungee plugins I mean
Travertine?
alr
https://velocitypowered.com
Velocity is not bungeecord so it doesn’t run bungeecord plugins but it does the same thing; it’s also a proxy for minecraft servers
The ridiculously scalable, flexible Minecraft proxy.
same drop in as Bungee to Waterfall?
No sadly not; it’s a different system.
oh, well thats a no go if it cant run the plugins, those are needed.
unless I can find replacements of each for velocity
Excluding VentureChat and MC global bell all other plugins should have velocity alternatives versions of them
On those two I’m not sure
I can perhaps switch up venturechat, globalbell though would be nice.
I’m too tired to help gosh
Velocity also does not work, guess its just not compatible with proxy
Direct to server works perfect, just not through any of these proxies
nvm to both of those, its fixed
I really don’t get why. Velocity is the proxy I help maintain and there is absolutely nothing it can do to make game logic malfunction
Oh
This last time, it was shaders interfering.

otherwise, working with Velocity and Waterfall with those options set true
Okay then that’s that
You’re basically now running Waterfall in velocity mode
Which is kind of ironic

Oh well. If need be, I will switch to the superior platform when needed.
This should just work fine
I’m always trying to push velocity because problems on bungeecord tend to be on the stupid side most of the time so forgive me

Hey so when I enable ip-forwarding in my BC config it still says "If you wish to use IP Forwarding, you must enable it in your bungeecord config as well!" BUT I DID! Why is this?
Anyone know what the deal is with this?
https://f000.backblazeb2.com/file/grahams-public/f216c083/D914VMu5PE.txt
I have waterfall running behind HAProxy and it's working well... I'm about to get on a plane but I can help.
What have you done so far
A key thing if you want to keep source IP's between HAProxy and Waterfall is enable Proxy protocol v2 in haproxy and in waterfall.
In waterfall it's just called proxy protocol but use V2 in HAProxy
And are you using it via opnSense or pfSense or stand alone?
what are you trying to accomplish with haproxy?
Probably trying to firewall his servers but can't use NAT which is my case.
Can somebody please help me?
99% of the time, that means that you're connecting directly to the servers and not the proxy
99% user error
Rip.
Airee, either something is mangling packets or you're using mods, there is an option in waterfall.yml to disable metadata rewriting
listen minecraft
bind :25565
mode tcp
balance leastconn
option tcp-check
server yourBungee:25565 check-send-proxy check send-proxy-v2
Yep that looks about right
But it happens every time to everyone
Yeah it should be bind 0.0.0.0:25565
operator failure rather.
How to fix?
could you send me the IP people are connecting with
well, it's not erroring anymore.
[02:50:17 INFO]: [Airee|/:42059] <-> ServerConnector [build] has connected
[02:50:17 INFO]: [Airee|/:42059] <-> ServerConnector [build] has disconnected
not working either though
i'm pretty darn confused as to why this is happening. the server i'm connecting to (build) is far less complex than the other 4 on the network.
Hello, how many players can be handled by a single waterfall instance? (Running on a 16c/32t cpu at 4ghz with 8gb of ram)
Idk if i'm doing something wrong but as soon as I hit 1500+ players cpu goes to 100%
Well I would say the limit is right around where you are. Thanks to the network control thread; thats all a Waterfall instance is realistically able to do
But if you're really loading one instance this hard you might want to consider checking out Velocity. It should be able to handle more with less system strain
Bet the devs are really happy with competition advertising in the help channel...
they are, because if everyone used velocity than bungeecord/waterfall, everyone would be better off
If CPU on one thread is your issue, you could spin up a duplicate waterfall instance and then add a new A record accordingly in your DNS host...
