#waterfall-help
1 messages · Page 23 of 1
waterfall is a proxy that connects multiple paper servers
-> #paper-dev
Soz for spam, I just still have the problem
well first off: you somehow exposed your backend servers to the internet and let people connect directly to it (the 25505 port), you should really fix that
for your force host issue: do you have the option enabled to force people on the default server? not sure if that interacts with forced hosts but it might be worth a try
is there a way to forward the MOTD of paper server to bungeecord? for individual host?
if you set ping_passthrough to true then it will display the paper server's motd on the proxy
@acoustic adder Should I have force_default_server at true? cos i have it at false rn
hi i want to ask how can i modify this message ?
what would be a proper way to hold somebody in the loading screen for about 0.5sec before they connect to a server?
I don't think you can. That's client based.
cant the server force them to load extra stuff for a while?
Disabling entity metadata rewriting in waterfall.yml might be what you're looking for
that should make a server switch appear to the client as basically just being akin to joining a new server
😦
grab the messages file out of the proxy jar and put it in the folder where the proxy is ran
Actually
That one might even be in config.yml
Ah, no, bungee doesn't do that, gotta grab the messages file
so i just edit the messages file and thats all
would it be possible to hold somebody from joining for a specific amount of time? Like what if you pause the thread or something, will that pause everything else as well or just their connection?
That would stall the worker thread
Look into AsyncEvent and more specifically the class LoginEvent
cant i hold them up somehow like a police search?
If you're tryna deal with new players, use the async events like tux said
You can't stall the worker thread, it will cause every connection on that event loop thread to stall, which is a Very Bad Idea
i am also trying to hold up players switching between servers
New players is ez
Existing players is a pain due to state and the proxy was not designed to let you have players in a limbo state
ty for the help electroniccat
How do you switch between servers in waterfall /server isnt a thing
/server does exist on waterfall, check your logs, maybe something prevented it from downloading the modules
this is the whole log
@reef fulcrum
everything seems to be working
the modules are there
but the commands don't exist
on the lobby
Please don't ping
ok soz
Pretty sure that our modules also do not say that they're by waterfallmc
Oh, they do
yea they do
I must have been used to running bungee and being lazy instead of actually shoving them in manually :L
But, make sure that you're connecting via the proxy
Does the proxy show that you joined?
nope
yes
it does
[lobby] has connected?
what?
well it does say my name on the lrft
more
What does it actually say in game when you run /server ?
/bungee ?
yes
run it, what does it say
Honestly, no idea
oof
If the modules loaded, the commands should be there unless you've got some plugin installed which ripped them out
That's on the server, not the proxy
Beyond deleting the modules folder, I have no idea
i'ma delete it
okay
restarting
nope
same
eh
gotta go to the shop
i'll come back later for this
Maybe /server is disabled in bungee's config.yml
someone just entered the server and started stealing peoples names and making them appear offline
I assume this has something to do with offline mode and bungee
myes
you're supposed to configure it so that you can only connect through the proxy
last time i checked the guide on firewall at spigot was a bit outdated and did not block stuff properly
had to do some googling to learn how to firewall properly
should proxy_protocol be false or true
what do you mean by open
opposite of closed
Uh, so am using bungeecord, but randomly got this error which made all players lose connection: overflow in packet detected! Cannot receive string longer than Short.MAX_VALUE (got 34348 characters). Anyway to narrow this down or prevent it? Is it caused by a client's incoming packet?
I heard that there is a method where BungeeCord crashes if anyone sends invalid handshake packet.
Is it fixed by waterfall?
have you only heard about it, or also seen it?
@twilit sand do you have stacktrace around?
Project Waterfall build #298: SUCCESS in 3 min 21 sec: https://papermc.io/ci/job/Waterfall/298/
crickets
Project Waterfall build #299: SUCCESS in 2 min 59 sec: https://papermc.io/ci/job/Waterfall/299/
@tidal musk I only heard about it on SpigotMC forum.
Furthermore why there are 2 separate BungeeCord forks on PaperMC with the difference that the other fork is also supporting 1.7.10?
Wouldn’t it make more sense if you had one fork including 1.7.10?
why not?
not everybody needs 1.7.10
but they might need improvements what waterfall has
It wouldn’t make any difference if your server is 1.8.8 with ViaVersion.. so?
