#waterfall-help
1 messages · Page 49 of 1
could you elaborate on "server query enabled without a password"
minecraft supports a query protocol which allows executing commands
.g minecraft wiki query
(phoenix616) https://minecraft.gamepedia.com/Minecraft_Wiki
oof
actually that's the wrong one, I don't think that one supports commands
this is what I was thinking of: https://wiki.vg/RCON
well make sure that it is, then check all your plugins for backdoor code I guess
and if you discover one redownload all plugins and server jars from trusted sources and check them for the previously found code to make sure they don't include
(I would start by searching for where that port was opened)
alright, thank you
after some time and asking around
im told that it's an uncommon luckperms exploit that happens when someone is able to run waterfall commands from backend servers
and i quote:
"so you need to disable the ability to run waterfall commands from in-game
what they are doing is accessing luckperms via the waterfall server
there's a waterfall config item that prevents people from executing waterfall commands from the minecraft servers"
waterfall doesn't have such an option, they are most likely talking about a plugin adding that functionality
https://pastebin.com/ghNnLnGP help pls
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.
how did you manage that 
https://pastebin.com/bWx6We6d This appear in the last update of waterfall
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.
connection_throttle_limit for what is this?
guys waterfall got netty crasher optimize?
!config
do backend servers have to be in offline mode while setting up waterfall in online mode?
or should they be in online mode as well
They'll be in offline mode, bungee handles the auth
alright, thanks
NoClassDefFoundError : com/google/common/collect/Iterables$4 @ com.google.common.collect.Iterables:569
error with the last update of waterfall
whodahek r u
.flags
Optimized & Recommended JVM Startup Flags: https://mcflags.emc.gs
hello. i have a werid error. i am currently running hexacord to test it out. i was running travertine and a forge with spongeforge modded server (FTB Revelation). Randomly the players got disconnected. MSG: Server went off bla blah connecting to fallback server. have the same problem on hexacord and travertine
did you disable entity meta rewriting in the waterfall config?
ITS Off yes. But i think i havent touched that config
What version are you running? Waterfall and on your servers What plugins have you got? Etc... We need more info than just a log dump.
Also read this https://github.com/SpigotMC/BungeeCord/issues/2806
Your issue is more likely a DDoS attack rather than a waterfall issue
Somebody is sending bad traffic to your server
Hi, why appear this? https://pastebin.com/pSQBisHE
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.
Cancelled ServerConnectEvent with no server or disconnect
a plugin cancelled the event, but didn't disconnect the player or send them elsewhere
what benefits has waterfall vs bungeecord?
we have various performance tweaks over bungeecord, a bit more API and have improvements to make the proxy more robust
okay, thanks. is there any wiki to check the full comparision?
No, the patches are on github, however
cool thank you cat!
are waterfall servers just accessible with a Forge launcher?
waterfall is a proxy, not a server, it supports forge 1.12.2 and below, modern forge all hopes of compat as-is essentialyl
i want to use waterfall as a proxy for paper (not forge) servers
or i must use normal bungeecord to do that?
That's not from waterfall
oh my bad
Is there a way to make forced_hosts work with restricted?
Like if you try to join a certain server through the forced_host setting and you don't have the permission for that you get sent back?
sent back = that the normal priorities list applies
Quick look at the code tells me that you shouldn’t be able to do this without plugins
But with plugins it should be quite possible
Du you happen to know one?
From the admincraft discord:
I am on last version of waterfall, all my 1.8 to 1.12 servers are not working
For example, I have a 1.8.8 papermc server connected to the bungee, when a player tries to join with a 1.8.8 client they get disconnected with that console error
GenesisIIX lost connection: Internal Exception: io.netty.handler.codec.DecoderException: The received string length is longer than maximum allowed (30 > 20)
Im getting this error even with a clean server install without any mods or plugins
And when connecting directly without going thru bungee, it works fine
I did my own research and found that the only part of the minecraft protocol that requires a string with a length of 20 chars or less is the Server ID in the Encryption Request packet https://wiki.vg/Protocol#Encryption_Request
Would that have anything to do with this bug?
The protocol doc describes the Server ID field as being "unused" but does third-party server software like Bungee/Waterfall use it?
@undone carbon No, that would require a length limit to be in place here on Waterfall, which it is not.
Also side note
Encryption is only bungee <-> client
Is this error from the bungeecord or the backend server?
backend server
Then I would assume that this is the arbitrary plugin channel identifier/name length limit md_5 imposed back in the day youre stumbling upon
let me check that
Hey there, so Its my error if you have questions qou can ask me directly it would be easier, @undone carbon was trying to help us hahah 😄
btw, i was doing support for @rugged acorn; i'll let him take it from here
Thanks again man :))
ye
well checked and the message then is more to the tune of Attempted to send a Plugin Message to a channel that was too large. The maximum length a channel may be is 20 chars.
wait
need to check the other side again
You mean a bungee plugin right?
No problemo 😄 ty XD
enable debug in server.properties iirc if you want useful messages
that much is true too
Ok debug is enabled, server is restarting, well see
So yes, this has to be a CustomPayload or plugin channel name
since the server is in offline mode it cant be the encryption
wrong direction for that packet anyways
encoder vs decoder and the location of the message are importanté
Im getting this with debug https://pastebin.com/MzVhRwYm
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.
So, yea, some plugin message
Youre saying the problem comes from a bungee plugin?
either that or from a client mod
Cant be client mod im testing it myself
well do you have a custom payload or plugin channel registered?
they're using via
It's more the protocol hack aspect of it
Well nice
You have a plugin on the proxy which is
- Seeing a newer version of the protocol, so assumes the newer limits
- using an older version of the server software which has a limit on the length of the channel names in there
I mean, I couldnt ask for better support 😄
I would assume its #1
it's 1 and 2, it's assuming newer limits and pooping itself coz the server itself doesn't allow that
On bungee we dont have much plugins, only the "official" recognized one
installing via on the proxy might be more compat but is not the ideal setup
Yeah thats what i have been told
otherwise, you'll have to find out what plugin it is and "prettyplz" the author
Well its not beyond people to change the limit on that old version of spigot either
Ok so I will try on a staging Bungee to see if I can target which plugin or even reproduce the error and Ill keep you guys informed, thanks for the support,
(I will also try to install via on bungee, whch one ahould I installl?)
