#waterfall-help
1 messages · Page 34 of 1
Which log do have i to send? it is without any plugins and mod still
still timed out with timeout: 300000
either with disabling entity metadata rewrite
can't really say, depends where the timeout is
if it's on the server, you'd need to boost the timeout there, no idea what thermos lets you do but forge does have a system property for that
if it's the client doing it, "good luck"
on the hub console it just appear "Pendragon_Altria lost connection: Timed out"
on travertine console: <-> ServerConnector [ThermosHUB] has connected
-> UpstreamBridge has disconnected
<-> ServerConnector [ThermosHUB] has disconnected
how?
There was a fml system property for it somewhere, no idea if thermos supports it though
Well, UpstreamBridge is the client closing the connection too
You'd basically need to work out what is causing them to both stop talking to one another
run wireshark over that stuff and see what goes wrong
may some networking went wrong
@Hyronymos 🥝 it's the outgoing traffic from my waterfall showing that even 13 players don't use 50MBit/s, lol
Hey, I realise it's not exactly a waterfall question, but does anyone have a waterfall/bungee network recommendation on a plugin that lets players keep their inventories when switching servers?
https://www.spigotmc.org/resources/mysql-inventory-bridge.7849/ seems like a good resource
locking essential features behind a paywall
I just use my own solution because I couldn't find one that suited my needs, lol
FOSS is still the best :)
yeah, all my stuff is ;D
Hi, I need to test some of my plugins live. Could it cause any problems when i open the firewall for my ip so some of my bots can connect directly to a server in offline mode while the rest runs on Waterfall online?
if you don't have your own IP then someone who is on teh same IP by your ISP could connect to it I guess... but imo that chance would be pretty low, lol
Yeah that is not an issue, but it won't fuck up waterfall or the server?
You could just open the port of the offline server and join directly
But just.... Don't forget you did that because it's terrible for production
Well yeah, I know, what i mean, could it somehow disturbt the proxy
A never mind, I will just try it
Ty
you probs wanna make sure that people can't connect to it through the proxy if it's a public server, but, not really so long as it's configured properly when it's shoved back in
Ok ty
How do i set a fallback server?
you add it to the priorities list
what
fallback no longer exists
Now the priorities thing is a list of servers that it tries in order
ty
and is there a way to send someone always to the lobby instead of to the server in locations.yml
yes, force default server
ty
.flags
Optimized & Recommended JVM Startup Flags: https://mcflags.emc.gs
Hello there :)
These flags are optimized for waterfall too ?
use the pins
i recommend G1 under the whole "it's well tested"
ZGC can work but has some caveats
Ok thank u 🙂
Hi, I don't know whether to write it on paper or waterfall, but I will ask the same thing. A while ago I used paper for my Minecraft server, but now I am creating a network, when using the AuthmeReload it does not log me automatically when changing servers, so I was reading in the papaermc.io forum the problem is with the network updates, has anyone had this problem that can guide me?
sorry the length
iirc cracked networks have authme on the lobby server and run no authentication on the backend servers
aside from ip-whitelisting the proxy
so they connect through the proxy, get put in lobby, are required to authenticate, and then can select the instance to connect to
Do you recommend flamecord for security against crash etc?
all of those anti-crash waterfall forks and related plugins tend to be of low quality
I would avoid them
As "exploitfixer" really helped me against a kid crashing my server, I thought that switch to flamecord was also a good option ^^
That's why I ask if your get feedback about it, but thanks for your answer :p
though with the quality of some of the patches, I wouldn't have accepted them into Waterfall were I still maintaining it, let alone Velocity
I mean sure, but at least we'd now a bit more details about what the actual crashes are, lol
there's a reason why I went independently with Velocity's anti-exploit patches
some of the improvements I just gotta get to, some of them are really more of a ballache to deal with with bungee
hello, I still have the same problem with 1gb of ram, after 1-2 hours impossible to connect .. logging in infinite and timed out and yet it puts the player and connected. I use DynamicBungeeAuth for protection and even with 3 plugins it does by cons with 0 plugins no bug, but this bug occurs with any plugins not one in particular.
i send this when it will do me
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
Hello, I want to change the "host" in the config.yml but it doesn't work. It always reset to 0.0.0.0
some control panels do that automatically
if you're on ptero, you don't wanna set that anyways
I'm on ptero but i need to change that :/
You don't
The things in ptero are isolated to their own container with their own network
the IP address is set on the docker container itself, so, that's 100% normal
howdy folks, doe any of you know a decent plugin that allows you to export metrics to prometheus for waterfall
how to do it????
end
Are you trying to shut it down or something?
I got an error on player login
anyone have issues getting some (like 50/50) vote/list sites to get # of players online from waterfall? (Do you just point them at the query_port of any of the servers?)
Do you know if I can change "Already connecting to this server!" message using waterfall?
Good afternoon guys
We're running the latest version of waterfall
And players have been having a really hard time to stay online or to even connect to our server
they mostly get disconnected while loading our lobby or will get randomly disconnected inside servers
We've been losing many players because of that, it it happens to everyone, even to me
I mean, you literally just came from #paper-help with logs where your server is stalling...
I haven't been able to log in for 2 days now
Waterfall is far from the issue here
Yeah, we're getting issues on both
One is crashing
and waterfall is have way too many read time out
as the image shows
read time out literally means "we weren't sent any packets in so long"
😦
Given that the server is stalling and even crashing, that's like... 100% expected to happen there
since it's a proxy, it hanging for 30 seconds would naturally kick out any players connected through it
So probably not a waterfall issue?
Also, what is the recommended Java version to run under waterfall
latest or LTS
How does one change the port for a bungee/warterfall server
bungee.yml -> listeners section -> host
Im using waterfall and I dont see this
disable_modern_tab_limiter: true
log_initial_handler_connections: true
throttling:
tab_complete: 1000
game_version: ''
disable_entity_metadata_rewrite: false```
thats what I got
config.yml
oh woops yes config.yml
:port
Guys, many of my player are getting disconnected out of nowhere and its happening pretty often
How can I optimize my bungee to improve their connectivity and decrease connection loss
I reviewed the config with my fellow co-workers and we didn't find any abnormalities
You need to work out /why/ the connection between the two is breaking, check the logs
if it's still the "read timed out" thing... that's the server
Im downloading the most recent log to check it out
This is what the log shows for the one of the players that keeps getting disconnected
Connection reset by peer
The client closed the connection in an unexpected manner
So it is a client error?
