#waterfall-help
1 messages · Page 42 of 1
wow
you are running an outdated ProtocolLib
also lots of unecessary plugins (VilalgerOptimizer, SAML, ProtocolSupport and ViaVersion?!?)
ClearLag...
ViaVersion is because we allow players to join on old client version...
yes...
viaversion does the exact opposite
please learn what the plugins you install actually do
Players joining on old clients isn't the issue at hand.
This isn't my server, I'm just an Admin stepping in trying to fix it.
protcol hacking plugins mangly packets
you have two different protocol hack plugins lol
im surprised it doesn't just crash immediately
@frank coral pls fix
I mean, ProtocolSupport + ViaVersion might work (one for backwards and one for forwards compatibility)
using ViaVersion is just really pointless in this case
yeah why though
but yeah, the error is ProtocolLob
*Lib
always update plugins when you update the server...
Using ViaVersion was what allowed us to have players join in 1.16.3 when the server is still on 1.16.1
Either way, that's not the issue at hand
1.16.1 is not supported anyways
oh, you are running an outdated version too
I thought you were running 1.16.3? at least that was your first message...
that's what I get for trusting users smh
The server is still running 1.16.1... the issue is updating to 16.3
It's not a trust issue mate, you just didn't read what I wrote.
well if you need multi version compatibility, as we've said, choose Viaversion + Viabackwards OR ProtocolSupport
" <X1Grant> Updating from 1.16.1 to 1.16.3"
you dont need two plugins that do the same thing
also you should read what we write
lol
Correct, meaning its running 1.16.1 and updating to 1.16.3
I am reading what you are writing... lol
well your server doesn't look updated to 1.16.3 to me
Because it isn't.... We are updating to 1.16.3
He didn't.
well come back once you use a supported version and a sane setup if the error still occurs
I said " Updating to 1.16.1 to 1.16.3" someone he thought we were already running 1.16.3...
Paper is already updated to 1.16.3
Drop the protocol hack plugins, and update to a dev build of ProtocolLib, and come back to #paper-help if you have issues
But players can't login because of the error I posted that started this conversatino.
well if you need multi version compatibility, as we've said, choose Viaversion + Viabackwards OR ProtocolSupport
first step is this
stop using two plugins that do the same thing
Ok
and use 1.16.3
Got it.
Thanks
He reverted to the 1.16.1 build because players couldn't join with the errors I'm trying to resolve now. Ill update ProtocolLib. I appreciate the help.
some plugins using internal classes will need updates from 1.16.1->.2/.3
if the error occurs on 1.16.3 then we will need the 1.16.3 logs, lol
but at this point any plugin that isnt garbage will have an update
Yea, ofc Phoenix my bad for grabbing the wrong logs.
He didn't mention he reverted already before he had to dip. Ill check for the logs he generated when he was working on it.
Drop the protocol hack plugins, and update to a dev build of ProtocolLib, and come back to #paper-help not here if you have issues
but yeah, if ProtocolLib wasn't updated then it's highly likely that it was that
well Proxi the error occured in the proxy server so this is the right channel 👀
the errors cause is a botched setup on paper however
Hello, I have a major question to ask about waterfall (bungee). I know that people can sneak behind the back door and hack into other staff accounts and spam bots caused by bungee. I am leaning to ask if anyone here know the best ways to secure bungee to prevent hackers/spam bots?
firewall
bungeeguard
not using bungee and using velocity + modern forwarding
ah I see ty.
@gaunt breach output of /bungee please. Also; do you use plugins like TAB or SkinsRestorer? Both are known to cause issues
I've been running into issues with a containerized waterfall instance, for some reason, the host of my game server container is invalid to waterfall/bungeecord, i was wondering if this is plain correct or an issue with my setup?
they are on the same docker subnet
and the proxy depends on the game server so the container is created before waterfall has a chance to check, which means that the host should exist
######### Proxy ##########
# https://hub.docker.com/derkades/paper
mc_proxy:
container_name: mc_proxy
networks:
- default
depends_on:
- mc_hku
image: derkades/waterfall
stdin_open: true
volumes: ['/srv/minecraft/proxy:/data']
restart: 'unless-stopped'
ports:
- '25565:25565'
environment:
MEM_LIMIT: "256M"
##########################
########## HKU ###########
# https://hub.docker.com/derkades/paper
mc_hku: # HKU
container_name: mc_hku
networks:
- default
user: 'www-data'
image: derkades/paper:1.16
stdin_open: true
volumes: ['/srv/minecraft/hku:/data']
restart: 'unless-stopped'
# ports:
# - '25566:25565'```
internal port for the game server is 25565
my surprise was when i did expose the game server port to localhost (address: localhost:25565), it was a valid host but it still couldn't connect me to the server
not with docker
i can run infinitely many servers on the same port because the ports are only seen within their own container unless exposed
and the container_name is supposed to be like a hostname/dns record, it's treated like a separate machine, like a vm but a container doesn't have a whole os running, just the software
same thing like if you're running 3 machines, on the machine itself the games can be running on the same ports, because the machines have different internal IP addresses
but the problem is that bungee doesn't see the host
oh i just remembered i had this issue before, and i could fix it by using the IP my server uses on my local network but I would still like to use the container names as hosts
saves exposing the ports
getting this by join waterfall ip
docker should add the hosts automatically without manual intervention @royal hawk
like
with other applications i can just use the container names to connect to the services
disable netty dns in waterfall.yml
Beyond that, ree at bungees logic
thank youuu
testing now
on the proxy...
disabling netty resolver did not solve the issue sadly
yes on proxy or /bungee on the lobby server?
You're on your own then
if you can't even get the bungee command to tell you the info
i must install bungee and waterfall?
bungee and waterfall are the same thing...
They both use the bungee command coz we're not here to just change commands on people
I mean, that command should 100% work
So, you'll need to work out why it doesn't
thats my proxy server with newest waterfall build so thats right or not?
i mean, that looks right
than i dont know why not work in console
for future reference, bungee can see the game server when it's exposed and connect to through network IP
I know
(10.0 is just my router's way of saying 192.168)
It's an issue with how bungee parses DNS
i'm sure this has been encountered numerous times before, did md_5 put out a statement or anything about this?
like wontfix or something
🤷♂️
OH
??
yea, that would be it
seems like md_5 doesn't wanna fix it
Well, _ is technically invalid
yea
DNS names can contain only alphabetical characters (A-Z), numeric characters (0-9), the minus sign (-), and the period (.)