I.e. minecraft.server.com A x.x.x.x 3600
minecraft.server.com A y.y.y.y 3600
is there a really good guide that tells you how to optimize (if possible) and setup waterfall? Im "new" with this and would like to learn
the bungee install guide, read the full thing and stuff it links to, e.g. the firewall guide
there ain't really much in terms of optimization given that it's just a proxy
okk thank you
How do i use SocketAddress to get the ip and the port of an paper server? I know i can use InetSocketAddress but he method is deprecated and I wonder if i can use SocketAddress to get the information.
check instanceof / cast to InetSocketAddress
ah ok. thanks
if it's not an InetSocketAddress, either choose to support DomainSocketAddress from netty, or just fail-fast
amazingly bungeecord deprecated the method returning InetSocketAddress
Well, if you wanna support SocketAddress as a whole you basically gotta
No idea how long Inet is gonna stick around for though
that's just great, because everyone is looking forward to connecting to bungeecord from a unix domain socket
so incredibly practical

I'm trying to run a Waterfall network thingy and every time I try joining (Everything is set up properly) "Could not connect to a default or fallback server, please try again later: io.netty.channel.ConnectTimeoutException" I have a good feeling it's because my Internal IPv4 address is the public IP too because of my provider
Does there happen to be a way around this?
use localhost?
Yeah, that's what I've been putting in for it
failing that, speak to your provider, not much we can do to help with routing issues on hosting providers
Alright, thanks anyways
What do you recommend to improve cpu usage with travertine?
disable entity metadata & tablist rewrite?
that should generally help, but, real thing is that the proxy shouldn't add on too much on its own, when you do get to the player counts where it matters, you're more looking into load balancing rather than anything
got 2600 online and ryzen 3600x is hitting up to 85-90% cpu with it
i belive i can improve it
@reef fulcrum is travertine build server dead?
Hey,
is there a site with optimized jvm flags for waterfall like aikar paper flags?
Or can all flags used also for waterfall?
pins
How do i do this with spark?
run it, make sure it's monitoring all threads, and then go look over what shows
i have players on dogshit mobile internet getting kicked even with playerconnection.keepalive being 120
whats should i look for in the documentation?
/sparkb sampler --thread * -timeout 60?
something like this?
maybe, depends what timeout does, if that's how long it should run for, make sure you can reproduce the issue in that timeframe
ahh
im gonna get one of the players with dogshit internet to join tomorrow and ill run this then
dogshit internet
I mean, if their connection is ass, there is only so much damage that plugins can do
yeah but my players dont care about having like 400 ping they just wanna play
cause india
Well, yea, but, network issues are network issues
ofc
keepalive is only gonna save your ass on keepalives, that option also does not touch the netty handlers, so read timeouts, etc, are not something you're gonna be saved from there
is it possible for minecraft to just not kick them
Not sanely
Like, you can remove the readtimeout handler, but, then you're gonna have players who DC who remain connected forever, etc
My server got hacked into yesterday and I want to ask does this plugin fix it? https://www.spigotmc.org/resources/securednetwork-renamed-ipwhitelist-1-7-1-16-protect-your-server.65075/
I heard about UUID Spoofing and stuff but doesnt waterfall fix that
We can't protect your backend servers for you
hence the bungee install guide having a firewall section
Configure your firewall properly or use bungeeguard if you don't have access to it
ok
Hi is there a way to connect vanilla 1.17 snapshot to bungee?
five miiiight have a PR on the issue tracker for that, but, we don't track snapshots
The last time I looked there was trampoline a fabric server mod to allow the proxy to connect to a vanilla like snapshot. Viaversion on proxy would be a temporary solution to allow the snapshot to proxy connection.
Yeah but i cant use fabric mods on spigot and paperspigot or pure vanilla that what is bothering me
Don't understand your problem.
- Fabric supports snapshots
- With trampoline you get support for connecting the proxy to you snapshot server (fabric + trampoline + Vanilla experience)
Fabric with only trampoline is pure vanilla experience. I never said to install mods that extends gameplay functionality.
I didnt understan the answer hah i m reading it on forum i understand now ty
I need to patch vanilla server to get it connect to proxy
Yeah, and that is done using fabric and trampoline here.
Yup thanx
how can i make a custom veryson of waterfall that makes a user 24/7 on server even while he is disconected from waterfall server
If he is disconnected he can't be on
To make a custom version, you have to fork the project in github, and learn how the patching system works.
i need to keep an account 24/7 on server while i'm disconected with my client
Why?
because the server has a queue
Asking about your attempted solution rather than your actual problem
ok
@zinc sundial that kind of thing isn’t supported and you won’t get any help with that Endeavour here
@round ocean Via on the proxy doesn’t magically make it support versions that it doesn’t have coded in. That only works on velocity (due to the way Via injects there)
@tidal musk https://github.com/Xernium/Waterfall/releases
And yes you’ll need FabricProxy
The latest version should work
And no ViaVersion+ViaBackwards will likely not work for the snapshot without some tweaks due to how Snapshot protocol versions work now
Thought they were up to date already.