Is it possible to alter this message as users keep getting confused when i they see this and I just want it to say something like "This server is premium only"
https://cdn.discordapp.com/attachments/479414867556696083/802637484936003624/2021-01-23_16.33.34.png
This is using travertine btw
I get a feeling that's from the client
Luckily this is for an app, so its used to generate codes (think of how namemc does website linking)
maybe I can switch it to offline mode then kick out non authenticated players with a custom message?
what I mean is, that that error was 100% on the client side, it didn't even manage to actually connect to the server
did you mess up the auth process 🤷♂️
Nah, its all online mode at the moment. Theres a plugin that runs on the LoginEvent (i think) sends a post request yada and kicks the client with a message
But if that message is client side... My only option is to set the proxy to offline mode and manage auth myself and kick with a custom message
Could not connect to a default or fallback server, please try again later: io.netty.channel.ConnectTimeoutException
This keeps coming on my server when anyone tries to connect.
My config is here:
server_connect_timeout: 5000
listeners:
- bind_local_address: true
force_default_server: true
forced_hosts:
pvp.md-5.net: pvp
host: 0.0.0.0:25565
max_players: 1
motd: '&1Another Bungee server'
ping_passthrough: false
priorities:
- hub
proxy_protocol: false
query_enabled: false
query_port: 25565
tab_list: GLOBAL_PING
tab_size: 60
remote_ping_cache: -1
network_compression_threshold: 256
permissions:
admin:
- bungeecord.command.alert
- bungeecord.command.end
- bungeecord.command.ip
- bungeecord.command.reload
default:
- bungeecord.command.server
- bungeecord.command.list
log_pings: true
connection_throttle_limit: 3
prevent_proxy_connections: false
timeout: 30000
player_limit: -1
ip_forward: true
groups:
md_5:
- admin
remote_ping_timeout: 5000
connection_throttle: 4000
log_commands: false
stats: 06156314-4c5f-4c2a-88cf-689455266bee
online_mode: true
forge_support: true
disabled_commands:
- disabledcommandhere
servers:
Server1:
address: localhost:25567
motd: Server 1
restricted: false
hub:
address: localhost:25566
motd: '&1Just another Waterfall - Forced Host'
restricted: false
I have tried putting the numeric ip in the servers section. same error. This worked on my other host, but after switching to a dedicated server, it doesn't. When I connect to the Hub using the ip for it(with bungee off) it connects me. What can/should I do?
anyone able to help?
I'm using pterodactyl panel.
The open-source server management solution.
but the thing is, I used this configuration with my other host that also uses pterodactyl
so it should work...
if you mean localhost, I have tried with a numeric ip as well.
Bind everything to the ptero interface, localhost is isolated to the containers
why cant i join a server? so im in the lobby and im op. i added a queue plugin to connect to my main server. so it works fine but i cant bypass the queue. would this be on the plugin? or the server not recognising i can bypass player cap?
I would imagine that would be down to the plugin's permissions
Oh wait you just added the proxy didn't you? Proxy plugin permissions are going to be handled differently
huh?
i have literally no idea what im doing
i got 2 servers, one lobby for the queue and the main one
it directs players to the lobby and auto places them into a queue (using ajqueue)
The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy.
That should explain everything you need to get going
done all that
imma just say fuck it and make some stupid way of doing this
ok so my genius plan failed
why is it if i sudo a player to /server it doesnt work
it comes up unknown command on their side
Did you configure the bungee permissions?
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.
i keep getting this when trying to connect?
is there any fix or explanation to this?
do you have any plugins installed?
yes but this happens on a fresh install as well
events don't really do anything without plugins though? 
is this a bot or an actual person lmao
this is really not the place for prejudice against bots...
okay i removed all plugins again
but now im getting [Proxy] Cant connect to server and no errors in console
this is another reouccring issue
wait i got this before it disconnected me
16
28 WARN]: [/75.181.95.160:54730|SweatShawp] <-> DownstreamBridge <-> [hub] - NativeIoException: readAddress(..) failed: Connection reset by peer
some players can connect and some cant
Connection reset means that something external to the proxy closed the connection in an unexpected manner
do you know any common causes?
ive been reading up and it could be the host amongst other things
but it does only happen on 1.11< for certain players
bad connection somewhere in between the player and the host
is it possible to change error messages? like server full
why is bungee forcing me into another server randomly? no plugins
check the logs
should shove you into your last server unless you have force default on, otherwise it'll use the priority list
https://gyazo.com/340256e9fa24cecd754a555776f5c4ad?token=3c191288f28c36bf2c57be8333827e5b
any command i use everyone gets kicked
tell the plugin author to use adventure or switch to velocity. it's a known bug in bungee-chat without a fix or cause
(or remove any hover event from your messages)
ok thank you
hi, my waterfall has a spike of CPU every time one player join, like 1% to 60/70% with and without plugins, any help? 🙂
it's running on i9 9900k
do you have any plugins installed?
might be normal then, I mean it DOES need to do some work on join (auth as well as relaying a bunch of packets)
hmm got it, thanks!
Other thing, tonight I did an event with more than 150 players in a very small prelobby, waterfall was using almost 900% of the cpu, after the event started, the used waterfall cpu went down more and more, what can it be due to? 🙂
use a sampler
not showing all too much useful there, try using a smaller interval, maybe potentially might pick something up, but generally 🤷♂️
All we can tell you is that mojangs API said sod off
is this happening to anyone else?
does waterfall support the same plugins as paper?
No, it's a proxy, not a server, they have two entirely different goals and do two entirely different things
ok, are there any good guides out there? I have not found much
I'll try that then
Is it possible to link a server to use a nether of another server?
same for the end
you'd need to find a plugin for that type of stuff
ty I will try to do that
Can I reload the config without restarting Waterfall?
Yes, please let me act upon the wealth of information you've provided
if you're talking about the well known issue with bungee chat, if we knew, we'd of fixed it.