Using one fork from 1.7 to 1.14 is optimum. If you don’t want to use 1.7, so keep your server on 1.8.8 (or higher) and use ViaVersion.
Problem solved.
so then you don't need to use the other fork (Travertine), nor even worry about it

leave labeling projects useless to their developers instead
No, I want to use 1.7. But Waterfall is updated more than Travertine.
So my question is why not one fork with 1.7 to 1.14 supporting all of these versions?
Ppl who don’t want to use 1.7 could create 1.8 (or higher) and use ViaVersion. These 1.7 players wouldn’t be able to join. So what’s the point?
And from the last topic, I found this:
https://www.spigotmc.org/threads/new-bungeecord-crash-exploit.397041
https://www.spigotmc.org/threads/crash-problems-how-to-solve-it.405088/
Yes I know that I am using Waterfall instead of BungeeCord. But I think this issue is also in BungeeCord and will therefor affect BungeeCord users. If...
travertine being off few commits from waterfall does not mean it's really "updated less"
there haven't been any significant behaviour improvements in waterfall
"crash exploit"
that's simply DoS and the cause of very verbose logging
nothing else
Well.. the payload exploit (book) is basically a dos too but could be fixed by limiting it. And my question was whether waterfall has a protection against it or not.
waterfall does not have any special workarounds for this
ok thx for the answer
I've asked multiple times for means to replicate these issues, especially with my health as it is right now, I don't have too much energy to throw into stuff, so
Some issues are basically unfixable, but I'm working on means to allow a plugin to see when stuff is going down so it can react to it
Travertine is a few commits behind, I generally aim to deal with them together, just, I'd rather focus on the main project vs tryna throw what's left of my reserves at travertine
Lmao
God, I wish
Project Travertine build #111: SUCCESS in 1 min 44 sec: https://papermc.io/ci/job/Travertine/111/
Configuring build.sh to point to an invalid ${what_name} nuked my entire repo x'D thank god for local version history
F
electronicboy with means to replicate these issues you mean a link to a crasher and / or the source code?
for example, yes
Project Travertine build #112: SUCCESS in 1 min 35 sec: https://papermc.io/ci/job/Travertine/112/
Project Waterfall build #300: SUCCESS in 3 min 25 sec: https://papermc.io/ci/job/Waterfall/300/
Hello. What is this setting in the Waterfall config.yml for? prevent_proxy_connections: false
iirc it's the same as vanilla's prevent-proxy-connections in the server.properties
correct
Does it block any connection? Or only send the IP to Mojang?
sends ip to mojang and mojang decides
iirc it's for blocking certain players from certain countries playing minecraft
don't they also try to compare launcher auth IP with the IP the server sends or something like that
yeah might be that
Hello! Is there some tutorial to learn how to use Waterfall? Or to configure it?
basically, follow the bungeecord tutorial
Ok 🙂
Is there anyway someone an set aliases for waterfall commands (e.g setting a command /server C to /C
No, but there are plugins like slashserver for that specific usecase
(might be alias plugins too, you'd need to search)
awesome, slashserver will do nicely.
I dont have permission to do /alert, /send or /find yet i have the permission in groupmanager?
The proxy can't see permissions on the server, it's basically not magical
How do I do it then
use decent permissions plugin or maintain two sets of permissions
config.yml, or, ^
(using config.yml for perms is gross, really, you should be using a permission plugin which is actually maintained)
luckperms can manage perms on both bungee and your servers
groupmanager
muh legacy
inb4 pex
.g pexbridge
(DiscordBot) https://www.spigotmc.org/resources/bungeepexbridge.10839/ -- BungeePEXBridge | SpigotMC - High Performance Minecraft: "Automatically sync your PermissionsEx permissions to bungee!"
enjoy
lol
i still use pex actually
even though literally everyone and everything is screaming at me not to
muh legacy
fossil
i'm ruining the environment
not enough information
uhc
Is there a way to 'send' the player to the closest proxy server based on their geolocation.
For example
1 player in EU
1 player in US
Both players connect to "play.domain.com"
But I have servers in US and servers in EU, how can I send the player to the correct server based on their latency to the proxy? We don't really want the US on EU servers as the experience wouldn't be as good as directly on US.