Yes but Via doesn’t translate what the bungee sees
It translates the output data of the proxy
On velocity it translates the input data afaik
Via on Bungee too?
On the bungee: client (version a) -> bungee -> Via on bungee -> server (version b)
So no; adding Via to a bungeecord instance doesn’t magically make it compatible with what Via supports
That's meh.
I remember that I just dropped via on Bungee 1.15 and ran a 1.16 snapshot with fabric but could be the time where bungee got an unofficial update. :man_shrugging:
If that did happen chances are that I did that
does anyone know where to find the waterfall modules?
wdym, find them?
I deleted them :/ I need to get the defaults one like the /server /find etc again
hey everyone are there any plugin about bandwidth limitation for each ip connection?(or per user basis)
anyone know what could be causing all of the players on my network to get kicked? It's happened twice so far, and when it happens there's a ton of stuff in the proxy console all saying that that there's a ton of requests coming from hundreds of various IPs. CPU usage doesn't get that high and it's not out of ram. Anyone able to help me with this? Here's what the bungee console looks like btw, none of the IPs belong to players:
bot spam
your proxy is overloaded, apparently, or you've at least got plugins doing stupid stuff
ah rip, was thinking that could be the case
bungee is a massive architecural flaw which I can't fix without breaking a ton of shit
it's definitely not overloaded, could be a plugin though
using your firewall to limit new connections in a timeframe is a good idea, as is scaling out horizontally
and, yea, there is defo.a plugin eating time in there, hence the nature of the message
are any of these on your no-no list by any chance?
as for why, god knows, but all that crap happens in the event loop which blocks other connections processing in the meantime
most recent one is vivecraft extension, might want to remove that
Hi, just wanted to ask if it's possible to run a forge 1.14.4 mod server and a paper 1.16.4 together behind a watefall?
No.
Forge 1.13 and newer is incompatible with Proxies
That might change in the next few days
Oh okay
So in the really near future? Then I would just wait.... do you have any more information about it? Like a PR or Issue or so?
Since I’m the one developing it
I mean
Velocity will have it first
I’ll port it to waterfall later
Nice
Is there a way to send a console command from my Bungee plugin to a specified instance server console?
How do I wget the latest waterfall jar now? I can no longer wget https://papermc.io/ci/job/Waterfall/lastSuccessfulBuild/artifact/Waterfall-Proxy/bootstrap/target/Waterfall.jar
Check kashike's post in #announcements
so how can i disable the versions motd at server list?
or what is it
nevermind i found out how 😄
the correct way is to get https://papermc.io/api/v2/projects/waterfall (versions, the highest) then get request URL
https://papermc.io/api/v2/projects/waterfall/versions/$highestversion and from there get the build number then https://papermc.io/api/v2/projects/waterfall/versions/$highestversion/builds/$buildnumber and then get the downloads.application.name matches waterfall* and then finally get https://papermc.io/api/v2/projects/waterfall/versions/$highestversion/builds/$buildnumber/downloads/$jarname
or if you want to use v1: https://papermc.io/api/v1/waterfall/1.16/latest/download
be warned ^ breaks without newest game version
@craggy solar
IM using paper 1.12.2 but the latest waterfall?
https://papermc.io/api/v1/waterfall/1.16/latest/download seems like it would work with wget
No need to get the latest version number then
yes I mean that will stop working as soon as 1.17 is released
the first way is the better one but stupid to set up in a script
Just have to update my script links now! ;p
Yeah same, I can easily wget a url lol
Dont want to have to parse JSON in bash just to get a god damn download
I'll write a small python script for that eh
sweet
Small python3 script for downloading the latest versions of Papermc projects using the v2 API - paperv2-downloader.py
@craggy solar try that if you want
Hello, can you send me the download link for WaterFall which makes minecraft version 1.8.8 to 1.16 please ?
here you go
btw you will need plugins to join with differnet versions
https://jo0001.github.io/ViaSetup/?lang=en use this to know what plugins you gonna need for versions you want to be supported
Is there a way to send a console command from my Bungee plugin to a specified instance server console?
yes
can you do paper plugins on waterfall?