Because its been happening a lot to many people, so it draws my attention to being an error produced by our side (the host)
can't say
We just know that the client closed the connection in a manner that wasn't expected
could be that the server didn't send something in so long causing a timeout
Any suggestions on what to do or change?
if you're still having server issues, "fix them", basically
I think they where talking about some stuff in that area, so 🤷♂️
I get this error whenever I start up waterfall
[17:02:39 ERROR]: Exception in thread "main" java.lang.IllegalArgumentException: Server lobby (priority 0) is not defined
[17:02:39 ERROR]: at com.google.common.base.Preconditions.checkArgument(Preconditions.java:359)
[17:02:39 ERROR]: at net.md_5.bungee.conf.Configuration.load(Configuration.java:160)
[17:02:39 ERROR]: at io.github.waterfallmc.waterfall.conf.WaterfallConfiguration.load(WaterfallConfiguration.java:49)
[17:02:39 ERROR]: at net.md_5.bungee.BungeeCord.start(BungeeCord.java:279)
[17:02:39 ERROR]: at net.md_5.bungee.BungeeCordLauncher.main(BungeeCordLauncher.java:62)
[17:02:39 ERROR]: at net.md_5.bungee.Bootstrap.main(Bootstrap.java:15)
It says it right there
You have a priority set of a server that doesn't exist
You changed your server names
I have no idea how to set up a priority
Update the priority list
It's in the config
Defaults to lobby or whatever
Just change that to the name you changed the server to
hub:
motd: '&1Just another Waterfall - Forced Host'
address: 144.217.73.93:56969
restricted: false```
this is what I have
i see
soryy I have not setup a proxy before
Then change your priority from lobby to hub
So now when I try to join
says its in online mode
Ive turned online mode off but still
You join through the proxy, not the backend server.
I joined off the port in the config
prevent-proxy-connections=true I had this in my lobby config
Must be it yes?
nothing to do with that
"server in online mode" means that the server sent a packet that it only sends in online mode
e.g. the server is in online mode, you're connecting the proxy to itself, you have a really janky plugin on the backend, etc
Ah so do I just connect to the lobby server
Sorry im just hella confused
I have no idea on how to fix
so I fixed the error with online mode
Now I have this If you wish to use IP forwarding, please enable it in your BungeeCord config as well!
Ive enabled this in the waterfall config
sounds like you're connecting directly to the server and not the proxy
Im connecting to the ip off of waterfall
It'll say that if you didn't set bungeecord to true in the spigot.yml of the backend server.
Its set to true
The error im getting is:
If you wish to use IP forwarding, please enable it in your BungeeCord config as well!
DO I connect to the lobbies ip or the proxies?
Then set up forwarding in the config
[10:23 PM] zzzCat: sounds like you're connecting directly to the server and not the proxy
it's a proxy
it proxies connections
So I have all my waterfall configs setup correctly. Im being told I have to change settings in paper.yml, bukkit.yml and spigot.yml
I just have no clue on what I need to change
server.propeties: online-mode=false, spigot.yml: bungeecord=true, paper.yml bungeecord-online-mode=true (or something along that line)
Thank you 😄
you should also make sure that nobody can reach your paper servers (e.g. by running on localhost or using a firewall)
I mean any linux distribution normally has ain inbuilt firewall
but that doesn't matter as you need to set it up to match your setup to only allow access to your paper servers from your waterfall proxy
but if you run everything on the same machine then simply listening on localhost for the paper servers should be enough
install microsoft windows firewall
on your minecraft server
get it now free of charge from spigot
So I was finally able to join
And I can switch between servers
but My fiend cant join
nvm lol
Thanks got everythign working
I have one other question
Any idea on how to set it up so when I go back to lobby it brings me to lobby spawn
Can I make it so everytime you log back in it brings you to the lobby server as well
for the first one you'll need a plugin on the lobby server
the second one is just setting force_default_server to true in the config.yml of the proxy
pretty sure i did all the settings correct, but getting 'Could not connect to a default or fallback server' when trying to connect. my ports all check out, and seem to work. as far as I can tell i have a priority server setup in the bungee config
cant think what im overlooking
video someone linked to me shows a default, and fallback server setting, but seems thats been replaced by priorities?
just realized, i may need to have more than one listed there
generally ranges from "is the server running", to firewall/networking joys
Only needs one
hmm
Just, if there is more than one, it will try to run down them
how are you running the server?
not sure i understand the question?
are you using a control panel? dedi? shared host? etc
i mean, i have config to use waterfall on main port, instead of lobby using my main survival, and then a minigame type server as a third
dedi
using ptero or anything like that?
no
i hjavent even setup iptables yet since im just trying to figurte this out first
General thing there would be making sure all the servers are running on 127.0.0.1 bar the proxy, and configuring waterfall to connect to them over that
ohhh damn, i did a stupid thing and forgot to remove ip from my server p roperties (since its an imported world from another server)
Hello guys my module /find doesnt work some solution? thanks
what issue do you have with it?
wrong channel
Good morning everyone! I'm having a small issue. So I have 4 servers, (one being bungee) and I have Dchat and Warpsystem (premium) on the servers, and basically what they are supposed to do, is that the warp plugin allows /tp to other servers without ease, and I mean to TP to players on other servers. It works on one server, but my skyblock server, it does not work so much. I can use portals and such to connect to other servers, although, I cannot connect warps around the servers
Is there anything I can look at to help fix these issues perhaps?
no
Hello everyone ! I looking for the waterfall documentation and I didn't find it on the web site.
setup is the same as bungee
ok thx
Sometimes the "end" command in console just doesn't work
And when I just use the stop server button on my server host it stops the server with the same way as normal paper servers and then it's bugging out
must be on your host side
When there is a lot of players on the network I am on, (>200) all the servers on the network (some are on different machines) have High TPS, low ping but theres a lot of jittering noticeable when people move (doesnt happen when theres less than 200 on usually). Im using Waterfall. Any ideas the fix to this or what could cause it?