This server is running Waterfall version git:Waterfall-Bootstrap:1.16-R0.4-SNAPSHOT:c1ac7d5:377 by md_5
That is right?
yes
ok hmm
ill just name the containers without underscores
that is the line what i get when type /bungee
:/
why would you use invalid domains in the first place?
docker is fun
than maybe the issue on the proxy is the one server what i have call all time im not whitelistet and on proxy maybe they show me this with the exeption error
yesssssss
kick packets aren't sent in the login phase, which is where your exception is coming from
Once again, replicate without plugins.
@acoustic adder using underscores is very common in docker
sounds like a dumb thing to do
dumb but the norm
i have no plugins when you mean me
You have no plugins on the proxy and on the server?
before recently i was forced to, to make sure there weren't two containers with the same name, i'd prefix the containers with the stack they belonged to
both new created
but i recently separated all my stacks into separate compose files so
You did enable bungeecord mode properly?
spigot.yml bungee: true
i can just give them epic names
and on paper.yml bungee online mode: true
if the "norm" is violating the standards then I would think a long time about whether or not I want to really use such programs
It's not really a docker issue
maybe this is wrong?
yeah, just don't use underscroes in domains?!?
docker doesn't care about URI standards because the container names aren't domains
More, people blindly following what others do because 🤷♂️
I'm sure docker lets you configure that
Docker doesn't set container names, that's up to the creators
(of the container instances)
if it looks like a domain and is treated like a domain then it for sure is a domain
thats what i meant, container names aren't URIs, and valid in most other circumstances
they're host entries
like localhost
never seen a host name with underscores 👀
i mean if you dont use docker you wont
you can use hyphens
that's a good alternative and will be using that in the future
maybe I just don't configure my docker in non-standard ways?
fair
maybe found no the problem with the package. Lobby servers ip is localhost but everytime i tried change server.properities after restart get again localhost so must writing my hoster to fix it
any idea why a server wouldn't be receiving the right UUIDs & IP addresses?
ip-forwarding is on, bungeecord is set to true
Online mode?
Check mc auth server status, they announced some downtimes lately.
the UUID/IP is passed from the proxy in the handshake packet
if it's missing that, errr...
Well,if it was missing that it would fail, are you proxying the proxy or something?
@reef fulcrum only tcpshield > travertine > backend server in vrack
users don't get the online mode UUID & get the IP address of the proxy server (192.168.1.1)
proxy gets the user's IP just fine oddly enough
even luckperms on the proxy, gets it just fine
Where was that message from?
o.O
i'm fucking clueless yes
what plugins are on the server? are you sure something on the proxy isn't meddling with the handshake too?
as, that's odd
even tried 0 plugins
with bungee mode, it shouldn't even let you connect without that
What was the issue? 
Hi guys, don't you know how could I fix this please?
not use a plugin that does that?
velocity doesnt have that issue :')
what does velocity do? kick the player?
this channel should be renamed to use-velocity-if-any-issues
Even better:
Buy-some-mcmarket-skidded-tavertine-fork-if-any-issues
eww
What was the issue?
Wondering this myself
who the hell cares?
Velocity is a fork of Travertine? If so, buy me in!
no lol
it's its own proxy
velocity 1.1.0 works for 1.7-1.16.3 tho
1.0 works on 1.8-1.16.3
So it's not a fork but supports 1.7?
yes
written by @unreal stag, who could definitely sell you more, as he originally started Waterfall, and writes Velocity
I wonder when we'll see a velocity fork on mcmarket for 90$, advertised as performance boosted bungee
Perfect, time to learn its internals and make my own fork aswell.
I have seen some claim that, Simon.
i mean ngl it's definitely already sold there
but no need forking velocity, he's happy to take in anything that doesn't fit as a plugin
i shall now cry myself to sleep
my client had a fork with an ip whitelist built in
Do you run a hosting or?
Michael, I have already fucked up my 1.8.8 fork by changing how Passengers work and lazily copy 1.16's method.
Oh, makes sense
👍
Not bad, Proximyst
someone please take a look
5gb bungee crashing due to out of memory
im not sure what is wrong
can you sort by retained heap on suspect 1?
or is that the third screenshot?
but from the looks of it: something is creating 4GB of server config objects? Oo
looking at the code that is creating and touching these objects at that location would be what I would try first
I guess you could get better results if you have a debugger attached to directly go to the object or something like that? not sure, never tried that one before
Wait should I use velocity if I have this issue?
but I don't need 1.7 support, so don't know
but wanna solve this issue
Do you think it's better opinion than normal bungee proxy?
- don't ping people
- velocity is a completely different proxy software which is not compatible with bungee/waterfall plugins
it claims to have improved speed, stability and security though
if only have really simple proxy plugins and can find velocity alternatives for them then it might be worth a try
@PhanaticD Maybe some plugin is trashing packetQueue
Check how many instances of PluginMessage do you have
So normal plugins don't work there? :/
And don't you have another idea how to solve my problem phoenix? How to kick players after they quit? Maybe isn't plugin for it? Or some option somewhere,
velocity has """normal""" plugins too...
chances are, what you need is available on it already
alternatives to what?
if plugins won't work on Velocity, if you have alternative to solve mentioned message that's happening to players
BTW it's also maybe due I'm using securednetwork, but author of it doesn't think it's caused by his plugin
only one way to find out..
@delicate phoenix the sad part is that some of these forks have already started doing that
Yes, go look in the Velocity Discord, I found some
This is why I’m starting to put them into Tuinity
Tux, I just built Velocity and then ran it. Can say I love it but still have to learn how that works given it does not support BungeeCord and have to rewrite my plugin for it, any hints to start somewhere? And I have noticed that by default it does not support 1.7.
Is Travertines Perfromance worse that Waterfalls
there are builds on the page, onium
it links to the jenkins which you can use for 1.1.0 builds (with 1.7 support)
the documentation should be enough to start, https://jd.velocitypowered.com/1.1.0/ iirc for 1.1 jds
Is Travertines Perfromance worse that Waterfalls
yesthought that said Velocity
Yes, go look in the Velocity Discord, I found some
I see, what the, 90$?
@gaunt breach output of /bungee please. Also; do you use plugins like TAB or SkinsRestorer? Both are known to cause issues
@royal hawk no i don't and my /bungee output is
This server is running Waterfall version git:Waterfall-Bootstrap:1.16-R0.4-SNAPSHOT:f7d11c8:376 by md_5
Proximyst, I got the git version (few hours ago) and am using it... Sure, let me download it ffom there then!