Yes.
no, you can't. a paper plugin won't run on waterfall because they are completely different server software
🤯
Heyo knowledgeable waterfall people :) I have another very specific question which I keep having problems with.
I have one physical server with a dockerized/containerized environment (different containers running on one machine) and I've set up my config.yml by providing internal DNS names (like survival1, skyblock1 etc.) which then just gets resolved to the IP. The problem now seems to be that if IPs change as they are assigned dynamically when the containers are restarted, waterfall doesn't know where to send people to anymore.
When doing "/server survival1" it's saying "Could not connect to a default or fallback server..." but I don't even see an attempt to be connected to the survival server.
Just wanted to make sure that assigning static IPs to these containers would fix the problem. I wasn't able to verify if the IPs actually change when the issue is happening, it's just my guess and that's why I came here, to verify.
(The waterfall container can also certainly resolve the domain "survival1" just fine which I tried by pinging 😄)
Ah so it could still happen but that might help as well? :)
Will that ever be an issue if I add external servers in the future? Just asking for future proofing ;)
It's dns
True :D
not exactly waterfall related, but what are the name of the bedrock proxy servers?
not like waterdog, the native ones
tried asking on some bedrock servers but they're all dead lol
@tidal musk Alien is the only one I remember. And then there is Nemisys
The second one seems to be the popular goto
Hey, what is a good way to protect my waterfall?
for example players will not be able to bypass ports etc.
@tidal musk Use a firewall
bungeeguard
i am now getting a weird error on my hub
not sure if i can post it here
Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.
generally bad JVM/hardware, check ram
(either that or something installed on the server mangled a class)
Well, that excludes, potentially, bad JVM
so startup flags?
i tried disabling paperbin
maybe this will help
still, it must be hostings fault, because problems doesnt exist on my other server on different machine
with the same jar
yup
it was startup flag
Lemmie guess, you removed stuff like Xmx
Well, I mean, Xms == Xmx
or the preallocate flag
no
i had two jars
paperbin.jar paper.jar
maybe because i dont have -noverify flag?
that flag is bad
why?
Since when does disabling a safety mechanism of something sound like a top notch idea...
The only argument for that flag is when you've got shit corrupting stuff but you wanna play the "i don't wanna fix it" or, "I know why my stuff is corrupted and I'm willing to live with that"
tf is paperbin
are there any guides for switching from a standalone server to waterfall? or is not really that complicated?
just follow the standard install guide
(guides are on spigots wiki, it's really just like flipping 2/3 settings)
okay thanks
Hi! Is there a link to directly download the latest and not a specific version?
(for Waterfall)
please?*
No, but sulu did have a script somewhere to pull the latest build from API
Either use something like this https://gist.github.com/Xernium/91d67367dbcfef4b18fbe91b74324e06 or the v1 api https://papermc.io/api/v1/waterfall/1.16/latest/download
Small python3 script for downloading the latest versions of Papermc projects using the v2 API - paperv2-downloader.py
v1 API is deprecated
Encouraging new people to use something we're planning to get rid of in due course is generally not a good idea
Ok but why the latest download for waterfall of the v1 is deprecated (sorry for my bad English I'm french ^^)
the v1 api is deprecated
And there is no easy one click link for v2
v2 solves a few issues we had with v1, and we removed the latest thing due to it causing some headaches, we generally do not recommend auto updates in any capacity
Ok ^^
@ zzzCat will you remove v1 as soon as 1.17 rolls around?
maybe
Don't think we can really do that, it's just pulling data out, it's just an endpoint we have to maintain which eats up extra resources and such
that command is deprecated too, just so you know
update the bloody readmes to point your download links to the bloody new download page, or the docs
don't f* leave pointing to a dead f* link
ik
Project Travertine build #168: SUCCESS in 1 min 59 sec: https://papermc.io/ci/job/Travertine/168/
Project Waterfall build #390: SUCCESS in 3 min 57 sec: https://papermc.io/ci/job/Waterfall/390/
👀
I have a network setup with 4 servers, three 1.16.4 paper vanilla and one 1.12.2 forge/sponge. I am trying to restrict the forge server but waterfall isn't giving me permission nodes for luckperms.