You can but seriously not recommended
In fact the console will tell you to restart ASAP
I mean restarting takes like seconds for proxies
hi
i try to join the server
i get this message: Exception Connecting:NativeIoException : readAddress(..) failed: Connection reset by peer
connection reset by peer means that the connection was closed in an unexpected manner
how to fix it then ?
i tried removing some bungee plugins
and i restarted server many times
but i still cant join
that's not a waterfall issue
That's generally a network configuration issue
If you're using a hosting provider, try speaking to them
otherwise, check firewall, etc
ok ty
Am I correct in thinking this will send a message to every single player on the proxy if I send this from a paper plugin
ByteArrayDataOutput out = ByteStreams.newDataOutput();
out.writeUTF("Message");
out.writeUTF("ALL");
out.writeUTF(message);
Player player = Iterables.getFirst(Bukkit.getOnlinePlayers(), null);
player.sendPluginMessage(instance, "BungeeCord", out.toByteArray());
at the moment seems to be only sending it to a single player
Iirc you'll need to tie it into something like BungeeChat
Why don't you do /alert or /alertraw in bungee console

@little garden You're correct, Message ALL Body is valid.
I think it's cool that you do getFirst online player, I'd probably use the message sender though unless it's console.
Also this probably should go in #waterfall-dev just saying...
I realized but eh
Ohh that's my bad, I clicked the wrong channel
net.md_5.bungee.util.QuietException: Unexpected packet received during server login process!
19a7017b227472616e736c617465223a
And I'm not using protocol lib, and theres no plugins interfering with packets.
All that is there is geyser bungee and floodgate bungee
I'm gonna try reuploading the jar files again and see
Nope
"all there is"
lists two plugins that massively impact internals
remove them and test again
I mean yeah i can connect without the plugins...
Oh well
Gotta ask the plugin devs abt it
Hello I was curious about this issue:
If anyone had it or it's the authentication of mc is down, besides i am running waterfall offline mode.*
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.
Hey sorry, but is there any way to get the waterfall performance improvement as well as the flamecord protection?
Your best bet on protection is going to be a properly configured set-up (such as firewalling). That being said, if you aren't relying on any Bungee plugins, you may want to take a look at Velocity, as it has its own set of improvements/performance boosts (and is well-regarded in the community). The biggest downside is not having access to Bungeecord plugins (although Velocity has its own set of plugins)
If people can provide info on stuff, we can look into it, however, most of what flamecord adds is a leaky firewall
so basically flamecord is a fake or it actually does something?
Basically, their software is a closed source rip of waterfall, they've added a few tweaks to it, one of which waterfall already has but in a much more refined way, beyond that, the only real of-standing thing that I've seen is their firewall, which is basically a huge ass memory leak
I’ve already got two more mitigations lined up that I am going to PR later today if all works out
I mean. not too much, really
disable query, firewall off the IP, work out where that connection is coming from and speak to them, could be some service you signed up to borked or was misconfigured or something
whenever i try to disable query ( while server is turned off ) it turns back on, so i guess the company is automatically enabling it for it's own reasons?
okay, ty, and sorry
Waterfall supports 1.8, Travertine 1.7
Oh okie, thank you!
Hey guys - we have a new server running on a 3900x (3.8 ghz) with 32gb memory (64 for server itself) on US East Coast- we notice tps drops as low as 17. Running 20-30 people on it. Would this be a good candidate for Waterfall? Can waterfall be placed in other geo regions like Europe for example or are they best placed in the same dc?
Waterfall is a proxy to link multiple servers together. Ideally a proxy is close to the servers, because the distance and relative lag from the client to the proxy and then to the server has to be taken into account. Possibly making two round trips over the world; to the proxy and from there back somewhere else is less favourable
linking how exactly? I doubt it can support the same survival world?
oh I did delete there yes - figured it would be a better question in here
No, you connect individual servers to a proxy. They don't communicate with each other
if you want to build a network with for example a factions and a minigames server then a proxy is right for you
thats what I figured ok - was hoping to use waterfall to maybe take some of the connection strain off the main SMP box but not the right fit
yea doesn't change overall bandwidth
Well youre relaying the same load by using a proxy
just where the front doors are
tps drops to 17 is like nothing, but otherwise check your /timings paste
and we should add, modern minecraft versions can be very heavy on the host system even if you have the most balling specs
makes sense thanks guys
I am using lastest waterfall:
Error when send command with luckperm via test donation:
https://gyazo.com/705f025f9753d0a4bac35046b58e0ec2
full: https://pastebin.com/hUUrJ1fb
The targeted player is a bedrock user on java*
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.
some chat plugin threw an exception
you'd need to speak to the plugin author ro something
Oki, thank you.
is waterfall a drop-in replacement like paper?
Yup
Does anyone know an alternative for https://www.spigotmc.org/resources/mysql-player-data-bridge.8117/?
Need one where advancements also get saved. Paid or free. Pls ping if anyone has some info.
what is the difference between travertine and waterfall
1.7 support
is that it?
Yes
k
What does this mean? XD encountered exception: net.md_5.bungee.util.QuietException: Unexpected packet received during server login process!
that something send an unexpected packet during login e.g. a client mod or a plugin
Hi. Help https://pastebin.com/ypppFpgU
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.
@finite mirage a plugin prevented the player from going to any server
Contact the author of the plugin causing this. It’s not a waterfall bug nor something we can help with. Tell them that if they cancel a server connection event the player will be kicked if there is no fallback server in line
Thank
Project Waterfall build #394: SUCCESS in 3 min 44 sec: https://papermc.io/ci/job/Waterfall/394/
Project Travertine build #171: SUCCESS in 1 min 58 sec: https://papermc.io/ci/job/Travertine/171/
Hello, I have a Waterfall server and I added two servers. When I log into Bungeecord, it does not redirect me to the main server, it redirects to the other server, and this causes a security vulnerability in my server. I've only experienced this on my own OP account. Other users are connecting to the main server.
That sounds suspicously like an offline-mode proxy; we don't support that. Secondly, are you sure you enabled force-default server?
the proxy stores the server the player was last on
.paste
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
are you using mods?
how can i fix?
try without plugins, if it is them, you'd need to work out which one
You can fix it by disabling entity meta rewriting too

That setting is in the waterfall.yml
depends on if it's a mangled packet or just a bug in entity metadata rewriting, I mean, I doubt, but, wouldn't be unheard of
but what do meta entities do?
players are entities too
I believe it is viaversion, every time a player enters from a metadata error
precisely viabackwords
what should i delete to solve this problem?
the server is not open yet
We don't know
-.-
remove via backwards. it's pointless anyways, people can just update 
@grim glade output of /viaver dump would be helpful.