@cold mantle checkout the dev/1.1.0 branch and build that or download the 1.1.0 builds from Jenkins - those do have 1.7 support
@crude pike after taking a look at that, i think it might be premiumvanish
@green palm how is the performance worse? this is the first ive ever heard that
hmm, i have no idea, but i read that as "is travertines performance worse than velocity"
@spiral crater performance will probably be more the same, probably a tiny bit better for 1.7 clients. But Velocity 1.1.0 still handily beats Waterfall and Travertine performance-wise.
oh i thought waterfall and travertine were being compared
which is what the guy asked xd
Phanatic, your dumps 100% point to the plugin message relay
bungee is fucking awesome and does this thing where if you send a message to all servers, it will queue up those messages; Now, ofc, plugin devs are 10/10 and go to read/understand the docs, and so they shoot messages around the entire network with this "ALL" target, which just queues them all up on the proxy until it can send them
Anyone any idea what could cause this https://hasteb.in/apefovum.css it's my first time seeing it
people trying to send dumb shit to the server
drop the traffic from teh offending IIs in your firewall e.g. with something like fail2ban
i'm using a shared host
then the answer is probably: use velocity
(or don't use shared hosting)
or ask your host to defend against such attacks
alright thanks
permission plugins will override what's in the config
You'll need to set the perms up in LP
yes
okay thanks sir
and I have a question too, am I need another plugin to secure my server? I have IP Whitelist
Use bungeeguard
IP whitelist plugins are honestly generally flawed in how their implemented
will bungeeguard work with waterfall? and am I need to delete ip whitelist?
Yes, and I would
okay
why I might need multiple proxy? can you say me a reason? I don't need but I wonder
load balancing
how can I find more information about this topic? what can I type?
aka, multiple proxies pointing to the same network of servers
the proxy can only handle a few dozen hundred connections before it stalls out essentially, so, you just spread them across proxies
hm okay
what port proxy should have
and the main spigot server
i made it 25565 for proxy, and 25566 for spigot
doesn't matter for spigot so long as ports match up, 65 being the default port makes sense to use that one for clients to connect
if i have multiple proxies, then what I must do, which one should be 25565
Having an issue with one of my players not being able to chat on one of my servers, but can chat on all of the other servers part of the network. Permissions appear all fine; to my knowledge, none of the plugins should be interfering with it - tried googling it extensively, but it doesn't come up with anything besides parent/child restricted accounts/disabled chat - but he's able to chat on other servers part of the waterfall network/other different servers.
Are you using any chat formatting/modifying plugin?
Anything as little as some factions id placheholder for as an example.
I am - venturechat. Even setting him to different ranks (that can chat) do not work. He's not muted, no background punishments that may be preventing chat, no neglected permissions preventing it.
There's your issue. Turn off venture, and see if they can chat with that in that specific server.
If yes, than please contact venture, and report the issue.
Yep, it's not venturechat - had already tested that. 😀 Only thing I've not done, is remove each plugin and then re-add one by one.
Just not sure why the identical server, exact permission setup, plugins, configuration, does not have the same issue. 🙁
Are you sure you don't have any other plugin that modiefyes the chat?
Please check config files, I have found weird chat modifying plugins.
For as an example, redprotect, an antigrief/protection tool have messed with whitelist and chat without I even knowing 
Essentials (with /mute), UltraPunishments (Also with mutes), Skript can manage chat, but no chat events are there - CoreProtect to my knowledge does not handle/manage chat besides logging.
It must be a plugin issue, copy your files to your local masine, and try removing plugins 1:1.
No other choiches left sadly.
I would suggest to check the perrmisons with luckoperms verbose, but since you are using ultraperms, we can't do that.
I'm using Luckperms - just UltraPunishment for ban/mutes/warns.
@leaden igloo Are the player messages shown in the logs? And are there any errors?
Let me check.
You should prob' use luckperms' verbose feature to see if this isn't a permission issue
What would I be looking for when running that? Turn it on & have them say something?
Will get back to you as soon as I'm able to test with the player.
sure, @ me when it's done ^^
@still notch Tested - he doesn't appear in verbose when trying to chat.
Does waterfall link the minecraft chat?
@winged bridge I don't think so, but you can use plugins like BungeeChat to link your chats
@winged bridge You can force global chat in the plugin's config file
:O really?
Yep, just checked on github: # Global chat information. GlobalChat { aliases=[ g ] default=false enabled=true ]
Yeah, i guess: LocalChat { aliases=[ l ] enabled=true # This causes BungeeChat to still log transparently passed messages logTransparentLocal=false # List of servers we want to pass local chat through to passThruServerList {
Just change enabled=true to false in the LocalChat part, then set default=true in GlobalChat part
🙂
😄
can anyone tell me what values they have for waterfall running 16.3 servers
remote_ping_cache: -1
network_compression_threshold: 256
timeout: 30000
remote_ping_timeout: 5000
connection_throttle: 4000
thank you!
default?
just looking for stable values, idk if those are default or not ill need to compare with whats on the forums, just having network issues where players can connect but as soon as they attempt to run a command they get disconnected immediately with no trace
Is it possible to add bungeecord.command.server without giving access to all servers?
were you they guy who just asked that in #md_5 on irc.spi.gt? 
did you set the restricted option?
bungeecord.command.server.<lobby>
bungeecord.server.<lobby>
also did you make sure that you didn't have all permissions when testing it?
if i set perm bungeecord.command.server it allows my to access all server
that gives you access to the command
i remove that perm, then set the bungeecord.command.server.<lobby>
you need to set the servers to restricted, and then it will check the perm
https://pastebin.com/0j3Z8pVV My bungee isnt working atm pls help
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.
you put Vault on your proxy
Vault is a plugin for bukkit spigot or paper but not bungee
Bungee is not a server software
It’s a proxy
Via and Luckperms as well as that staff plugin there are designed to (also) run on bungeecord
so just remove vault?
Yea but this shouldn’t hinder your bungee working
ye I did it and it still isnt up, it was running last night idk what changed tbh
Aside form that error the log didn’t show any other issues
https://pastebin.com/0YqYhpa2 Heres a new one?
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.
define "not working"
quick question
if i restart bungee server will it disconnect people that are in other server?
if by "other server" you mean another bungee server, no, why would it?
hmmm,well when i restart my bungee,it disconnect poeple that are on survival and my skyblock
nvm
Hi guys, I now discovered it happens when someone join server and before he is joined there (he see void mainly due to for ex. bigger view distance), then he is "joined to server" even if he left... and then he can't join, he is stuck there.