Do I use bungee's permission nodes or are they waterfall.command blah blah
Everything is under bungeecord.command.<command>
It doesn't work that way for me, does luckperms have to be configured in a database?
Nevermind /lpb was a thing.
first time using a waterfall proxy. how does the plugin setup look? do i have to install for example essentialx or luckperms for example on the waterfall as well, do i have to install any plugins on waterfall server at all?? all tips are appreciated just don't roast me too hard i'm new 😄
Plugin install works the same way it works on paper/spigot.
However plugins made for spigot/paper won't going to work on your proxy, because waterfall and paper are two different server softwares
The plugins you can use on waterfall are plugins you can use for bungeecord.
You can find them in the spigot bungeecord section. Some plugins like luckperms has a version specially made for bungee
alright, so there are actually plugins for the proxy server. are they necessary though?
do you personally use anything?
It depends what you use the proxy for, but there are useful plugins like luckperms bungee that ads an option to set perms with server contexts etc.
Look around the spigot bungeecord section, and you will find useful/fun ones to mess with.
I personally use Velocity currently, but yes I used to use a ban manager, an antibot, a motd changer, luckperms bungee, tpcshileds bungee plugin, and a few stuff.
If you are starting a network try looking into Velocity, it's better than waterfall/bungee. It does not have many resources avalivable yet, but that will change as time goes on
why do you prefer Velocity over waterfall/bungee?
Because it preforms better, easyer to use, has many useful features, and it's not whritten in crap unlike bungee. You can learn more about it in the velocity doc.
how are the plugins there?
typo medley
anyways, proxy plugins only matter if you wanna do things at the proxy level
The installing process is the same.
i don't think Velocity is the best option for me. it's more resource hungry which i don't have a lot of
Nope, velocity is more lightweight than bungee
man its soo hard to decide such kind of things
If you need performance and innovation go with velocity, if you like having more resources avalivable and a larger community behind you go with Waterfall
larger community we talking about how many exactly
By large I meant the paper community and the spigot community
oh
Velocity has hunders of users too, but community size souldn't matter, more like how many resouces/guides/support you find for the software you use
Velocity has better performance than Waterfall for sure
i was thinking of sticking with waterfall and later change to Velocity, but in later stages it might be harder to transform into Velocity. i appreciate well written guides and support
I lately changed my proxy to velocity from waterfall, it wasn't hard at all, all I had to do is wait for my nececarry plugins to add support for Velocity, or find a better alternative.
well then i should stick with it
oh by the way
do you have any sort of authentication on your server?
I used to run a cracked server, but I don't do it anymore. I used authmereloaded on the backend server, and authmebungee in the waterfall instance.
Only you can decide, read both softwares doc, spend a bit of time on their forums, and decide after
yeah saw authmebungee as well when i searched on google
Authmebungee is an addon to the backend server version to prevent bungee related commands to be executed before login & don't allow the user to join any servers from the auth server if he isn't authitenticated
wait by backend server we are talking about the proxy server right?
Nope, the server behind the proxy
you mean the vps?
AuthmeBungee is an addon to AuthmeReloaded
No, the backend servers are the servers behind the proxy.
You join the waterfall/velocity instance, and you get sent to the spigot/paper server that you specifyed
Player -> Proxy -> Backend Server
Lobby, PvP, Build are backend servers
ohhhhhhhhhhh
They are behind the proxy
i get it now
its called backend because its backed by the proxy
so the proxy/bungee is the "frontend" because thats what the player uses to connect to the back ends
Exectly
aka that's what the player sees
Yep. And the proxy manages where they get sent to upon joining, how they get sent to, and other stuff regarding their connection like packet compression
yeah it's a lot more clear now for me when it comes to those terms
so authmebungee prevents users having access to backends unless they are authenticated in the waterfall
makes sense now
Not exectly.