Do you think its because my Proxy is only 1GB maybe? Should it be more I read online if it has plugins on it etc it should be 2-3GB for a few hundred players
@grave thistle Are all the players in one place?
Nope
Hello there! I'm having an issue where the client side doesn't respond at all upon joining for 5-10s. The thing is that when joining and typing in chat while in that period, everything shows normally on the console. I have tried removing all the plugins, issue still present. Direct connection to the server works fine, so it is the proxy.
It has 6gb allocated, other servers running on localhost containerized using docker
What is an ideal chat plugin which allows DiscordSRV to still work but links chats between servers?
Good morning everyone! I'm having a small issue. So I have 4 servers, (one being bungee) and I have Dchat and Warpsystem (premium) on the servers, and basically what they are supposed to do, is that the warp plugin allows /tp to other servers without ease, and I mean to TP to players on other servers. It works on one server, but my skyblock server, it does not work so much. I can use portals and such to connect to other servers, although, I cannot connect warps around the servers
Is there anything I can look at to help fix these issues perhaps?
does viarewind and such work with waterfall?
yes
tjat
2LS one on spigot I’ve heard works fine
any1, can we revoke gemstone on mmoitems premium plugin?
ask the premium plugin author?
what netty exploits
Response please I'm in hurry
You wanna know what would be faster?
using the search...
there was a netty issue, that's fixed
netty was not really the real issue, however; that's just people having no clue
So there is no reason to use flamecord.. thank you
Heya all of a sudden the playercount has stopped working lol, doesn't go above 0, no plugin has control over it so im a lil confused haha
Travertine latest
generally down to ping passthrough getting the count of the first server, or, god knows
Ah that may be it i did screw with tha recently
Would you mind if i sent you the config via hastebin because i can't seem to see any issues
The thing is that ping passthrough will only cause it to get the count from the first server
Ping passthrough is enabled
Would the query be an issue since i recently enabled that on all servers
no, nothing to do with query
Quite simply ping passthrough passes the ping through to a server and gets the count from that
Downfall is that the player count will only reflect the first server
in many cases, an MOTD plugin on the proxy is basically essential, or, disabling ping passthrough and just doing it on the server itself
Wouldn't disabling ping passthrough cause more issues though :p
no
Just, you'd have to deal with MOTDs on the proxy itself
There is no perfect solution to all this mess
MOTD's is easy i can just use SLP
well, not MOTD, but the thing on the server list; but, yea, just dealing with all that on the proxy is the best option overall, only way you're gonna get accurate counts anyways without more messing about
Thanks that worked
no
):
I did consider allowing that, but
- You have to handle UUID/IP forwarding
- Muuuch more state mess to deal with
- There is often no real actual gain
fair points
I like the fact that waterfall is a block of water
actually, LEGACY_STATIONARY_WATER
tell its ordinal as well for sure
use multiple front facing WF's and use round robin DNS OR PROXY protocol and put nginx in front of them
is there any anti ddos script for linux ubuntu?
Yeah, no. You can’t do DDoS protection on a single host and those scripts are nothing more than snake oil.
ohh okay
Just get a host that does DDoS mitigation, or hide your server with a DDoS protection service
Aikar, do you have any recommended flags for bungee?
I found the server flags, but didn't find anything for bungee
pins...
is there any anti ddos script for linux ubuntu?
@violet shard Use cloudflare spectrum
Yall not able to pay for spectrum or something?
what a shame
What is the price of spectrum? There is no concrete price there.
Would be nice to have it (:
Furthermore.. if player uses Cloudflare spectrum, wouldn’t be the ping to waterfall worse?
- 200$/month for the business plan
it's literally useless for the majority of servers
Like, sure, if you're hypixel who are literally sitting on a metric fuckton of cash, sure
For your average server, $1/G is fucking expensive
not to mention all of the limits...
This might be a stupid question but what is the average usage (GB) of Minecraft for that?
and, yes, tunnling your connection through some infastructure elsewhere on the web will increase latency, there is some "we have better routing" which might help, but in general is a load of horse manure
I have no idea if they charge for outgress or just ingress, but, I can say that I used to often hit a few hundred GB on a server with like, 40 players, majoritly US only
Q: If I code slow logic(e.g. access to DB) in Command::execute. Will it make any lag about player's connection?
(bukkit is yes, but bungeecord is..?)
It'll lag part of the connections. You shouldn't do it in BungeeCord.
Netty is built on being non-blocking, you'd be best to, I don't know, not block the event loop.
If it work in non-blocking. I think we can block any event.. is it wrong?
Don't block. At all.
A couple of events extend AsyncEvent so you can safely block those
But you have to still do those blocking actions on a different thread
All that those events do is just defer whatever action happens next until plugins are done processing
I see. So, You mean that BungeeCord has no main thread. But you should not block if not AsyncEvent ?
But you have to still do those blocking actions on a different thread
That applies even for AsyncEvent
I swear I wrote about this a long time ago
Oh, yes, here it is https://www.spigotmc.org/wiki/common-development-pitfalls-bungeecord/
The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy.
In BungeeCord, most events and/or commands are run on the network I/O threads. Due to the nature of BungeeCord (where all threads are multiplexed for high efficiency), performing blocking operations (such as I/O, fetching remote URLs, querying a database, ...) will hurt proxy performance.
I understood clearly. thanks.
null version?
you mean offline-mode?
god knows, nothing on that line really should be null
unless something threw an exception and bungee 10/10 ignores it
can't really say, it tried to send of a ping to the server
best guess would be check your server itself
[TabList] Task #317 for TabList v5.4.1 generated an exception java.lang.IllegalArgumentException: Team name '100000cowner' is already in use
oh
Can I resolve conflict any default minecraft command as bungeecord command?
e.g. /ban -> /foobar:ban instaed of /minecraft:ban (foobar is BungeeCord plugin)
that doesnt make sense
if it's a bungee command the backend wont get it
likewise, you cant run a bungee command from the backend
Hi guys, is there a way to remove "WaterFall" and just leave "none" there?
@unique linden Do you know anything about that?
- Don't ping randomers
- Not nice to remove credits for stuff
- if you /really/ wanna, use a plugin
Hi, I have that error, anyone know how to solve??