Any idea to solve it? Tried Velocity, but still doesn't support my plugins :/
99.9999% of the time, that's caused by plugins
alright
i cant join
my server
[16:21:30 INFO]: Senghow[/69.420.69.420:51520] logged in with entity id 45 at ([SHub]-1017.4848057869897, 30.0, 428.50078734119126)
[16:21:31 INFO]: Senghow lost connection: Internal Exception: java.io.IOException: Error while write(...): Broken pipe
[16:21:31 INFO]: Senghow left the game.
a
b
A pipe broke! You need to call a plumber to fix it See if this works without plugins
broken pipe means that the connection was closed in an unexpected manner
Go check your logs on the proxy, etc
A pipe broke! You need to call a plumber to fix itSee if this works without plugins
It doesn’t, they tried on the ViaVersion Discord
[16:23:16 WARN]: [/ipees|Senghow] -> UpstreamBridge - read timed out
[16:23:16 INFO]: [Senghow] disconnected with: ReadTimeoutException : null
[16:23:16 INFO]: [/ipees|Senghow] -> UpstreamBridge has disconnected
[16:23:16 INFO]: [/ipees|Senghow] <-> DownstreamBridge <-> [hub] has disconnected
[16:24:05 INFO]: [/ipees] <-> InitialHandler has pinged
UpstreamBridge - read timed out
So, the connection timed out
You'd need to work out why
.895
Not the exact same issue, but, it's similar enough that most of the advice there is going to be applicable
a
first thing to do is to check the internet
ethernet is better i meant if its good on both sides
Cool, but that doesn't mean shit if you're having latency issues for some reason, etc, etc
pretty sure my wifi isnt high on drugs
What about the server’s wifi (if you use a host)
Okay, so, heres the deal
I gave you a list of stuff to go check
if you're just gonna be dismissive/argumentive over step 1, you're on your own
ic
1gbit != wifi will never act up
Hey everyone!
Today we were hosting an event and we started to receive lots of PreLoginEvents from lots of different addresses, taking down our bungee
Any idea how to mitigate/fix the issue?
Shamefully, there is a level of "the proxies sole job is to accept and process those connections", I would work out why those events are taking so long as that will only propagate the issue much more than it needs to be
We are sure this is an attack on the bungee, we believe they are using cpu expensive payload that makes the bungee start timing out other connections
We are using one of the latest versions of travertine btw
I don't see how a CPU expensive payload would apply there
It's literally saying that the plugins listening to events are taking too long
bungee timings wen
There is also that issue with "a proxies job is to accept connections", there are some things which can be done to try to improve it, but, at the end of the day, it's where the weakest link is; Configure your firewall
never
😠
Timings doesn't work with threaded environments
make it 😠
It's literally saying that the plugins listening to events are taking too long
@reef fulcrum Are you saying that one of our plugins is causing the issue?
Again, this is affecting the whole bungee instance
It's definetly not helping the issue at all if plugins are stalling
We don't believe we have any plugin that would stall this event
This all happens on the event loop for the networking stuff, so those events taking seconds to process will impact the proxy
use a profiler
Any suggestions for profilers?
ideally something like yourkit or visualvm that attaches, but, there is spark if you need a plugin variant
The issue is we are not really able to replicate this, since this is an attack that went on for a few moments on the afternoon, and is no longer (and has never before been) an issue
i'll get spark loaded and if i see it happen again try to profile it, but i'm sure this is an exploit being used to take down the proxy rather than a plugin issue
either way, thanks for the help
The issue is that exploits aren't easy to deal with
All those clients have done is sent a handshake to the server and a login request, which contains the username, as seen in there
There is not much going on there, but, when you deal with multiple dozens of connections at once, especially when you have stuff like the event calls taking too long, that heavily compounds the light work
Recommend looking into limiting how fast connections in general can happen to the server with a firewall, iptables afaik can do it if you've got access to the OS
What i wonder is, would it be possible to mitigate this by enforcing limits/thresholds on new logins or server list pings (because i've seen those used for attacks in the past too)? I'd rather have players who are faced with a "Sorry can't process your login right now, try later" than literally timing out everyone online
The issue for that one is how far we have to allow connections to go in the first place
That's better off being done at a different level to the proxy who's job is already to handle a few dozen things as is
DDoS attacks aim on the weakest link
The weakest link used to be the connections to the servers, now with DDoS mitigation services, that's no longer a weak link, it's cheaper to get bots to do the job, which means that the next weakest link is the target, in this case, it's the software which has to deal with the logic to ensure that players can join in the first place
This seems like a very specific layer 7 attack, i do not see how any layer before the bungee is going to be able to differentiate this traffic from legit players
a firewall rule can not be made to ignore "minecraft login packets past this threshold" because they're all just TCP packets, it can't filter what they are
There is 0 signs that there is mallicious traffic being sent
the only real thing you can do at a higher level is aim to reduce the impact to the lower levels
We already have several layers of ddos protection and firewalls, wich are all being bypassed, because this specific attack is not sending "that many" packets
This is more the number of connections than the size/amount of data being sent per connection
Firewalls/DDoS protection services generally don't do well at mitigating this type of stuff, there was that DDoS protection service, but, afaik they're somewhat also having issues trying to resolve these ones
you can see the logs however, we're looking at 5 connections per second at most?
There was the flood of connections opening before that
all from different ips
chances are that was how many connections it took before the entire network system just started chugging
those are legitimate players however
this was after a restart where many people were rejoining the server
At that rate? That suggests more that you overloaded the system than an attack
The proxy is not really good at dealing with large numbers of connections opening at once due to all the work it has to do to process those connections, e.g. contacting mojangs servers
this has never happened before, we've had x3 times the players we had today, this is not an overload
More just looks like you need to look into load balancing than anything
it was way too specific to be an issue of performance, really
But, if these are "legit players", your entire story of "these are people using an exploit" goes out the window
there are many connections of legit players before the attack
there is a very specific point in time where the attack starts and all server becomes unresponsive, a few seconds after, you're timed out
We can't see the difference between legit and unlegit
it is the same time where yellow text starts appearing on console
I can't even see where the first warning messages connection was opened as it's not in the screenshot
i'm aware the software can't tell, what i'm saying is it should probably be limited regardless and just have the least bad of the outcomes, that is, not letting players (who may be legit) in
rather than timing out everyone online
I have very little interest in adding rate limiting to the proxy given that it would just appear like a "connection refused" rather than a pretty "try again" message
There is an event which I added to waterfall which would make it easy for a plugin to deal with that if they really want to
what is such event?