Authmebungee prevents you to enter other servers in the proxy, from the auth backend server.
but
Becuase authme only can be installed in a backend server
thats only if the backends servers have ips right?
if the backends are local then they cant connect to them?
because they don't have ips
proxy: actualip:25565
lobby: localip:25566
survival localip:25567
If your backend servers are running in the same machine your proxy do, setting the backend server ips localhost in the bungee config is a good idea, as nobody can join directly to the backends, only thru the proxy.
exactly
Proper firewalling is never an issue anyways
Don't forget to close the backend server ports using a firewall
one moment let me try closing them real quick
i remember i closed them then the waterfall failed to move the player to them
If you are using ubuntu, starting with ufw isn't a bad choice, check out a tutorial if you are new to firewalls
ye i use ufw
Must be a misscofiguration, if someone joins your proxy, from now on you can send them anywhere you would like
yeh because it should be handled internally
Btw, spigot has many really useful guides, it worth reading a few of them even if you don't have any issues.
mind linking them?
i've learnt a lot from you so i will surely grind through the guides
Sorry I can't really do that right now, as my internet isn't capable of loading Discord correctly, not a website, but you can find good ones by going thu the treads people made.
.g guide website:www.spigotmc.org
I mean that was a stupid search, but you get the point.
yeh i can just go on the spigot forums will probably find some there
man the optimization helped with the timings
Same, it really does
Close them :P
Your proxy config is missconfigured.
First, may can i ask, all of your servers are running under the same machine?
yes same machine. using pterodactyl
Ooh, pterodactyl is a different story, but let's find the cause of the issue. You have your lobby server running, under port 25567, and added to the bungee config like the following:
lobby: localhost:25567
Does your lobby server has bungee mode on (spigot.yml), has the server-ip: localhost, and ths server-port: set to 25567?
The backend server has to be in offline mode as well.
they all are
just gonna test localhost in the bungee config
although server ip in lobby config is 0.0.0.0 but its the same right?
Don't forget to add your lobby server into the priority list as the first entry in the bungee config.
No. 0.0.0.0 means all interfaces, you don't want that. Use localhost.
ohhh
then that was the issue let me try that
now its localhost both in bungee config and lobby/survival config
Sounds good.
Make sure you join thru the ip/port of the bungee, and not trying to join thu the adress of the lobby server.
There's also an option to always send users upon joining into the lobby server, I can't remember it's name, haven't used bungee in a week
i change server-ip in server.properties for the lobby to "localhost" and after i try to connect to it it changes back to 0.0.0.0 automatically
That's the limitation of the panel.
Check your panels settings, there must be some settings to disable the 0.0.0.0 change.
found it
I'm more experienced with using the panels than configuring them, becuase personally I never used them for my own servers.
Nice 👌
and that should go in waterfall settings as well
Yep. The only setting where your ip and port must not be localhost, is your proxys adress.
What is the issue you are getting?
[22:43:56 INFO]: [Ryanize] disconnected with: Could not connect to a default or fallback server, please try again later: io.netty.channel.ConnectTimeoutException
It you are connected to the proxy, but can't connect to the lobby, that means you didn set the send to x server to true in the bungee config, and didn't set up the priority list right.
.g bungeecord config explained
The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy.
here is my config
Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.
which part am i missing
Looks fine for me, nothing really.
The issue is must be with the Lobby server than I assume. Could you please double check that you have the same ip/port in the server-properties file, that you have in the bungee config/ that you have configured with your panel?
Try 127.0.0.1, instead of 127.18.0.1, I don't think the second one points to localhost
Have you fixed the issue? 👀
Try using ssh with screen/tmux, a bit more easyer than using a panel 🌮
easier to manage all the resources for a newbie like me
the issue is the ports
The open-source server management solution.
following this but nothing
I don't think the server.properties thing is what you're looking for
I just allocated the 127.0.0.1:port IP to each of my backend servers and left the server-ip=0.0.0.0 in the properties
are you using pterodactyl as well?
Yeah
how does your firewall and port forwarding look like
oh and ye what about your bungee config
did you use 127.0.0.1 or something else?
The only ports exposed to the internet are the 25565 (for the proxy) and a few misc. ones for plugins and admin stuff (like SSH/pterodactyl web server), along with the one command you specified in the guide you linked. The servers themselves are allocated the 127.0.0.1 IP but the config file uses 172.18.0.1
I'm very rusty on my firewall knowledge at this point but doesn't that use a top-down priority? So would blocking everything using ports 25566 and 25567 first prevent it from respecting the new rules you added with pterodactyl0?