[19:47:37 INFO]: disconnected with: ReadTimeoutException : null
I know its a losing battle, but still feels worth trying; does anyone know of any tools to block or limit clients with forge mods connecting to waterfall?
previous to 1.15: just kick prevent them joining/kick them. forge sends that info to the server that it's forge and has mods
in 1.15: no, forge doesn't seem to send that anymore or at least there is no native support for that in bungee
oof I thought I remembered being able to do it in 1.12! Ok, thank you kind llama bot
beep boop
I am confused on how to find 1.8.9 Waterfall. Is waterfall for all versions?
As it is just a proxy, but I am not sure
Waterfall supports 1.8 to 1.15 ^^
And if I want to have multiple versions on the same network does that work aswell?
Yes
Hello, what is the best way to communicate between Paper and waterfall?
depends
depends on what?
what your goal is
for player related data, you can use plugin messaging channels
if you just want to send some data about online players just use plugin messages
the drawback is that those require a player to be online
so plugin messaging channels are the same as in spigot?
paper is basically spigot but not garbage
they are the same as in bungee
most bungeecord aspects apply
okay cool, one more think i asked that in #waterfall-dev where do i find the maven links for waterfall?
just take a look at the repo
.g papermc waterfall github
(DiscordBot) https://github.com/PaperMC/Waterfall -- PaperMC/Waterfall: BungeeCord fork that aims to improve ......: "Waterfall is a fork of the well-known BungeeCord server teleportation suite. Waterfall focuses on three main areas: Stability: Waterfall aims to be..."
sorry but i can't seem the needed information for maven
such as the repository and the dependency
thats the dependency, but it can't find it, am i missing the repository?
Hi, what's this error please?
https://pastebin.com/exReAfET
basically something tried to register a time twice, possibly on two different server iirc
I tried to connect from Lobby to Svanilla, there were just 3 plugins on Proxy (tcpshield, ipwhitelist and luckperms)
it's most likely one on the paper server behind the proxy
hm, I tried to find issue with mobs lags on one server so I removed all lugins from this server... and mobs were lagging still... so I decided to try remove also plugins from bungee, but then I couldn't connect to this one server (there were all plugins already) and on bungee were only needed. Hm, I will try to discover which plugin is also needed
And btw this is what for error? "NPMBC_Z7 "? Why it's in bungee console?
Does waterfall have any way to query the servers behind the proxy? Trying to get player counts for specific servers on the network
Because that's the name of the team that's being duplicated
There is already an issue open on the tracker, it's one of those stupid bungee limitations which fixing is gonna be fun...
there is a ping method somewere to fetch that
Assuming you mean from an API perspective
otherwise, best you've got is using forced-hosts to provide some "if this host, connect here" type stuff
(no idea if that works for pings though)
zzzCat it's issued by bungeetablistplus
I will try to report it to author too....
don't know why I can't connect on server when this plugin is missing... it's just bungee side
is it able to use load balancer forwarding for waterfall?
haproxy, yes
vultr load balancer is not compatible for minecraft?
is there any guide for haproxy?
needs to use the proxy protocol, basically
if vultr supports the proxy protocol, which they really should... it will work
theres a settings for it proxy protocol
so i need to enable the proxy protocol in load balancer?
you don't
i already enabled proxy protocol on load balancer, its can ping by CMD, but cant ping the bungee server
if you mean you want bungee to actually load balance players across servers, use a plugin
why
either you didn't enable ip forwarding, as it says, or something on your server mangled it
if I enable ip forwarding, people will be able to connect to servers behind proxy
no?
Leaving your servers open to the public is what causes that
ip forwarding just passes the players UUID/IP over to the actual backend servers
why a waterfall proxy cannot connect to the paper servers?
I just told you what you need to set for that...
Has waterfall documentation like paper to see what are specific options in waterfall.yml doing?
Is it possible to have Waterfall not logging or showing any IPs? i.e. no-log proxy
You can code a log filter for that or edit the source code directly. AFAIK there is no other viable way
Probs just grab out the log4j config, remove the file appender
how does this work ive been trying to find a way to make a mc 1.12.2 vanilla to have bungeecord setup to stuff like rlcraft with spongeforge and skyfactory and the sorts. is it possible to do that here
vanilla does not support bungeecord
btw vanilla i mean no modpacks but still spigot or paper or stuff like that
you can't run modpacks on paper, you'll want sponge for that
mean paper for hub modpacks for bungeecord server
well, whether or not that works depends on the mod packs
some don't like it when the mods used aren't present when first joining a server so you need a separate modded lobby server for them or directly connect to the modded server
ok i got bungeecord working somehow ;).
I'm trying to seperate my players into two different servers, so that I have EU players on my german hosted server and US players on my us hosted server.
If they join the same ip e.g. play.mcmcmc.com and go into that waterfall to a hub, go into the portal of e.g. US server and teleports there
does that mean that the US server directly handles the connection to the player? So he has the latency to the US server and not to the Waterfall in Germany?
alright, thanks. Do you know of any solutions to this - other than having two separates IPs?
you would need two proxies, one in germany and one in the US and connect them via a load balancer or regional dns (not sure if that#s the correct term)
Ok! I'll look into that.
why does waterfall/bungee chuck the entire version into the debug screen? https://github.com/PaperMC/Waterfall/issues/517
seems extraneous and obviously isnt great for people on small resolutions/large scales
idk but that seems awful
well if there is no command I can see why
I mean I can see why if someone got really drunk they might think that was a good idea
still seems really dumb

Sounds like you're saying Mini did it.
"nothing we can do about it" besdies just removing it? 
yeah if it were me I'd make it just say Waterfall
yeah, not sure who thought leaking the version was a good idea
I mean of course security through obfuscurity isn't a good approach but it helps against the dumbest enemies, lol
no it's just retardedly large
wait what. is the issue about the text being too long or about leaking the version
🙃
just being too long
I would agree that's a bad spot for the version though too
hell the slot is intended for brand info
is the version part of the bungee brand 😄
how is the unicode support, can I put taco emojis there?
if it works in the chat then probably there too
o boi. totally forgot 1.16 is bringing unicode support
can't wait to replace all the worldedit messages with emoji
poggers
more young users for worldedit
more retardation in enginehub discord inc

that escalated quickly
🏓 ➡️ 🔨
We message? Lol BAN!