if you don't mind the connection refused messages, doing it with the firewall instead of in the software would be a much better option
ConnectionInitEvent, which is the waterfall one and will let you do stuff off the event threads if you want to, bungee also added ClientConnectEvent, which they take pride on their naming of that event but is done on the network event loop
thanks for the info i'll try to look into options, we've also found https://www.spigotmc.org/resources/pinglimiter-protection-against-layer-7-attacks-bungee-waterfall.72596/updates wich could help, issue is we won't be ever able to find out if it works, we'll only know if it doesn't work which... would be bad to find out
That's irreelvant for your issue
I mean, those two rando things might potentially help, but, at the same point, I could take advantage of that to take your server down
well the idea is that by adding this the bot used to take the server down would need to be more sophisticated
It won't
Those events will still be fired up to that point
best case is it's a plugin making it 20x worse and that hits in at the right spot that it makes it better
:/
worse case, the bots make the trivial changes to work around that, or, take advantage of that to take your server down in more fun ways
the pre login events will simply get cancelled by the plugin, is it just too late to do so?
https://github.com/Futurixel/PingLimiter/blob/master/src/fr/futurixel/zencraft/proxy/pinglimiter/PingLimiter.java#L110 and they literally flood the scheduler with tasks in the event that that type of stuff happens... nice
You're literally at the point where the prelogin event is taking 5 seconds on the network threads, that means that the connections are being allowed to at least that point; a plugin like that /might/ help in some respects, but, it has its own potential to cause its own set of issues
do i just replace the bungeecord.jar to waterfall.jar to change it?After that will i need to redo the setting ?
yeah just replace the jar
I started to getting "OutOfDirectMemoryError" when try to connect to the proxy after some hours of running. I can fix it temporally by restarting server, but after some hours, it'll happen again, and I didn't found fix for it.
We can't really help debug that, something is leaking memory somewhere
It's randomly happening, that it'll disconnect all players and I can only restart it, don't know what is causing it.
the actual full error might give some picture of what's trying to allocate that memory, but, as for what's actually taking up that memory, we can't see
how many players? are you having a flood attack?
I have no idea, there is defo something funky as there is no way I can see that stalling on 8 players, but, something is eating native memory
best guess would be replicate without plugins, otherwise, need to get into proper debugging which is "fun"
It actually doesn't matter how many players, it is happening constantly after approx. 24h, it can happen, when on the server is like 2 or 15 players, same issue happen again
The issue is that all I can see is that it's listening to a ping from the client, the number of threads processing those requests is consistent, so, they shouldn't ever expand beyond like 128MB iirc given the default netty thread count of 8, you're eating almost a gig in native memory
You can add -Dio.netty.leakDetectionLevel=advanced to try to help picking up any leaks that might be in waterfall, but, that doesn't cover direct memory as a whole, nor is this something that I see often enough to remember, hence the err over plugins; That might help stuff, but, beyond that, we'd need means to reproduce it; There are debugging steps to take, but, we can't really help with those
Project Waterfall build #378: SUCCESS in 3 min 33 sec: https://papermc.io/ci/job/Waterfall/378/
Project Waterfall build #379: SUCCESS in 3 min 8 sec: https://papermc.io/ci/job/Waterfall/379/
I had been getting the "Don't know how to turn {...} into component error" a few weeks ago. I switched from bungee to waterfall and that seemed to fix the issue... until now. Started getting it again.
It's a bungee chat issue afawk, we have no means of reproducing it so debugging it is basically a no-go
current understanding is that mojang changed how something works and bungee doesn't reflect that properly
When a player joins, they leave and rejoin. It says that he is already connected, but can still join. Does anyone know of that?
I have bungee enabled in sponge config, I have forge enabled in waterfalls config yet I am getting this error on the Sponge server [Netty Epoll Server IO #1/ERROR] [FML]: NetworkDispatcher exception
It is saying I am timing out
Do you have forwarding enabled?
Do you have both
modules.bungeecord and
bungeecord.ip-forwarding
Set to true in the sponge config?
Yes I do
Is there an error in the bungeecord console ? If you have error logs please upload them and share the link here
.paste
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
.... [16:41:20 WARN] [ViaBackwards]: Ignoring outgoing plugin message with channel: legacy:fml|hs
.... [16:41:20 WARN] [ViaBackwards]: Ignoring outgoing plugin message with channel: legacy:fml|hs
This is all I get in the console
and
.... [16:41:18 INFO]: [Korgen0|/REDACTED] <-> ServerConnector [REDACTED] has connected
Uh so idk how to say this
But Via doesn’t work with mods
It runs on sponge yes
But it will defer mods
No its running on the proxy
And I am connecting on 1.12.2
The correct version
For the sponge server
Yes.
Since this is legacy it will discard the forge data
Remove Via from the proxy and try again
If it then works please install Via on the servers that are not forge based instead
Alright I'll try that later
When less people are online my server
Thnaks
Thanks*
@mint remnant this may help you
iptables -I INPUT -p tcp --dport 25565 -m state --state NEW -m recent --set
iptables -I INPUT -p tcp --dport 25565 -m state --state NEW -m recent --update --seconds 10 --hitcount 5 -j REJECT
iptables -N MCLIMIT
iptables -A INPUT -p tcp --dport 25565 -m state --state NEW -j MCLIMIT
iptables -A MCLIMIT -m limit --limit 10/s --limit-burst 20 -j ACCEPT
iptables -A MCLIMIT -j LOG --log-prefix "MCLIMIT: "
iptables -A MCLIMIT -j REJECT
first two rules make a IP only able to make 5 new connections per 10 seconds
second set makes only 10 new connections allowed per second from any IP, burstable to 20
Hello I have two questions to ask.