I don't think that feature exists built-in, but it could definitely be done with a plugin, with relative ease I think
it might be built-in but since I don't actually use waterfall I'm just basing my take on things I've heard
Okay, thank you anyway. I'll see if anyone else has done it here. Would be interesting to check out. I'm assuming this is how massive networks such as hypixel are handing it?
I need some help
I can't connect to any of the servers
My server domain won't work either, the other server has offline mode set to false
where do I find the thing to change this: https://cdn.discordapp.com/attachments/530302630979829771/653034907597144084/unknown.png
The list is annoying to deal with I would prefer it Waterfall 1.8- 1.14.
especially seeing as it gets bigger each version added.
You should be able to change that with the game_version property in the waterfall.yml
Meh
☑ Terminator 2: Rektment Day
That's what I get for using dumcord on mobile I guess
i giveup
I actually was faster, just double checked if what I wrote was correct :P
I'm a Java dev, I looooove verbosity :P
Hello. Is there any way to set which server player will connect in LoginEvent? So I can set the server where will player connect? If no how can I do that?
You probably wanna use a reconnection handler instead of events
Those are specifically designed to control where a player goes when they jojn
How can I use it? Do you have any example?
.g waterfall javadocs reconnect handler
(DiscordBot) https://papermc.io/javadocs/waterfall/net/md_5/bungee/api/class-use/ReconnectHandler.html -- Uses of Interface net.md_5.bungee.api.ReconnectHandler...: "Packages that use ReconnectHandler. Package, Description ... Classes in net.md_5.bungee.api that implement ReconnectHandler. Modifier and Type,..."
Thanks. It called after LoginEvent? Or at the same time?
probably after, I don't recall
You'd need to go check the impl, too head spinning to care to look
How can I set the server? I think this: "void setServer(ProxiedPlayer player)" (Save the server of this player before they disconnect so it can be retrieved later.) is not for that I need.
so i have a minecraft server, using waterfall as a proxy and many other servers (mostly 1.14.4), i also have a 1.8.8 server what has viaversion on it.
when i connect to that 1.8.8 server through waterfall, its laggy (delays on teleport, joining delays, respawn screen takes 5 seconds etc, get it?)
but when i connect directly through the port to that 1.8.8 server, there is no delays and its completely fine
What causes this and can this somehow be prevented?
have you tried turning off entity metadata rewriting in waterfall.yml?
that did not fix the issue
something in waterfall makes it laggy but i don't know what
your next option is using a profiler and figuring out what is causing that
though i bet that viaversion could cause that... theoretically protocol hacks (shit has to be rewritten on-fly before sending to clients or server) add their own overhead
and with poor configurations (poorly set JVM flags etc.) issue could get even worse
through water it sends positions like million times a second
that didn't happen when joining directly from the port
again this only happens when connecting through water
else its fine
and how are you recording that info?
i got it from my friend who sent in a bot
It's pretty hard to say exactly what is going on without actually pulling out stuff to debug it
That doesn't say anything then
I have no idea what information that is logging out
Specifically, what packet? Is what I'm asking...
if that's the client bound position packet, you'll have to find out what is sending it
To my understanding, that won't be induced directly from the proxy itself, but likely a side-effect of something going on
Once you determine that, we can move further on the issue
Its almost like a client updates its position 20 times a second
I don't see anything abnormal there without timestamps to see those were all sent at once or something
bot is kinda dumb idea for debugging
as bots do not usually behave like vanilla client does
instead, turn on debug logging for minecraft client and you shall see what packets get sent / received
Its almost like a client updates its position 20 times a second
yeah thats the issue
its not the client updating it
its the server sending the player their position at that rate
which only happens when they get teleported
but right now it just constantly tells the player where they are
which is not a vanilla behaviour
playing lol
Waterfall doesn't have the ability to define movement or spawn packets as they are not defined in the Protocol class to be interpreted or reencoded by the bungee cord.
yeah, you'd have to add your own custom packets to do that
it's not too difficult to do tbh: http://git.io/JeSGY xD
half of that is older Bungee version support, lol
which is not a vanilla behaviour
did you try with plain 1.8.x?