Hey, I use waterfall in front of a 1.15.2 vanilla and a 1.12.2 modded server. Problem is that a player can't connect to the vanilla server if he last logged in on the modded server. Waterfall redirects directly to the modded server and then the client fails because mods are not installed. How do I make waterfall to connect the player to the next server in line when the first fails to connect?
priorities list should take care of that but ideally you'll want a plugin to route the player to the correct server on login
Weird, my priorities in the config.yml` look like
priorities:
- vanilla
- modded
well vanilla comes before modded so it wouldn't try to connect to vanilla when failing to connect to modded
you might want to enable that it forces a join on the default server instead of connected to the logout server
Still not sure how this would work, no server will work with modded and vanilla clients, so I can't have something like a lobby, right?
.optimize
Yea, upstream shoves the whole version of the proxy in there
Pretty sure that the server software does too, just, proxy string is super long ontop...
I have already debated on removing the versioning, ranges from "i can see it maaaaybe being handy", but, Then I don't really see in what context it could be handy that you couldn't just /bungee
so I'm assuming waterfall is like paper bungeecord?
it's a fork of bungeecord, yes
That's nice
any recommended antibot please? 😦 got 3 days 24 hours bot attack
can bot attack cause lag in a server and crash it?
yes
yes
any recommended antibot please? 😦 got 3 days 24 hours bot attack
@violet shard
I've heard of https://www.spigotmc.org/resources/2ls-antibot-the-ultimate-antibot-plugin.62847/
and also TCPShield: https://tcpshield.com
tcpshield is great, unless it makes your ping increase by 8x
won't fit for all audiences
in my case, 10ms became 80ms
i got a problem
so i am trying to connect a domain to my minecraftserver
but my public ip is showing instead of my local machine ip
when i put local machine ip it says you cant put it but when i put public ip it works?
How do you plan for people to connect to a server if it's using an internal IP that never actually reaches the internet?
nono
what i mean as local machine ip is like my ipv4 from the internet, i made it into a static ip
wat
assuming that you're doing port forwarding
correct?
if yes, then use your external ip
yes
what you'll see from e.g http://crymyip.com
its crying :C
ok ill trydat
if connecting via external ip does not work for you
you might need to toggle NAT Loopback
better let someone else test connection to your server
ok
i have to wait 48hr for the domain propagate to my ip ... :C
it's almost always way quicker than that
@tidal musk you could switch to cloudflare DNS and get near instant propagation ™️
some asshole domain service providers do slow propagation on purpose to sell premium dns
thanks it worked
Aegis bungeecord recommended?
Yes, we highly recommend software breaking the license of the software where the author was literally yeet'ed from fucking mcmarket for his shady business
Use some common sense
There is some stuff I still need to touch, but, quite simply, all of the "improvements" I've seen from these things are stupid and often create their own issues
ohh
mind you, MC-Market is full of bad stuff, except for the few dozen things that are relatively decent
The forum where you find anti-cheats... and cheat clients
relatively decent things are usually the stuffs get posted on both SpigotMC and MC-Market (Premium Section)
Anywhere a homie can learn how to use a VPS? Most videos don't really help
generally janky terminal implementations
any recommended antibot please? 😦 got 3 days 24 hours bot attack
@violet shard Use XProtect, I had this issue. I tested a lot of plugins but this just worked for me
I cannot get from my 1.12 server to my 1.15 server. it is there every time.
You've posted in every channel at least 6 times, We've told you to remove plugins until it fixes.
which plugin to remove?
We don't know, just remove one by one until its fixed. Prob anything that fucks with packets such as ViaVersion, Protocollib etc
ok the plugins on the 1.15 server? because there are none on it;)
Then ur bungee
But I need it, I'm not starting to remove plugins like luckperms because I need it.
We're not saying to remove them forever
this is part of debugging
listen to us or stop asking
it's getting really boring
Then stop posting in all the channels if you're not going to take our advice, Z litterally told you a day ago
._.
Yeah thats us rn too
broken packet
And what can I do? or do I have to ask the plugin?
Is there any documentation for Waterfall?
regarding what?
The configuration files
not really, but seeing as there are only like 10 lines it's probably easiest to just check the patch descriptions which introduced them
most of waterfall.yml is self explanitory
okay
Throwing documentation up is on the todo list, just, I wanna shove it on the papermc docs, and that already needs a huge reshuffling, imho...
hello guys, i don't know whether this is the right place to ask, but when i try to run my waterfall server it comes with an error and freezes, i don't understand this error or what is means. i tried other waterfall versions, on different machines, same error. what do i do? sorry if this is the wrong place to ask...
.paste
Please paste large logs to a pastebin: https://paste.gg
uhm, in what way? ima take a look and see if it's obvious
server name as a number? er
it's upset that something is declared as numerical that shouldn't be
Processen kan ikke få adgang til filen, da den bruges af en anden proces.
Either your file perms are busted, or, your badOS is being bad, don't run servers on the desktop at all, windows is stupid here
Like, really stupid
omg, thanks so much now it works! damn how did you get that from that xD
named a server 15...
2 second look at the code
Line number matches up to where it's tryna get the server name as a string
bruh, well thank you very much
yeah, the proxy is way more strict about value types in the config
if it's supposed to be a string just always wrap it in ''
okay, good to know, thanks for you guys help
Since I use the waterfall and have a lobby, it's easier for players to try to join my server from the lobby than from the Minecraft multiplayer menu
and Connection handler is using a lot of performance
Is it a way to prevent players from trying to join too fast?
Without waterfall
Aikar's Timings Viewer - View Timings v2 reports from Paper and Sponge
With waterfall
Aikar's Timings Viewer - View Timings v2 reports from Paper and Sponge
waterfall already has a connection throttle itself for limiting players from joining
connection_throttle: -1 ?
yea
Which number I need to set?