- When i setup bungee and have the config direct set to /hub (It said unknown command, I believe I must be doing something wrong?) I have to set to that on priorities. I am looking for a way so players can type /hub to spawn back to the hub.
priorities:
- hub
- When a player is on a different server then try logging off then logging back on takes them back to the hub and I am looking for a way to prevent them so that when they log back on they will be on the same server instead of the hub. force_default_server is set to false as default so shouldn't that not force them to the hub when they log back in?
forced_hosts:
exampledomain.com: hub
I am looking for a way so players can type /hub to spawn back to the hub.
use a plugin for that
ah I see hmm, and as for other question feel free to (@) to notify me! Thanks.
force hosts override that setting, given the whole "forced" nature
w h a t
if you connect using a forced host, you connect to a server defined in the forced_hosts section
if force_default_server is off, then it should connect you to the last joined server
Yep that what I was saying. force_default_server was that to false as default so shouldn't they be connected to the last joined server?
if you connect using a forced host, you connect to a server defined in the
forced_hostssection
regardless offorce_default_server
you connect thru pvp.myserver.com, in the config it's a forced host should connect you to pvp, and it will connect you to pvp no matter what
you connect thru myserver.com, and if it's not a forced host, and force_default_server is off, then it will connect you to the last joined server
Yep but it doesn't.
My is set to false.
My guess is something interfering it or a bug?
check for proxy plugins first
None of my plugin should be interfering with it.
you never know
is "exampledomain.com" actually your server name? Well, what you've got configured there?
Nope that not my domain. It just an example lol
Hello Cat, May I send you the link of pastebin through DM?
What I mean is, is your domain in there pointing to the hub?
Coz, if so, once again, that will override that setting
(i mean, it's not even overriding that setting, it'd doing what you told it to)
yes
ah kk ty.
As this question when l1ttle mention i need a plugin for /hub to spawn back to a hub? https://discordapp.com/channels/289587909051416579/555462214606389275/760012473435684924 Is there any good recommendation plugin does it?
oh nvm I found one I believe.
How can i protect my network? How can i install firewall?¿
Or something more secure
search for guides on iptables or ufw
The most secure way is removing the internet acess of your box 
please update upstream of travertine
Project Travertine build #162: SUCCESS in 2 min 4 sec: https://papermc.io/ci/job/Travertine/162/
hello
i have a problem with 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.
i cant get in the server , its even down
.318
im using the latest tho
see the issue cat sent
mhhm scoreboards
aint even using a scoreboard
anyone knows some famous plugins that uses scoreboard
also it happens rarely like once a day but the time it happens is different
Hello, I was wondering if WaterFall bungee have a setting that will put people back to spawn point from (/spawn) when they log out of the hub and rejoin the hub?
The proxy itself doesn't have a setting for that but EssentialsSpawn does if you have that on the hub
ah I see. ty.
Hello!
mihai: you either have the same team on two servers or have a plugin adding duplicate teams
Guys, don't you know someone how to solve issue with "Already connected to this proxy?" Something that can kick players when they leave, to prevent it? Can't use Velocity, there aren't all my plugins supported yet
Find out what plugins causing it
Hopefully it will be due to plugins
in my many years of dealing with people expecting magical fixes for that issue, 90% of the time it's been attributed to plugins, 10% of the time I never heard from them again
This is literally a "nobody can debug this for you", we don't know what you're running on there and what it all does
did geyser mc and paper mc partner up?
no
Why would ip-forwarding be disabled by default? Just curious because you'd want the real IP & UUID to be forward to the backend servers, right? This is just one of those shower thoughts 😂
Because backend server might not support that and bail out on garbage in hostname field
backend servers do do that unless configured for forwarding, part of it might just be the aspects of getting a lazy testing setup, as well as maintaining compat with the protocol by default
You have to configure all that anyways, so I hardly consider it much of a "out of the way" thing to configure
hey
i am having a weird issue with waterfall
i dont know if its my server instability or not but im unable to return to the lobby if i am on the vanilla 1.16.3 server
this is what it spits on the waterfall screen
this is what i get in game
all the other servers work extremely fine
some plugin broke your proxy
no thats a plugin
🤧
@high dagger a plugin youre using wasn’t updated to deal with 1.16 yet
anyone know how i can prevent this? I have travertine latest build.
i've disabled it in the config
restart your proxy then
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
probably just isnt reading the right config then i'd say 
“playerbalancer” is being dumb
Hello, i am having some trouble trying to find why doesn't my networks functions, and i'd like to know if, even if the proxy doesn't manage to send the player to a server, does something happens in the logs when a player tries to connect? Just so that i know if it's a network problem
Hey there, I was wondering if there is any docs or diagrams about how waterfall works
not really, at it's simpliest it's a dumb proxy with some "smart" hacks to coax it to work
inspecting source is the best bet, as it's really all down to the tricks to coax it to work
velocity would also be a better source review as they do stuff in a less convoluted way
e.g. no entity metadata bs
Enabled Waterfall version git:Waterfall-Bootstrap:1.16-R0.4-SNAPSHOT:691ffb1:379 why is the sha hash wrong? Apparently 379 IS the latest build, but I needed the website to know that. master is on 27e2885
oh, that makes sense, ty
running latest waterfall
someone attacked my server today and now it reports back, that mojangs auth servers arent aviable and no one can login
restarting the waterfall proxy didnt solved the problem, i had to restart all servers
https://pastebin.com/J7biGAz6
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.
@frosty finch looks like someone flooded your server to prevent people from authenticating, you will probably have to change IPs and consider using a rate limit on logins.
is there a plugin for limiting those rates?
i did, havent found anything yet, but thanks for the hint, i appreciate that
Hello!
I've seen an old video, where mineplex had an animated motd icon, as well as animated lines. I found a plugin for this purpose, but it's abandoned.
My question is, that why doesn't animated motd exists nowdays? Mojang blocked it? Is it to hard to make work? It was an exploit?
Because mojang fixed it
Fixed it? I tought it was a feature, not a bug.
It was a bug
You're not supposed to be able to keep sending that data to the client
it's supposed to be a one shot and close
What's the server that Waterfall uses for force_default_server?
it uses the priorities list vs joining the last server they where on
thanks!
Optimized & Recommended JVM Startup Flags: https://mcflags.emc.gs
is there a tutorial for all the options on waterfall >?
on the paper docs, yea
@cold haven waterfall flags are pinned
Fucking weird shits going on. If a player uses /server build from anarchy, it says you are already connected to this server. if said player uses the command in superflat, it takes em to anarchy. ive restarted waterfall, and checked my config.
how do i troubleshoot this?
oh and i can somehow go between worlds just fine somehow
but 2 other ops cant
did you /greload or something?