e.g without via* plugins?
what happens when you have via installed and connect with 1.8.x again?
just to rule out that protocol translator ain't doing some weird shit
either way you should hop into viaversion discord for further help
if the proxy is causing the server to spam packets at it, there is an issue with something that's causing it to do that
I'm not going to waste my time tryna debug with 1.8.8 specifically, I don't have the health capaciTy to even lie about tryna get to that, but, basically; What you need to do is work out why the server thinks it should be tryna send that information
i tried to connect directly through the port to the server with 1.14.4 and 1.8.9 and then the performance was as we expect - lag free
but using 1.14.4 and waterfall, it was poor
So, 1.14 to a 1.14 server produces the same issue?
So, you're literally telling us nothing new there
so how am i gonna find out "why the server thinks it should be tryna send that information"
and btw via mods sent me here, because they don't have an idea whats the problem
easiest option I've got involves a debugger and potentially modifying the server to produce some info
Expectation Connecting:QuietException : Server is online mode!"
I checked every server.propertie and they are all offline
Any idea what the cause might be
A server is in online mode. lol
I've never seen a case where that shows without a server being in online mode.
I didnt modify anything, experienced that after I updated to the latest build
Do you get the message in your server consoles stating the server is offline and insecure with big yellow text
At startup
Yes
[07:50:32 WARN]**** SERVER IS RUNNING IN OFFLINE/INSECURE MODE!
[07:50:32 WARN]The server will make no attempt to authenticate usernames. Beware.
[07:50:32 WARN]Whilst this makes it possible to use BungeeCord, unless access to your server is properly restricted, it also opens up the ability for hackers to connect with any username they choose.
[07:50:32 WARN]Please see http://www.spigotmc.org/wiki/firewall-guide/ for further information.
[07:50:32 WARN]To change this, set "online-mode" to "true" in the server.properties file.
The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy.
I don't think it could be programmatically turned back on. Did you also set bungeecord: true in the spigot.yml's and ip-forwarding: true in the bungeecord config?
Yes, everything is properly set
And after upgrade to the newest build everything suddenly just broke
I haven't heard of a bug in the latest builds but I personally don't run them either. I have custom forks.
What is the server throwing the error, you should probably toss the server.properties here
but I believe that you did turn online-mode to false
The server did not throw any error
Since the user didn't even connect to the server
Nah, the server the bungeecord was attempting to route you too on connect.
first lobby, or /server command
what was the target server, look at its server.properties
How the hell you generate a config a day in advance. lol
Unless you're an aussie or something
Ah. but that config looks correct. I see no abnormalities. So my second guess would be that the routing is incorrect.
It clearly says CST xD
eg, its not actually sending the request to the server you think it is.
Like on a shared host
How would it be incorrect if all I did is updating the Waterfall.jar
It's just a hunch, that config is correct.
Sounds unrealistic to me xD
It is unrealistic, but unless something is setting online-mode back to true after server launch there shouldn't be an issue with that spawn server.
Nope, I checked the server after it starts
Offline all the time
here is the paste of the bungee config
9/10 times, either a server hasn't been setup properly, or people have the proxy connecting to itself
and are you using imboxmc.com or ab.imboxmc.com, if so you need to look at your survival server.
Which is exactly what you've done from the looks of it
Your survival server is tryna connect to the port that you have the proxy itself on
survival:
motd: |-
§e=§6-§e=§d?§9——§3——§b——?§6§lImbox§2§l MC§b?——§3——§9——§d?§e=§6-§e=§r
║§d§l社区§f?§b§l国际§f?§c§l正版§f?§6§l温馨§f?§e§l有爱§f?§a§l稳定§7║
restricted: false
address: localhost:26042
He's right
oh shoot
bind_local_address: true
host: 0.0.0.0:26042
max_players: 25
In the end you were connecting to something in online mode 😛
talkin about packets then its this one https://wiki.vg/Protocol#Player_Position_And_Look_.28clientbound.29
State: PLAY, Packet ID: 0x35
no idea why it should be sent more often than without waterfall/bungee
proxy doesn't even intercept said packet
(unless you have some odd plugin which does)
a plugin on the proxy itself?
idk wbu but i wrapped MCProtocolLib
and registered all missing packets on bungeecord like that
that's top tier
but cross-server server selector was defo worth it
cross-server server selector 😮
- disabling all plugins on the proxy does not solve the issue, it still is laggy
As already stated, it's unlikely to be proxy itself sending those packets
Literally impossible unless added by a plugin.