Ok, so 5000 should be good?
yea
Thanks
btw, why GC runs so more often on this timings when I have assigned more RAM?
uptime is comparable
ZGC: 262 avg(0.44ms every 75.35s) (16GB ram, 5,5h uptime)
ZGC: 602 avg(0.39ms every 40.83s) (24GB ram, 6,8h uptime)
why the fuck are you a) assigning 16-24gb to a proxy and b) using zgc
it's survival server, i'm using zgc because it's better
i have enough ram so why not?
quite simply, ZGC has a lot of weird caveats, it's recommendation is a bit... eerrr....
works great for some, silently kills for others
that's the issue
he's asking about an apparent survival server's memory issues in the proxy channel
Is it possible to switch bungeecord on demand or what to do just to switch now, if one goes down? I just had my VPS go blackout due to datacentre issues but all my backend is still running and I want to serve people.
There will always be a single point of failure no matter how you spin it. You can shift that point using load balancers or multiple bungees or dns load balancing or a combination of all of them
yes
You can't make people invisibly jump between proxy instances, way too much state involved
Best you can do is just loadbalance
you can do that
if bungee 1 goes down, all players connected to bungee 1 will disconnect, but can connect again to bungee 2
If you combine that with DNS load balancing it might work even better. That however only applies if your users don’t use an anonymizing DNS provider like CF
👀
what would a provider like CF have to do with anything?
You just use round robin
the dns mechanism works well with RR, and worse case is only people being "weird" get impacted if their DNS provider is breaking that, in which case, glhf tryna deal with them bar going for multicast BS
(not that multicast helps too much for small downtimes, really)
@ electronicboy round robin works yes. IP based targeted DNS tracking is what CF blocks.
dont get me wrong
its very good that they dont tolerate it
Well, yea, but, nobody really gives a shit about IP based dns in this avenue
I heavily doubt that anybody sane in this environment is doing their own DNS
any body know a ddose protection plugin for bungee or paper
ddos protection can't really be done at the plugin level
unless we're talking some really shitty level ddos like dig packet spam
Has waterfall something like /restart on paper?
That’s nasty. Is there a reason? Inf loops are bad practice after all
Or ist it supposed to be started as a systemd service 🤔
I mean /restart is barely properly supported on paper so yeah, would have the same issues on bungee lol
if you want to do something more complex then you shouldn't rely on in-software commands to manage the status of that software itself anyways imo
I personally just have a loop that exits if a stop file is set to 1, otherwise it just restarts on /stop and /end ¯_(ツ)_/¯
That’s a possible one, somebody could ultimately use a plugin for writing into that file ^^
Problem is I want to delegate the possibility to restart the proxy to mods without giving them access to the machine
But that could be the trick. I'll dig into that, it’s simple to write to a file on a command and then issue the api call to stop the server
yeah, that's tbh safer than using a mechanism like /restart as it would have to start the new server instance out of the existing one before that process is fully shut down
I would like to know how spigot is doing this but there is no public code
Anyways, thank you for the input Phoenix!
all the code is public?
Really? Can you get me a link, because I couldn’t find it (not on github at least)
while not really matching this channel take a look at this patch (paper might further modify that though): https://hub.spigotmc.org/stash/projects/SPIGOT/repos/spigot/browse/CraftBukkit-Patches/0030-Watchdog-Thread.patch
Thank you very much! And yes, it does really not match the channel. I appreciate your help
found one ddos protection, https://tcpshield.com
It is possible to limit waterfall to specific version protocol, instead of supporting the majority of it?
there are plugins for that iirc
hey i need some hel;p
so I read the guides I watched a video on how to do it and I think i did it right but whenever I try to run the bat file the cmd prompt flashes for a second then dissapears
hello?
well then your batch isnt working properly
Can I prevent from direct join to my server without additional plugins?
For example when lobby server is down, players can connect directly to survival.
and bypass login

That sounds like an issue with a cracked/offline network. You won’t get much help here with that. I suggest you go look at whatever resource you use to secure your network
They offer DDoS protection. And I would not advise you to go with them.
why not Tux
basically, it boils down to terrible service and scummy business practices
you cant really prevent it
you can build a honeypot system or be like hypixel and just have tons of different proxies and dont really care if one goes down
the best bot protection is using an online-mode proxy
granted, it raises the bar a bit, there are plenty of compromised Minecraft accounts out there, but the lazy people will be stopped very easily
yeah but still dosnt really protect since you can get hundreds of compromised accounts very easy if you know how
honestly just wish they'd switch to MSA already
compared to what we have now, MSA is a massive step up in security
MSA?
Microsoft accounts
having little security is also WTF
well they could finally offer 2fa I guess but I don't see how the whole privacy issues that Microsoft has would better the siutation
tbh mojang should just force people to use mojang accounts, that would probably solve half of all the "hacked" accounts, lmao
Hey! I have a few questions before I dive in the waterfall.
- Is it possible to sync inventories across selected servers?
- Can we keep the servers in online-mode?
there are several different plugins that can sync inventories
and you will have to set your paper servers into offline mode. waterfall will do the authentification and forward that information to all the servers (if you have it configured to ip-forward)
Thanks!
Hi, could I disable this message somehow? >.... [16:10:48 INFO]: Attempting to send 1 players to Lobby
.... [16:10:48 INFO]: Send Results:
.... [16:10:48 INFO]: EVENT_CANCEL: 0
.... [16:10:48 INFO]: ALREADY_CONNECTING: 0
.... [16:10:48 INFO]: ALREADY_CONNECTED: 0
.... [16:10:48 INFO]: SUCCESS: 1
.... [16:10:48 INFO]: FAIL: 0
Smells like a plugin issue
Hey, I'm very new to waterfall and I'm trying to set up my first server but I get this error while connecting to the lobby server
even though in the config I have
if you guys know what I'm doing wrong and I'm sure it's basic, please ping me!
and yes, I have restarted all the 3 servers
alright, I figured it out and it turns out I have to connect to the bungeecord server instead of the 2 servers I want to connect between ;-;
getting this error, while in the lobby server offline-mode=false does anyone know how to fix that?
reboot it
already have
all 3
maybe I'm starting them in the wrong order
just doesn't work, could it have something to do with the ports? none of them are 25565
How exactly does the config option log_initial_handler_connections in the waterfall.yml file work? I have it set to false but the console and log files are still getting spammed with InitialHandler has pinged
AdvancedBan
Hello, i want to create a bungeecord server to distrube load on different server. Is it possible to create one without a hub? So instead i just want the main to be our survival server and from there you can be connected to the ones hooked upon it. Is this possible? And if so are there any downsides? I haven't seen anyone do this before
it's completely possible and there are no differences or downsides because it's exactly the same thing as far as bungeecord is concerned
there is nothing special about what people usually setup as a "hub server". It's just another backend server and can be whatever you choose
Of course, it may be that certain hub-feature plugins assume your hub server is a single backend server and not a survival/hub server
Thats great to hear, this will give me so much possibilities. Thanks 🙂
and we dont use any hub plugins since survival so i think that will turn out fine
why waterfall can cause cpu spikes?