I didn't mean you had to delete them entirely, just put them lower on the list so they're processed later. But honestly just try what you have there real quick because that'll tell you if it's a firewall issue anyway
ye
will try it like this
this will determine if it's a firewall issue or not
ye i can connect now when they are not denied
Do you still have the default rules set to deny all incoming connections and allow all outgoing?
Because you wouldn't have to directly specify denying the ports anyways
The backend servers ports must be closed, Ptero has issues if it requires you to open those ports lol
i did not change anything regarding that
At the top of the ufw status it should tell you the default rules set
maybe it's ufw status verbose
Yeah all your ports are sent to deny by default so you don't have to explicitly set it
omg
its working
it was the priority
now that they are moved down its working fine
ahh im so happy
learning by the moment
appreciate the help really, thank you guys
i promise i will read the spigot guides as well Simon haha
Glad to hear that it works ❤️
Great 😄
its 1 am here gonna go to bed. been a long night, optimizing, aikars flags on the servers and learning a lot here
have a great day/night
Go to #paper-help for that question
Have a good night!
Glad you got it working!
❤️
I am trying to start waterfall for the first time, and am getting this issue when I look in latest.log, also, no window opens when I run, a console dialog flashes on the taskbar for less then a second before going away. Error is here: https://pastebin.com/nKbDpK7C
Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.
Something apparently managled a packet or sent something the proxy was not happy with
We do not support protocol hacks
Waterfall) The command “/send” doesn’t work.
what i have to do?
@tawdry tundra does it say command not found or no such command? In that case check to see if the cmd send jar is present in the modules folder. If it’s not then please download and update to the latest version of waterfall. When you first start it after updating it should download the missing files
no
/server command works well.
/send command doesn't works, saying I haven't permission.
@royal hawk
I'm using the latest version.
That would imply that you don't have permissions then
Do you have a perm plugin? if so, manage the perms in there, otherwise, make sure that you're in the admin group in config.yml
i put luckperm plugin in waterfall plugin folder
but it doesn't work
I'm in admin group
and...can i use Userlogin plugin?
Well, the admin group is useless if you're using luckperms, you need to configure your permissions in there
.g bungeecord commands
(DiscordBot) https://www.spigotmc.org/wiki/bungeecord-commands/
The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy.
thanks
UpstreamBridge has disconnected
my computer(server) can log in
but other users
can't
upstream bridge means that the connection between the client and the proxy died
specific error matters, but, that's generally networking issues
If i send you my config.yml, can you fix it?
ServerConnector [server] has disconnected
UpstreamBridge has disconnected
Error enabling plugin LuckPerms
Forced host server pvp is not defined
ServerConnector is the logic that connects players to servers, check your server logs, see if there is anything useful there as to why it disconnected them
EOF implies something closed the connection in an unexpected manner, you'd need to work out why
Then why can I connect to server?
if you're home hosting, I'd guess it's a "you're not connecting over the slow internet connection"
thanks
discord connection, Skript, Userlogin, Worldedit
Using FastUtil Long2ObjectOpenHashMap for block connections
Shouldn’t have an impact on networking
I am already gonna say; that plugin has nothing to do with it
Hey wan't to ask has anyone here used Bot-sentry in their waterfall server and has it done it's job?
2ls antibot works fine for me with bungee, for velocity epicguard does the job
alright thx
I am trying to setup different subdomains for each server on my network, and the internet keeps saying "point the domain at the waterfall server" when I don't know what to do to achieve that.
I use google domains and all the other things use godaddy or other systems.
Any help?
@prime ledge if your server port is 25565 an A-Record to point at the server IP suffices. For anything else you need an A Record to point an SRV record to
I've been trying to find one explaining google domains, they don't have the openness of godaddy or other domain panels.
Google should have a Documentation of setting up DNS records
DNS records fundamentally work the same everywhere so if you can’t find anything for google DNS specifically it’s most likely similar
.g google domains minecraft server
The bot seems to be able to find things easier than human eyes, as I scowered google for something and got nothing for google domains