I'm getting that when I connect to skyblock from FakeLobby
And I can't connect to SkyBlock
.paste
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
Does anyone know how to hide these bungee commands from tab completion?
if you don't have perms to see the commands, you won't see them
beyond that, I have a plugin on the forums somewhere
let me try
pls send me link
oh sorry
It's on the forums, go look for it.
which forums
@proven flax make it -1 to disable in waterfall.yml
I just want to hide certain commands, is there a way to do this?
if you do that
on the paper forums in the waterfall releases section
https://papermc.io/forums/c/waterfall/10 I searched but could not find
I couldn't see a plugin that can edit tab complete
see the plugin releases section
my english is not good
okey i find
no not this
i find your plugin :D
yes, it's really a suitable plugin for the feature I'm looking for.
I wish he could also block commands would be nice. Thank you for the plugin.
I run a server with papermc & waterfall using bungeeguard and someone was seemingly able to op themselves
Waterfall Log: https://paste.gg/p/Walrus2020/92173bf0e1334a1bbf5583f3de26f395
Weren't able to op but was able to give themselves this
1, Bungeeguard is missconfigured
2, Properly firewall the backend servers.
3, Use velocity modern fowarding.
4, Secure your mysql port if you use it
5, Something's backdoored
Could be yeah
5, a little more difficult to figure out however
@zinc raven what plugin(s) are you using?
Well, there's nothing that stands out, doesn't really rule out a backdoor though. Probably not a back door as long as they're all from well known authors & are all downloaded from official sources. If you wanna be sure though, DM me a download link for your plugin folder and I'll look through them 🙂
That would actually be awesome thanks mate I'll do that
Hello I have a question is it possible to create timings for bungeecord and how long would it be possible to create or program something like this =?
No, use a sampler, bungee is threaded, timings is not designed for that
But if someone tried to create something like that, how long could it take?
This information is really important and more when you have many attacks or a large amount of public
Would it also be possible to do something like this with a plugin or is it not possible?
Basically, timings is designed for something which ticks, bungee doesn't
there is no concept of a loop to monitor, so, outside of tryna hardcode a time frame, it's generally gonna be meh on the outset
not to mention, you'd have to make timings able to monitor multiple threads
I get it and see the consumption of the plugins that use the ups and downs of CPU and Ram when you receive x number of users or an attack also not possible?
It seems that using bungeecord has its limitations
you'd really need a sampler
Hi
How can I protect my server from Bot attacks
I'm using Waterfall
And (Windows
)
Talk to me by Private
Found the exploit
hey, i have a question: is waterfall (and travertine) closed-source? i want to have the source code so i can maintain the fork on my own
.g waterfall github papermc
(DiscordBot) https://github.com/PaperMC/Waterfall
2ls antibot is fine I quess, but the best protection is turning online mode to true.
i want the source code i'm lazy to apply patches 😭
We don't publish the code as-is, you need to apply the patches
ok
Is there a 1.16 alternative to this?
Unless this feature is built-in to paper and I'm not aware 😄
wrong channel, and plugins like that generaly don't break assuming they're built well
does waterfall come with any built in viaversion like system?
No; but Waterfall supports 1.8-1.16.5 natively like bungeecord, it doesn’t translate however
Hi Guys!
I have two problems with waterfall
The first is that when a player joins the CPU used by the waterfall it jumps from 5% to 90/100% (I state that the cpu is an i9 9900k) even without plugins
The second problem is that when there are many nearby users (such as an arena or a waiting lobby) the cpu used reaches even 700% until the players move away from each other.
I do events for my subscribers on twitch, and this happens when there are already more than 50 players nearby, as soon as they get to about 100/150 waterfall it saturates all the cpu.
Could anyone help me? Even in private chat 🙂
Run spark when you’re under that kind of load
Report back when you have that ready
Additionally, is there any kind of indication in the logs that you are being attacked? Might also happen in the background
Going even further; if you don’t need waterfall/bungeecord for its plugins consider checking out velocity
I’m open if you want to shoot me a DM
?
I just don't have DMs open for anyone 
this can be fixed by turning off entity meta rewriting and tab list rewriting
waterfall.yml
The java $env and the java $home shouldn't matter as long as the process is privileged for the appropriate IO/sockets. Also that error message doesnt say anything, the crucial part of it is missing
... okay so I looked at it and Bungeecord trims the important part
nice
I assume the proxy has no way to connect to the server its trying to put you on
if the server is on the same host as the proxy ensure loopback traffic is unrestricted if youre using 127.0.0.1:port or if youre using docker or another container solution; ensure that those ports are appropriately forwarded and you adhere to their networking
now I'm getting confused
This error: io.netty.channel.AbstractChannel$AnnotatedConnectException
-> Cause: Your proxy (Waterfall) can't connect to the server you're trying to go onto
Fix: Resolve your networking or firewall issue
This error: https://srcb.in/rhGuJas1kN
-> Cause: Bungeecords rewriter
Fix: waterfall.yml
disable_tab_list_rewrite: true
disable_entity_metadata_rewrite: true
If your issue is fixed now then thats good I guess
My ping and other players ping spikes up to 2-5k when teleporting, any ideas why this is happening? TPS is at 20
bad internet?
Nope, the server and players internet are fine
Pining the server is fine, it only happens inside the server
Bad performance of CPU

Insufficient of RAM
I'm checking to see if it might be caused by my anitvpn
Unlikely tho
And btw why don’t you allow people to use VPN
It prevents bots
Actually it will
and alt accounts

Fair enough
But it couldn’t block all the vpn IP address
Cuz it’s not cleaver enough
Yeah I know it wont do all
Maybe it will just detect the ASN and some popular VPN’s IP address
Could this issue be a tab complete thing? I just tried to do /gm3 /gm1 /gm3 /gm1 really fast and my ping shot up
How many Ram do you have in server
On my proxy there is 1gb
i9-9900k
1GB isn’t decent but I believe it won’t cause the issue
How many players r in the server?
How many plugins did you install?
How many cores did you enable for the server?
There is like 60 plugins on my server, I can still trigger this lag with only a few players
I have a dedi with 4 cores I think
Although I'm looking at the status page and the proxy server's ram looks like its capping
try to move your server to one of your another computer with higher storage of rams
Test it again
To see if the issue still exist
Just chiming in here; you do have compression disabled from server-proxy right?
network_compression_threshold: 256
Also 1G is more than enough for well over 100 people
is that what you are talking about? Should that be -1?