I literally just listed two...
oh ok
is there some sort of waterfall timings
@tidal musk Dude, why you are using swap memory. This thing slows everything, it has slow access time because HDD is not designed for random access.
just do swapoff -a and you will notice performance boost
sometimes you want to have swap
due to your problem, I bet that you run offline mode and you got attacked by bots - well, best solution is to turn on online mode. Second one - way better hardware & firewall setup that will block proxies
yeah and then there's Kubernetes - won't even launch if you have swap enabled
lol
(unless you pass a command line switch)
sysctl vm.swappiness
if it's 100 then your system will move rarely used programs into swap
rarely used running programs*, to be correct
Hey idk why but my waterfall proxy is just fulling up with ram
and just stops/freezes/become unresponsive when it reaches the limit
Only 15 players connected and Ram usage: 709 / 1024 MB
Plugins:
usually it never excided 512MB
hello anyone?
This might be a bit off topic, but could someone explain the difference between paper and waterfall?
Waterfall bridges servers together (optimized and improved BungeeCord), Paper is an optimized server based on Spigot
Oh I see thanks
Hey guys I keep getting kicked out by the waterfall proxy https://hastebin.com/ezikimopem.md
waterfall.yml, there is an option to disable entity metadata rewriting
Roger
(Assuming you're on a modded server, if not, probs a bad plugin)
there is a PR, no idea if it's been updated yet
Waterfall doesn’t support pre6
I do have viaVersion / ViaRewind on waterfall
contact with via support then
1.16 when
help
it just using tons of ram
for 5-6 plugins and 20-30 players
is this a bug?
in new version?
Java will use the memory you tell it to use
No it's just increasing the ram usage
when I start it's bellow 100 mb
yes because that's how java process uses memory
but this will excide the ram limit
and stop working
exceed you mean?
limit it then
when java considers itself to be running on an end-user machine, it will be reluctant to allocate more memory
if it considers to be running on a server grade machine, its first choice is to allocate more memory
but before I updated it works fine and never went more then 200-300 mb
replicate without plugins, then take a heap dump
sure you can
before we even start on that path
pid changed
have you tried to limit the max memory like skeem suggested
no but I will try that
@primal bridge if you still need it, look in the Waterfall PR section, youll find my waterfall fork that supports 1.16 snapshots there
Via* wont magically do that for you
also permission granted to poke me on ViaVersion if you still need help with that
Found ur repo
thanks I'll try it
@royal hawk do you have like a CI where I can get a jar
or do I have to build it myself
sec
https://jenkins.krusic22.com/job/Waterfall-Xernium/ build #13 is pre6. It has yet to pull in and build pre7 @primal bridge
What is waterfall
it's a lot of water falling
What is waterfall
@tidal musk bungeecord fork
@primal bridge apologies about snoopy. I don’t know where he gets his attitude from. He maintains a modified travertine or was it waterfall fork with 1.16 support, but frankly what he’s doing is copying the work I did- Not to insult him in any way- But some of the things there especially concerning networking may still be outside his league
I saw it. He usually behaves well. Some of the logs are just cringeworthy to read
XD
wait you guys keep logs of deleted messages? well duh. Gotta be able to ban-hammer the idiots that think posting some sort of invite spam scam or messages with malicious intent is acceptable
what
what he’s doing is copying the work I did
lmao snoopy does not know a shit what he's doing, search for the messages what he's put here and you'll see how clueless he is
votifier in 2020
which snoop is that micro, or by what nick do they go by here
nuvoteifier
snoop gives 2 user matches
lol
justsnoopy
it's not giving any results for that one
Justsnoopy30#2834
I don't ever try to insult peoples intellect or lack of knowledge. Though Snoopy is on another level with being persistently annoying sometimes

got a question: what version works with 1.13.2?
Need some help, it seems like Travertine (Waterfall with 1.7 support), also which happened with HexaCord doesn't forward packets right. Servers need as you know to be in offline mode, Trav is set to online and also ip forwarding turned on. This happens alot with mods like Immersive Railroading, MTS (IV) and FVTM. Server running Magma 1.12.2. Logs will be sent under this post
The mod devs argued, that because of the offline mode, it doesn't work because the mods are checking for players, but so it is set to offline they are offlineplayers and they have different variables, so this isn't working. So my thought is that maybe Travertine or Bungee itself is not sending packets/player thingies right.
I have no clue what to do, please ping if answering, that I can see the answer upon that many discords
.paste
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
(Now in the right channel xD 🤦 )
But, the offline mode thing is solvable with sponge, it will allow for proper UUID/IP forwarding
entity IDs, which are the common actual issue, can be fixed by disabing metadata rewriting in waterfall.yml
So those problems can be fixed then by enabling that option?
Next time I will for sure paste it onto pastebin, sorry for that 😄
Is 1GB enough for proxy with max 70 players?
yes
[15:26:48 ERROR]: Task BungeeTask(sched=net.md_5.bungee.scheduler.BungeeScheduler@511816c0, id=14361, owner=skinsrestorer.bungee.SkinsRestorer@5bdaf2ce, task=skinsrestorer.bungee.listeners.LoginListener$$Lambda$406/0x00000008403efc40@454782fd, delay=0, period=0, running=true) encountered an exception java.lang.OutOfMemoryError: Java heap space
I think it isn't normal?
Well, 70 players on a non-pluginified server...