/server literally just grabs the server from the config and sends you there, so 🤷♂️
nope
im restarting
every time im restarting
i have never in my life used greload
heres my config
all servers are on the same machine
When you /server, it just connects you to that server, outside of some plugin or it somehow managing to hit the priorities system because of some oddball failure, 🤷♂️
best guess would be check the logs, and if you have plugins, try without them
imma turn off all plugins and try
ok it is a plugin
imma try enabling geyser first
do i have to setup iptables for waterfall ?
yes
yes
why do u think you dont need to
i... i disabled all plugins, enabled em one by one to the point where im back to all plugins enabled and its working
what the fuck
what
edit: viaversion wasnt installed on the build server and my both builders were on 1.16.1
what the fuck was the cause of my issue? like what
maybe the trick was to do 15 restarts
how can i make waterfall redirect you to the default server if one other servers closes ?
do i need a plugin ?
yes, one was literally just shoved in #755916528901161031, there's also stuff like MoveMeNow
i just want something that disconnects you
can i use the ufw firewall instead of ip tables ?
you just need a properly configured firewall
You need to ensure that your backend servers aren't accessible
be that binding them all to localhost or some interface (in the case of docker), or, firewall (or, some plugin if you really must...)
bungeeguard + iptables was what you recommended last time to me
but ufw should work as well right?
ufw is literally an iptables wrapper
cool. thanks!
I tried to remove my plugins and it's still happening https://discordapp.com/channels/289587909051416579/555462214606389275/760536054272491540
@meager citrus all of em?
also, please share your config.yml
And make sure if youre using something like via version its installed on all servers
And if youre not, make sure all servers are the same version of minecraft
I have it on all servers except bungee
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, i set up a waterfall server. The lobby is running 1.16 with viaversion and viabackwards. It works fine, now i want that people with mods can join my forge server (1.15.2) over the lobby server. They can join with 1.15.2 on the lobby, but then cant join the forge server. The error in the console is this:
[14:36:39 ERROR]: [StaticRed] <-> ServerConnector [modded-1] - encountered exception: net.md_5.bungee.util.QuietException: Unexpected packet received during server login process!
1bc7017b227472616e736c617465223a
I set forge-support to true in the config.yml
@amber pine Forge past 1.12.2 is not supported by any proxy. There’s some experimental work I’ve heard of but it is only for Velocity.
best is to just connect people directly to the forge server instead of going through the lobby
Hello!
I just made a fresh tavertine server, and getting this issue:
[/172.18.0.1:54888] <-> InitialHandler - encountered exception: net.md_5.bungee.util.QuietException: Unexpected packet received during login process! 0100000397030000
have you tried velocity?
No plugins on tavertine, no plugins on the backend
Nope, I would like to use waterfall
do you have any plugins or mods installed?
No
so just a plain paper server and a plain travertine sproxy?
Yes.
Why would you not use velocity if you have no need for plugins
Velocity is just as simple and better, it just doesn't have plugin support yet
you worded it quite badly, lol
more like it does not have very large plugin ecosystem yet
Yeah Im a little tired
Velocity is perfect for your usecase. While not many developers focus on it yet, it doesn't really matter for you since you are pure vanilla
Better?
no
Then you write it 
I don't know if that will help you, but Waterfall can still crash or overflood or whatever. https://pastebin.com/tZxww53N
It was kind of a dose, but only on the bungee. After I tried the so-called "Aegis" all problems were solved. But I hope that there are other solutions because I don't want to use this fork.
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.
@meager citrus when do you face the "Youre already connected to this server" error?
Nini, see the pins and configure a proper firewall
Those forks generally don't fix shit other than on the surface and create their own types of issues
Also,
connection timed out: sessionserver.mojang.com/143.204.212.175:443
That's nothing to do with waterfall at all, that's basically network level issues, overloaded?
yeah maybe overloaded by some exploid or something i try the pins configure now lets see. thanks for helping
Anyone know how to get command names to auto complete? I set up tab completion for command arguments just fine, but can't seem to figure out how to get the actual /command to autocomplete
unless you disabled tab completion on the server, it works fine
I've tried both true and false for disable_modern_tab_limiter and doesn't make a difference. Not seeing any other settings related to tab completion in waterfall.yml or config.yml
you want that on true, otherwise it breaks tab completion on modern clients
/command to autocomplete relies on the server sending those commands
the proxy hijacks the command definition packet to add its commands
Even the default commands /server, /glist, etc don't auto complete
can you run them?
can you complete /version? do you have plugins on the proxy?
Yeah all paper/spigot commands auto complete
One plugin on the proxy which is the one I just started to develop to do a basic /hub and /servers command. So nothing canceling events
https://discordapp.com/channels/289587909051416579/555462214606389275/762149482024271874 @tender axle when someone join server when it's loading and then disconnect / is kicked for AFK, tried to remove plugis from bungee and sub servers and sitll was happening
Those timings are from paper not from waterfall. Waterfall is a proxy not a server
Hmm, way to make the player list on ping hover like without proxy?
use a plugin, basically
Anything in particular the netty update helped with?
there was a few changes which looked mildly interesting for us
Nothing super OMG< WE GOTTA though
hello, why is this drop in tps due? thanks
https://timings.aikar.co/?id=4a88eedfaa184d14b237a2aef751692f
Aikar's Timings Viewer - View Timings v2 reports from Paper and Sponge
Wrong channel & outdated
ah sorry, what is the channel?
#paper-help if you fixed the latter
what should i fix? I do not understand
update to a supported version
1.16.3 is
ahhh ok, can't I get support if I'm not on that version?
Correct
what does this error mean. im on the latest build of paper and waterfall as of 10 minutes ago. I didnt change anything, just restarted my server and this happened
if i need to give more info lmk what you need
there is no error thrown on my paper server just in the waterfall console
you have a plugin creating duplicate scoreboards
I ran into that issue when using a scoreboard plugin alongside OldCombatFixer or whatever - the code they used to stop player collisions also used the scoreboard and I guess it would create duplicates?
Hmm my players can't connect on my Proxy since I update it : https://hastebin.com/itoxosofog.properties
That's a plugin thing
Some plugin cancelled the connect event, but, didn't give it a server to send you to
.318
.g papermc/waterfall 318
(DiscordBot) https://github.com/PaperMC/Waterfall/issues/318 -- Issue #318 · PaperMC/Waterfall: "Proxy does not handle duplicate scoreboard teams across connections #318. Open. superkarli9 opened this issue on Nov 30, 2018 · 18 comments. Open ..."