@harsh vigil the default is 512m and it’s plenty. Per player you’ll be using less than ~3 megabytes per player. Where it gets hard is plugins
Also for networks compression should either be disabled or higher
setting it to -1 did not work
512 or 1024 at max
I can cause this ping spike by simply going back and forth betwen gamemode survival and gamemode creative
Well for one even if the server takes a hit it will remain at 20tps. Ping spike as in server-proxy or proxy-player?
The second one shouldn’t ever be an issue because the proxy does keepalive ping separately from the servers
It's like proxy-player, it will randomly select some peope to lag out when I create the lag by going back and forth
Now here is where it gets interesting
The Tablist shows what the server reflects
And the server either knows the real player ping through forwarding
Or it doesn’t
So I can spam /ping and it be just fine, I did a MTR and it looked fine as well
So when you say ping do you mean actual delays or just that it shows a low connection in the Tablist
On paper servers the command handlers are detached form world ticks so any fired commands will work regardless of other network states; same is true for bungeecord
i can do the /ping command and there wont be any ping spikes. I do any kind of teleporting, changing gamemode and ping will skyrocket
let me do a screen grab
my ping is 100 to the server, I go back and forth between creative and spectator mode and then I get this ping spike. This will also randomly impact the pings of others. No one in this example though. I can consistently make this happen.
Can anyone help me setup a bungeecord server?
-_-
do you know whats the command to test the time latency?
no
mesing with compression ain't gonna fix the latency values
in general, mc's measurement of latency is not just measuring network latency but also how well the client itself is performing
On proxy startup it takes this long for luckperms to connect to mysql [LuckPerms]: Successfully enabled. (took 6007ms)
it's expected to sky rocket when doing things which essentially cause the client to stall in some manner, joining servers, switching works, etc
what about changing from creative to spectator
that shouldn't induce anything afaik
I can 100% consistently cause ping spikes for me and random members by going back and forth between creative and spectator with 20tps
tps is 100% unrelated to the keepalive stuff
I do want to loop this post back in because its pretty much the same issue https://www.reddit.com/r/admincraft/comments/l44urj/atrociously_high_ping_issues/
reddit
there isn't even anything that happens when switching between those two modes
I wonder if it is something to do with PremiumVanish or TAB?
any idea
looks at the channel name
sorry
hi for some reason bungeecord is passing through the lobby's motd as its own motd, and even though i changed it on the lobby, it's still not updating for the client
i also switched to waterfall which doesn't remedy the issue
disable ping_passthrough to use the proxy's motd
Hi, my server cannot download the modules from the papermc Jenkings. It gives me a 403 error. Any way to resolve it?
Whenever I try to go there myself, i get the following error:
Access Denied
larssieboy18 is missing the Overall/Read permission
Update
Thanks, that worked 🙂
phoenix, modules are not deployed separately at the download page.
yes, I know
Then what should they do at the download page lol
15:43:24 phoenix616| whelp, reading the actual log helps 👀
🤦
did that
it was already disabled
help me
as anyone had this message spammed in bungeecord console?
Query - Incorrect magic!: /36.182.78.175:5683
it seems to be some kind of denial of service using query they even spoof the ips
I got 10600 of those requests per second
disable query
I did its fine now I am just asking if it is common
there's been a handful of people complaining about it
What do you recommend for stopping bot attacks?
for query spam? disable query.
worse case, configure your firewall to limit connections
The query issue is solved now I am asking for the classic bot attack
How would I rate limit connections?
you can limit the rate of new connections, etc, with iptables
Im going to look into that thanks
Project Waterfall build #395: SUCCESS in 3 min 40 sec: https://papermc.io/ci/job/Waterfall/395/
Project Travertine build #172: SUCCESS in 1 min 37 sec: https://papermc.io/ci/job/Travertine/172/
What should I set as the network compression threshold (both in waterfall and in spigot)
because when the players load the world the cpu goes up a lot
pretty sure it's usually suggest to set it pretty high (or maybe disable via -1, not sure if that works) on paper and leave it default on waterfall
Hey, I'm getting this error when I try to join my new survival server: java.lang.IllegalArgumentException: Team tm-sb-0 already exists in this scoreboard
How do I fix it?
titlemanager
ugh LOL
so do you have any idea what's causing the issues?
like specifically?
corrupted data, etc?
it uses the same team names between different servers, there is a ticket open on the tracker but long standing bungee issue
hm
no tm for me then
I'm having another issue it looks like
After joining the proxy I get this
.paste
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
that screams network issues
you can disable async dns to use your machines DNS stuff vs relying on netty in waterfall.yml
Alright
can help in some cases, but, if it's network related
Thanks
Is there good documentation for Waterfall somewhere? If there is it eludes me.
most of the documentation is generally covered by bungee
Right, after doing config and having everything setup I'll connect to the server, nothing loads in, the proxy boots me, and then I wait 30 seconds before getting ejected.
From what I've read this is a VPS issue with some providers, and I'm hesitant toward going with a minecraft host as they're typically horrifically overpriced. Does anyone know of a VPS based in the USA, preferably east coast, that can actually work?
Explain your setup in greater detail; I can’t make heads or tails of this
Right.
So, I use serverspace.io for the VM, it has 6 cores, 16GB of RAM, and a fast enough connection, the issue I see with Waterfall (it also happens on Velocity) is that somehow their network configuration is a little off and it times my players out when they join, it's dependent on ping, since the US players can get in and everyone else can't.
I've tested all configurations on my personal server, no issue there. It's only with a cloud hosted VM.
And since the project aims for 7 pixelmon maps, I'm reluctant at best to actually buy 7 servers, each of which will be $30 USD a month, at that stage it's smarter to buy a US based server.
As far as the software goes? Spongeforge 2854, Waterfall 395, most recent MySQL, no mods or plugins active for testing in any configurations.
Hey, my Waterfall server was working but I restarted it to change priorities but now all of a sudden, it shows "Pinging" and then says "Can't connect to server", anyone know why that's happening?