I have two plugins only
of plugins is irrelevant
if they're throwing a ton of memory around, it's gonna run out eventually
because it's a 10/10 plugin known for persistently causing stupid issues
like memory leaks?
excessive memory usage has been attributed to it a few dozen times, yes
ok, good to know
The plugins idea is alright
It’s approach is wrong
So is the logic used on many things
i am sorry, am so bad in enlish
please tell me, what are the advantages of waterfall, unlike papermc?
Waterfall is a proxy. It is a modified version of Bungeecord. It is not comparable to paper
Oh, ok, thanks
can i have help for install waterfall on my vps debian
.ask
If you have a question, please just ask it. Don't look for staff or topic experts. Don't ask to ask or ask if people are awake or available. Just ask the question to the channel straight out, and wait patiently for a reply.
.ask Do you know anything about how to install a plugin on a single player minecraft world, if so, please help me do it.
If you have a question, please just ask it. Don't look for staff or topic experts. Don't ask to ask or ask if people are awake or available. Just ask the question to the channel straight out, and wait patiently for a reply.
(DiscordBot) Error: API returned 401
- Wrong channel
- You don't
you'd need to run a server, can easily run one on your own machine and connect to it locally assuming you have the resources
Sorry if I used the wrong channel, though, thank you for answering me. I'll see what I can do.
help me, I setup waterfall and whenever I connect to it I get this
what does this error exactly mean?
connection refused means that there is nothing running on that port or firewall
waterfall is setup on 25575 by default
I can connect to the other server which this is supposed to forward to, but ofc it says I need to enable ip forwarding
that's an indicator that the server is on
so doesn't that mean that it's actually connected..?
it means that either you didn't set waterfall to listen on the correct thing, or it's not running
another question, I put my proxy ip to 0.0.0.0, but when I put it to the main one, 117...xxxxxxx:port it says that it couldn't bond to that ip address. why so? I'm using an online hosting service
depends on the network setup
how tf do I fix my problem then
it depends on how you're hosting
if you're using pterodactyl, be warned that it deals with stuff in a different manner, generally wanna leave the port allocation, etc, alone
if you're on some standard shared host, check ports, make sure it's running, etc
What is the main difference between waterfall and usual bungee?
Sorry, I am unable to found it by myself
waterfall is the same concept as paper applied to bungee
better perfomance?
yes
ok thank you
Well, to the problem of mine shown above: Is there somewhere a fix for that issue? More and more mods keep having those issues, doesn't matter which versions. Zora no Densha in 1.7.10, in 1.12.2 ImmersiveRailoading, ImmersiveVehicles and RealTrainMod. So they have to fix the issues with packets all along, that is kinda bad. So I wanted to ask, if it is somehow fixable that the forwarding of packets is working on Travertine side. We don't want to stick with Sponge, as the plugins there are in my opinion not that good and also for our whole network. Also it is kinda bad that the mod devs have to stick with that problems.
EDIT: Seems that some Sponge servers will have those problems too.
You can't have UUID/IPs working without sponge
At least, not unless somebody did it with a forge mod, but, don't think Ive seen that
Sponge also can't fix issues with mods being janky
if there is a forwarding issue with waterfall, we'd need more information, but there is heavily a "not all too much we can do" given that some of these mods are just programmed in a way which will not play too well
disabling the entity metadata rewriting will fix some issues, but, potentially cause its own set of issues which mods which don't handle the client resetting state properly
https://github.com/DonBruce64/MinecraftTransportSimulator/pull/463 here you can see how it is coded. You might need to go over to the APacketVehicle thingy, there you can see the changes which have to be made.
Yeah, that option didn't work out and fixed nothing sadly.
without rewriting, those IDs should be forcefully updated on the client
(There are some hacky proxy plugins which seem to break that feature, however)
Well, on that side we sadly have to contact those mod devs to get it updated. Nonetheless, thanks for help. What information would be missing to go deeper in the issue?
I'd need proof that waterfall is mangling something that it shouldn't be when metadata rewriting is disabled, making sure that no plugin is actually messing with it
And how can we get the proof? By recreating?
Because logs and so on are empty of errors or packet information.
that would more be debugging what the server is sending and the client is receiving
Is there a debug mode somewhere?
not that I'm aware of
I think that the client has some stuff if you enable debug logging, but not sure if it actually produces anything useful for this
(DiscordBot) https://wiki.vg/Debugging -- Debugging - wiki.vg: "Feb 18, 2017 ... Debugging. From wiki.vg. Jump to navigation Jump to search. The Notchian client and server (and additionally the ..."
oof, I am not that experienced in that kind of things xD
I could maybe make my server available for testing, because I cannot turn on/nor know how to enable those debugging features. Too complex for me Xd
Wouldn't produce anything useful, really
Maybe one of your collegues knows some way
you would have to dump packets at the server before it hits waterfall, and the client after it hits waterfall
for the client there's a patch you can apply, and likely a forge mod but Im not familiar
on the server you'd need to fork it
also go a lot smoother if you had waterfall running in the debugger
then you could watch every single thing it does and how it screws with it
including all the stuff from the mods
but it would likely be very laborious
I apparently have no clue how to do all this stuff
Even with this site noted from zzzCat
what does it mean when these show up even tho nobody's actually online nor trying to connect?
InitialHandler - read timed out
InitialHandler - NativeIoException: readAddress(..) failed: Connection reset by peer
something opened a connection to the proxy
ah
can I please get help with this https://pastebin.com/QqGhj5U4
is your key correct?
yes
First thing I checked
this too
thats on bungee
wait maybe I did the keys backwards...
nope
Project Waterfall build #347: SUCCESS in 3 min 14 sec: https://papermc.io/ci/job/Waterfall/347/
Project Travertine build #143: SUCCESS in 1 min 29 sec: https://papermc.io/ci/job/Travertine/143/
👍
It's failing to connect to the voting plugin on the backend servers or something
Im redoing the entire thing
its rlly weird
idk why this is happening...
I redid it and it broke again
oh\
It worked
It was because I was using the test button
wtf? what does the test button even do then? lol. It works on paper perfect and is just weird on waterfall...
whatever thx!
Does cpu matter in the case of bungeecord? Are there any minium / recommended specs to use?


➡️
❓ ☑️ 🚫