^ for the scoreboard one
Yes I found the issue thank you !
any way to disable DNS caching and make it resolve DNS everytime someone tries to join a server?
ratelimits or anything should be an issue, I run an internal DNS server inside of a VLAN
Unless netty has a system property for the cache, no
f
Is it possible to connect a 1.12.2 forge server to a waterfall proxy?
If im trying to join the server from my proxy i get disconnected instantly and just stay in the lobby. Nothing in the proxy and on the forge
Yes, just, if you want IP forwarding, you need sponge
did you enable forge support? not sure if that's on by default in waterfall
ah just instead of forge spongeforge?
sponge is a mod
ok makes it even more simpler
Didnt know of sponge really yet
Do i also need it as a plugin then?
yeah i thought there may be a plugin version of it
Did install sponge now
Still the same problem im justing getting disconnected instantly nothing on the logs of the server or proxy
Is it maybe a problem im trying to join from a 1.16 server with viabackwards onto a 1.12.2 forge server?
as I said, forge support is hit and miss due to mods
yeah ok, then i will make it a standalone server
without any form of log output, can't really say, could just be some mod blowing up
Well, the cleanliness of that basically says that the server intentionally closed the connection
Is there anything I need to watch out for when switching from BungeeCord to Waterfall regarding events? Currently running into odd issues like the ProxyPingEvent not getting fired
No, there is one event where we fire it in places that bungee doesnt' for the sake of exposing it/consistency, but, otherwise, no
Help in waterfall #379 mc version is 1.16.3
<Client Error message> The server you were previously on went down, you have been connected to a fallback server
<Waterfall log>
[18:42:57 INFO]: [MC] Registered player Hate_Game |
[18:42:57 INFO]: [MC] Refreshed UUID-Name lookup: d5654cbc-ffe1-4bca-922b-738be0b77765 |
[18:42:57 INFO]: [Hate_Game|/122.32.51.42:14214] <-> ServerConnector [towny] has connec|
ted |
[18:42:57 WARN]: [/122.32.51.42:14214|Hate_Game] <-> DownstreamBridge <-> [towny] - ove|
rflow in packet detected! Cannot receive string longer than Short.MAX_VALUE (got 121668|
characters) |
[18:42:57 INFO]: [Hate_Game] disconnected with: The server you were previously on went |
down, you have been connected to a fallback server |
[18:42:57 INFO]: [MC] Un-Registered player Hate_Game |
[18:42:57 INFO]: [/122.32.51.42:14214|Hate_Game] -> UpstreamBridge has disconnected |
[18:42:57 INFO]: [/122.32.51.42:14214|Hate_Game] <-> DownstreamBridge <-> [towny] has d|
isconnected
server keep kicking me in minecraft 1.16.3 no matter what the server version is
But when I connect my server with 1.15.2 version, I can flawlessly join my server
how to solve this?
Something sent a bad packet
[18:42:57 WARN]: [/122.32.51.42:14214|Hate_Game] <-> DownstreamBridge <-> [towny] - overflow in packet detected! Cannot receive string longer than Short.MAX_VALUE (got 121668 characters)
I think this is cause but how to solve?
You need to work out what's causing it, we can't see anything useful here bar a bad packet
replicate without plugins, etc
.g papermc/waterfall 318
(DiscordBot) https://github.com/PaperMC/Waterfall/issues/318 -- PaperMC/Waterfall: "Proxy does not handle duplicate scoreboard teams across connections #318. Open. superkarli9 opened this issue on Nov 30, 2018 · 18 comments. Open ..."
Can anyone help me with this?
I have it enabled on all server including the bungeecord server but it still shows this.
You need to connect to the proxy, not the actual servers
you did something wrong while configuring
gimmie a sec
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.
smp has an extra space before it
your forced hosts thing is also mangled, just, remove it, you don't need it anyways, force_default_server will do the "send to the lobby" thing
no
thanks for the help :)
sorry to be a pain, one last thing when i try to connect through the proxy i get this?
connection refused
Basically, either you misconfigured it, e.g. wrong port, or, firewall or router said nope
No, connection refused means that the connection was explictly refused by something
im just simplifying it
Yea but like how it's linked
So, you probably just want to make the proxy use that
how can i make it so getting kicked on one server will always send players to another server instead of just disconnecting them?
requires a plugin
I wrote this very simple one a while ago: https://github.com/Minebench/KickInfo
.g bungeekick
(DiscordBot) https://www.spigotmc.org/resources/bungeekick.1310/ -- BungeeKick | SpigotMC - High Performance Minecraft: "BungeeKick works kind of like MoveMeNow. When players are kicked from a proxied server, they are pushed to the lobby instead of quitting the ..."
You can try that one too ^
does the same as mine just without titles 👀
people want big flashy text
How is auth handled when using waterfall? Since premium is disabled, I guess I'll have to use my own auth plugin, however how is this handled, shall I ask for credentials on every server they join or should it (can it be?) handled by waterfall?
The proxy deals with auth and passes the UUID and IP to the backend servers, can't MITM auth, so it's dealt with by the proxy
You just need to ensure that only your proxy can connect to the servers, firewall or bungeeguard
i have a question, i have been trying to activate this proxy option in cloudflare so my ip is hidden, but when i do that i cant connect to my server, is there a way to do it and still be able to connect?
no
No you should not proxy your server IP through cloudflare, it doesn't work for minecraft servers
Use TCPShield instead
In theory it does work with Cloudflare Spectrum but that's very expensive
Thank you zzzCat
There are plugins for that type of stuff, scroll up people where talking about some plugins which do that kinda stuff, no idea if they cover your needs specifically, however
it's not
We can't tell what caused a player to be booted
it would be pretty dumb if /kick just sent them to the lobby
I mean, 2 second read of the description, yea, I guess?
What exactly is waterfall?
Waterfall is a proxy that links multiple standalone mc servers together
Kinda like bungee cord? Is it a fork of paper that runs exactly the same except with these features?
It's a fork of bungeecord with additional patches, fixes & enhancements
Technically speaking, is there anything Velocity cannot do that Bungee/Waterfall can? Or any benefits to using Bungee/Waterfall instead other than plugin availability?
Not really, they're proxies, only real difference is in APIs, bungee generally goes out of scope because "pretty"
e.g. not having the proxy fuck around with scoreboards would make fixing shit like #318 much easier
That's what I thought, I really appreciate it.
@cursive flicker Only difference really is in plugin availability. In general Velocity tends to be the superior technical solution but unfortunately too many people need to use legacy brain-damaged plugins.
how do i enable ore obfuscator on paper/spigot i know its built in but i dont know how.