Hi there, I was wondering if there was a way of using waterfall to split the map (nether, end and overworld) over three servers and use the vanilla portals to move from server to server? (So the player wouldn’t know what’s happening)
So set the server restricted in config, they don't have the perm but can still access the server, what am I missing? 😮
address: *************
motd: ""
restricted: true```
Does anyone have a solution to get notified whenever there is a new build of Waterfall?
I used to have RSS feed on the ci/job but since that is closed for outsiders, there seem to be no alternative to get notified.
Are you sure that they don't have bungeecord.server.creative or are inheriting it from anywhere?
Is there a way to place/remove a block when a server goes offline/online?
nothing to do with waterfall
you'd need to have some way of dealing with that on the servers
.895
tnx man
not supported on bungeecord/waterfall
maybe, go look on spigotmc
i did find one but it's super outdated
¯_(ツ)_/¯
Hey anyone know how i add waterfall last version to Multicraft in a waterfall .jar.conf file
What reverse proxy (Other than nginx) should i use for a waterfall network?
haproxy
Well, if you're expecting the proxy to be able to do anything than general load balancing, there is basically no real options as the proxy needs to be aware of the mc protocol
Hi, I need help with WaterFall, when I start my BungeeCord server everything goes well, but when I finish starting, I try to connect and it comes out as "Disconnected", and when I refresh the list of servers it appears as off even though it is on, I correctly set the ip but nothing, and I don't get any error in the console
Alright, also when im trying to use bungeecord or waterfall i can't connect to my server and it just says as disconnect message "Disconnected"
Same thing here
fix your network setup
if there is no message in the proxy log then you didn't even reach it
with network setup do you also mean iptables?
anything that might block your connection
Hi is there any way to hide players on vanish with query-result?
There would be, if the vanish plugin had an api
if so, just intercept serverping, remove vanished players and ur done
You'd need some form of plugin for it which would somehow need to sync the state from the server itself
is there already something like that?
🤷♂️
:<
2 second ctrl-f of the page shows they have bungee support already
you'd need to speak to the plugin author, we can't provide support around premium plugins
are we talking about PremiumVanish all the time?
I think I talked to him once and he said something like
"Unfortunately there is no support for server queries. It's the server or bungee itself that sends it and there is no API for me to interact with."
okay I'll be in touch with the author thanks a lot!
I have question. Waterfall don't have exploits who can went down server?
every software has such exploits, the question is are they widely known
Waterfall exploits being repaired?
Depends on the kind of attack and the target of the attack
give example
denial of service (dos or ddos) to overload the proxy; denial of service over the proxy against the server(s) directly, disassociation attacks ....
there are quite a few things that can never be ruled out
as for attacks against waterfall, we usually do our best to mitigate such attack vectors
aa ok abd packets exploits from clients are fixed?
tbh if you care extremely about a stable experience just use velocity 👀
That strongly depends on what kind of exploits you are referring to. Whatever affects the proxy or integrity of the servers behind the proxy we tend to cater to. We do not fix bugs in ingame mechanics with Waterfall
the biggest servers probably run their own software/forks
But velocity can’t run bungee plugins?
well yes, it cant. But big networks usually code their own plugins anyway. Besides there are alternative plugins for a wide range of bungee plugins for velocity
It kinda can run them 👀
runs
K
What is plugins for protect waterfall?
Protect? Elaborate what you mean by that please
Hmm w8... i cant speak good english xd
When i have waterfall, So i cant need flamecord and exploitfixer?
Or other forks who probably fix exploits
I’m not sure what exploitfixer does but what I am certain about is that flamecord should be avoided. The fixes it claims to have should already be in waterfall
Aaa okay
So flamecord is useless
And how to fix bot attacks? What do u preffer?
Bot attacks is something networks normally don’t have to worry about. A good ban plugin should do the trick there. Bots are nearly exclusively a problem with offline mode server/networks and those aren’t supported by paper or the wider community.
Aa okay
Single waterfall with for example only luckperms how can have online players?
Do u have lini to this version? https://media.discordapp.net/attachments/802992289566687292/806649011779403816/unknown.png
Link*
DoS exploit going around where they connect 40-100k proxies from foreign country's with no intention on completing the handshake. Best fix is to block the whole subnet/range from joining either through your hosts firewall, or through IPTables. Just an FYI
Russia: https://www.ipdeny.com/ipblocks/data/countries/ru.zone
China: http://www.ipdeny.com/ipblocks/data/countries/cn.zone
Can't be blocked efficiently through Bungee, and somewhat bypasses OVH's anti-DDoS. GL all
@real wraith see https://github.com/PaperMC/Waterfall/pull/603
should already be mitigated
@royal hawk Nope. It stops one of the exploits though which was spamming tons of crap to the server
If you have something else to report that needs fixing, post an issue on github
just saying nope doesnt help anyone @real wraith
What could they possibly do from layer 7?
The biggest thing is gonna be the logger and the general typical # of connections in play
The connection just stalls. Zero data. It times out
that sounds like a typical DoS
You can reduce the logging, but, then you kill the ability for people to be able to monitor stuff in general
a DoS is not an exploit- its a given for any network application
Most proxies I've seen in ages though. Seems fully blocked by just dropping some of his ranges
Only real good fix for stuff like that is blackholing ranges or basically using a connection limiter on the firewall to try to tame the damage of new connections
FivePB chill out. DoS is an exploit. It literally means denial of service. Exploiting something to cause disruption of service to normal users
people use the word 'exploit' too liberally
Yeah. It seems mostly from China and RU. I blocked only around ~10 ranges on a friend's server and it seems too weak to cause any issues now
sure, it's exploiting the typical nature of the internet, but, to call it an exploit per-sei somewhat implies that it's taking advantage of some bug or something
Well it's taking advantage of your machine's inability to handle that many connections at once. Therefor exploiting the application's inability
🤷♂️
Call it what you want, I don't give a fuck. Just spreading the word so people don't get fucked over
Even if your application can hold the onslaught, youre straining your own connection possibly to the fullest so you end up unreachable anyway

What is Waterfall?
guys
when attack get from mcra*** server says is offline but players can join normaly
i got latest waterfall
Unless there's something more interesting in the logs, there's nothing for us to go off there at all
i got botsentry
People already have issues with servers erronously showing as offline in vanilla, 🤷♂️
probably thats why
some plugin is creating duplicate scoreboard teams
delete scoreboard.dat ??
I mean, you can try, but, probably not 🤷♂️
Deleted but created again
anyone having an issue with getting disconnect from servers connected to the lobby with "the server was closed" or "[Proxy]" errors
no, that's not a global issue, that's pretty much 99.999% something with your setup
do you have a suggestion on where to look
i am using latest waterfall, with the lastest paper for a lobby
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
am go shop though
and so foar the issues is with stoneblock 2 server
one sec for pastebin
i am running servers on VMs thru esxi
and in the waterfall i have them directed to the internal ips for each server
hm, fancy pants setup
what
it's not usual to see someone come in here talking about esxi
Failed to set channel option 'CONNECT_TIMEOUT_MILLIS' with value '-1' for channel '[id: 0xafa5c787]'
screams bad config
so would this be server_connect_timeout: -1 line in the config for waterfall
yes
is there a recommended setting
Not really, default is generally fine for most
10-30 seconds depending on the server type
afaik there should be no real reason to need to change that
Forge generally favors longer times
well ill try 30000 and see what happeds
is there a way to force the clients back to the lobby from waterfall or do i need a plugin on each server
Wdym?
so if the server restarts or crashes i want players to be set to lobby instead of server select list
If the server they’re on crashes they get kicked. If you don’t want that then you’ll need a plugin to change that
Yea
ok cool thanks
It looks like that when I use a hostname with a SRV record that I can't use the forced host mechanism anymore.
What's up with that?
the client doesn't pass the hostname but that resolved address from the SRV iirc
Hm ok. I'll try that
Indeed
RIP my DNS entries
xD
I wish there was just a DNS mode that auto adds *.xxxx CNAME xxxx
Where is the link for the latest version of water fall? It seems the link I've been using for my auto updater is broken: https://papermc.io/ci/job/Waterfall/lastSuccessfulBuild/artifact/Waterfall-Proxy/bootstrap/target/Waterfall.jar
is their a forum or guide on the best way to secure waterfall
I have traveled to the far reaches of the intrawebs and have dived into the deepest of forum seas. All i want is to connect 2 identical modded forge servers and i havent found anything that helps in the slightest. Can you guys point me into a direction where i can get help in doing so or help me in doing so?
Hey, what's the new DL url for waterfall?
Hah no way. @lethal igloo we have the same question
@lethal igloo Alright so what did you do for YOUR automation need? This looks like it needs scripting rn.
I used to use that link, but now I'm not sure what link to use. Unfortunately, I'm using a server panel that requires a fixed link, I can't put variables into it.
Yeah. Exactly. Same.
so if you really want, you can use the v1 api which includes a latest endpoint. However this is deprecated and will be removed just like access to jenkins. That would look like this https://papermc.io/api/v1/waterfall/1.16/latest/download.
https://gist.github.com/Xernium/91d67367dbcfef4b18fbe91b74324e06 some things like this is also an option; if you just need the latest jar
Small python3 script for downloading the latest versions of Papermc projects using the v2 API - paperv2-downloader.py
Uses the v2 api
hey guys, what happened here? https://pastebin.com/rq26iKGw
i suddenly got disconnected, any suggestinos are appreciated tho!
some plugin created a duplicate scoreboard team
or basicalyl something along those lines, see the issue tracker.
oki
@ zzzCat would you accept a fix PR that does hotswapping of scoreboards if the issue arises?
the issue is the state desync it induces
you fix the proxy from screaming but you totally botch the state
the issue is that the client still has the old scoreboard
best I got as a fix would be to basically queue up the scoreboard packets until after we've cleared off that thing
I mean, if you can ensure that data retains consistent, I guess so
Yea, which is one of my biggest concerns with stuff like that as it blows the memory usage out of the water potentially
yea, the entire thing is fucked and I basically just, le sigh
ngl, I keep getting more and more tempted to flip that option by default
I mean
We are running in doors here
The only real fix is to treat normal (fast) switches as real switches
Plus
I know about a dozen ways to DoS anything based on bungeecord- just based off of the little amount of care put in by md_5 and you can rest assured- I have no clue how I should fix any of that
Man this is so screwed

Yea, best I got is basically start rewriting a whole chunk of stuff which I really don't have the energy for or the inclination to tell people that, yea, we just broke ur shit
The only real option at this point; well in my book at least; is to retire Waterfall in favor of Velocity
Hard-forking bungeecord for say adventure is possible
But that’s gonna get worse and worse as time goes on
I’ll attempt to mitigate a few more things later this weekend
But I doubt the band-aid solutions are gonna hold
I’ve tried that. You might as well reimplement a large part of the API if you’re doing that
It was miserable if I have to say
just👏 use👏 velocity
supporting adventure 1st party is something I wanna do, but, the entire nature of md's neglect of the API is just, urgh...
Imagine if the API actually returned things like the prefixes instead of the json content...
I mean at this point you could hard-fork the API and include a definition in the plugin yml of which impl to use
Paper+adventure or md’s
you can't really auto-magically pick just based on some setting
and am not doing commodore level BS in waterfall
or just reimplement all in the paper package
^ better option
If I have to say, that option is both cleaner and better in this case
Reimplementing the events at least
@forest halo I’m so sorry for the ping but is there a master plan for how adventure should be implemented in waterfall?
@royal hawk master plan? nope
fix korobi
?
korobi very ded
website ded
also why is korobi.io no more and some random blog site instead
domain expired
y u didnt renew
u didn't want?
get korob.io xD
cuz sum1 bought it and usin it as amateur Minecraft blog lol
report them for squatting 👀
o o f
I just assumed kash didn't want to pay for that unless he really just forgot or something
I remember when suddenly, vc.lq
I don't get this, they are able to host buy a domain, host a blog but don't know the difference between java and bedrock (which their parents had to buy)? This has to be a troll lol https://minecraft.ceo/⛏/gnVjPa47.png
rando tld
Well then what should we do; just add adventure fields to the bungee/md_5 api where Text is used? That’s easy but I don’t think it’s the best of ideas
And @Deprecate all else

If nothing else I can do that this weekend
After that next would be brigadier
That's what we are doing with Paper, only deprecating String AND BungeeCord-Chat methods
what's wrong with just using the adventure platform?
coverage, basically
We want to natively support it, no need for reflection and shit then
but velocity already has native adventure? 
and having a proper component API in waterfall would be nice
Fairly easy to transmit
Like, the API literally just returns json strings in some places
Brigadier is next up after that
Let’s get on that!!
@Deprecated use adventure link Component
We're talking about Waterfall, phoenix
Yea
does waterfall also throw a warning on startup if there is deprecated stuff used like Spigot does?

No, as there is no legacy support BS
Yea tho it’d be a classic md_5 move
Why /server and /send command not working?
Do you have the appropriate permissions and are the modules for those commands installed correctly?
What about permissions?
define "not working"
Was about to add that
I type /server and /send and I get Unkown command. Type "/help" for help.
/bungee works
logs?
.paste
A sensible, modern pastebin. Share text and source code snippets with no hassle.
Screenshot of the F3 screen when connected please
corrupted jars, apparently
The modules go in the modules folder not the plugins folder
Ah, yea...
delete them from the plugins folder
waterfall already manages those for you
They are
Clearly not: Error enabling plugin cmd_server
what's in your plugins folder?
The plugins folder is empty

After updating the jar this problem started happping
What are the start flags and what’s the path you’re starting it from
screenshot the contents of the plguins folder
that's specifically enabling plugins, not modules, it shouldn't be picking up the module jars from there
oh, wait, they do show as plugins essentially due to how it works, ignore that
try redownloading waterfall
something is busted somewhere
I haven’t changed anything in my PRs that would explain it
Okay I will try that
So yes your jar(s) appear to be corrupted

seeing stuff like that which needs to butcher the classpath to work properly is not re-assuring
We must be both blind to not see that right away
Lol I download the new jar file from https://papermc.io/downloads#Waterfall it now works
I would assume so; did you download the last one from songoda?
Don’t ever download these things from unofficial sources
Moreover in this case- I’ll investigate what they added to the projects
As it is a potential security and safety hazard
Well at least the jar is unmodified
Still though; strongly advise against using non-official sites
Does anyone know why trying to setup a network with a 1.8.8 sub-server on Ubuntu doesn't work? (It works fine on my PC, which is Windows 10, but not on any of my VPSes, which are Ubuntu -- one is Ubuntu 18.04.4 LTS and the other is Ubuntu 20.04.2 LTS)
Logs:
[04:43:17 WARN]: Event PostLoginEvent(player=xNitrate) took 291ms to process!
[04:43:17 INFO]: [xNitrate] disconnected with: Could not connect to a default or fallback server, please try again later: io.netty.channel.AbstractChannel$AnnotatedConnectException
[04:43:17 INFO]: [/LAN IP:59840|xNitrate] -> UpstreamBridge has disconnected
Basically, I set it up where the network is the latest WaterFall version with a patch to allow 21w05b clients (it also didn't work with just straight WaterFall 1.16.x). And the lobby is PaperSpigot 1.8.8.
Like I said, it works fine when I host it (both) on my PC (Windows 10) and, for some odd reason, it works when I proxy the connection from one VPS to the other (different networks completely), but when I host it on only one of my VPSes, it gives me the AnnotatedConnectionException. Which is different from the exception it gives when it plain cannot find the server (if the sub-server is offline, port-blocked, etc.)
Any help would really help me out.
Thanks!
P. S. This error happens whether you:
1. Are on the WaterFall server (connected) and you are trying to go to the 1.8.8 server.
2. Are trying to connect to the WaterFall server when it has the 1.8.8 server as a "default" or a "priority".
Or 3. Are being sent from a working Paper server to a 1.8.8 Paper server while being connected to the same WaterFall network.
All of these being when they are all (the WaterFall servers and Paper servers) being hosted on the same Ubuntu VPS.
don't use the external IP for the same host is my guess
?
all that tells us is that the conection failed \
Huh?
The picture is just redundant information that I'm trying to reiterate to say that the exception is different from the "pinging timed out" error.
Which is likely a routing issue if the only difference between the two servers is their host
what IP are you using in waterfall to connect to that server?
You're not understanding my case.
They are on the same network.
I use 0.0.0.0:port
you're saying that waterfall cannot connect to a server on the same VPS?
Yes.
Use 127.0.0.1, 0.0.0.0 is just special to mean "all interfaces"
But, it can to one not on the same VPS.
Just so you know, I do know a lot about networking. :P I almost got my A+.
But, I'll try it.
I was just using 0.0.0.0 because that means localhost and it is on the same VPS, thus localhost.
oh, 0.0.0.0 is supposed to route to local, not something you should apparently rely on, some OSes apparently forbid it, some deal with it weirdly 🤷♂️
Hmm. Okay. It worked on my Test VPS, but it is giving me a "readAddress(..) failed: Connection reset by peer" error on my production VPS.
And it was really weird, when I first logged on to my production VPS, I did /server <1.8 server> and it worked for like 1/2 a sec then it kicked me to the lobby. Now, I can't even get that far.
firewall?
I don't think so.
Plus, I have it ufw allowed anyway.
This is on the server I was trying to connect to. It was being spammed.
ancient server version doesn't support modern java
I have Java 8. Let me see if that works.
Yo. It works.
Dude. Thanks. I legit don't know how to thank you.
I've been trying to deal with this problem for about 1/2 a year now. I'm so stupid not to use the 127.0.0.1 thing... I just thought it was redundant.
But, yeah. Thanks.
There is a way to lower the use of waterfall CPU when so many packets are sent (such as when loading the world). How I understand this depends on packet compression.
On waterfall network_compression_threshold: 256
and -1 in the server paper
Depends. Disabling the on-server compression will only lighten the load a little
Waterfall still will itself compress and decompress incoming packets
Additionally, Waterfall is probably as efficient as Bungeecord gets. If you want anything that is ever slightly less heavy I suggest you look at Velocity
Though, as is, Waterfall uses rather little processing power and is perfectly capable of multithreading/ using multiple cpu cores efficiently
so I really dont recommend modifying the compression on the proxy itself
The problem is that when 200 players switch servers for good or bad, it uses about 7 cores. So the only ways to solve are: either use more bungee or use a tail system to lighten the load.
we are above an i9
make a cpu profile with spark while youre under load and post that here
a single join in a vanilla server (without plugins) uses about half a core
yea thats a bit too much.
How do I assign the permissions for /server /send etc.?
read pvt pls