I've literally already said that you need to work out why the server is spamming position packets if that is the issue
Go grab a debugger out or something
such dumb as i am, would there be a guide how to do that?
that's an issue I haven't experienced before and I haven't used such debuggers too
and I don't know if its an issue only for me or if its a wide-spread issue
Can be used on the server and the client to monitor packets.
It's not widespread of it would be knocking out my networks right now.
so i just need to put that to the startup script or what
that's the first step, yes
So now we need to wait for Bungee/Waterfall for 1.15 :I
i dont think we have to, last time it was just a number change
But still someone need to do it
ah rip
nobody stops you from doing it yourself ;)
Skils in java stops
sad
paper 1.15 when
Yeah I just opened the bungee github
so there's no bungee yet either
maybe donating will speed up the process
😄
or, you know, creating a PR ;)
PR on bungeecord?
Yeah let's just create another PR that will go in the sea of all the PRs md_5 did not care about
I mean PRing to non-open source is kinda iffy but still, why not ¯_(ツ)_/¯
md has explictly stated that he doesn't want PRs for protocol updates
I'd rather donate to the project
money buys coffee
coffee makes bungee 1.15
see?
When someone on my network joins my survival server and then leaves it from there i get a message in my hub server that he left the server. But he left it in survival and that's anoying to get. Is there any way to fix that?
remove the plugin doing that?
kek
maybe use a spigot plugin for that stuff
Project Waterfall build #301: SUCCESS in 3 min 7 sec: https://papermc.io/ci/job/Waterfall/301/
Hey guys does the latest waterfall commit include Bungee #1426?
How can I ask if he is a cracked player, so or prologinevent?
Project Waterfall build #302: FAILURE in 45 sec: https://papermc.io/ci/job/Waterfall/302/
oof
Whoops
cat, log4j module has still 1.14-SNAPSHOT as a version
Did I forget a rebuild... o.O
log4j module is not added in POM patch
Yea, I already fixed that locally
👌
I apparently forgot to do another rb before I pushed \o/
Yippee, build fixed!
Project Waterfall build #303: FIXED in 2 min 59 sec: https://papermc.io/ci/job/Waterfall/303/
💸
Project Travertine build #113: SUCCESS in 1 min 19 sec: https://papermc.io/ci/job/Travertine/113/
thank you cat ♥️
@reef fulcrum does 303 now contain the upstream bungee #1426?
sorry i read on jenkins and it shows bbuild failed, apologies
ok i read the detail. its clear now 😛
yay cat
does the latest Travertine have 1.15 support?
yes it does nvm
i cant download it
huh
{"error":{"code":404,"message":"Could not locate build"}}
yeah already got it no worries :)
sorry but what is traverine
an extended Waterfall with 1.7 support
People asking for travertine are generally the people who have a huge broner for 1.15
also because it wasn't available yet for waterfall ;D
1mm pener won't create big eruption so 
megameter
millimeter
manlymeter
Guys i wish to make waterfall ready for ipv6 as i have it on my servers internet how would i configure ipv6 into the config.yml like can i see an example?
i want both ipv4 and ipv6 on the server
address: "[::1]:25500"
wont that only work for ipv6 though?
oh no the problem is i want waterfall to listen on ipv4 and ipv6
don't know from top of the head but
if you can set the bind address
try '[::]`
if not, use HAProxy or nginx
both can do ipv6 fine
hmm would cloudflare work?
yeah its a ddos protection service for free based on nginx
not your-average-tcp-proxy
yes but i have had them working on another network for years
web != random tcp stream, so cloudflare won't help you here
unless you wanna go hypickle tier and pay for CF Spectrum
.g cf spectrum
(DiscordBot) https://cysticfibrosisnewstoday.com/2019/01/21/diverse-spectrum-health-across-cf-community/ -- The Diverse Spectrum of Health Across the CF Community: "Jan 21, 2019 ... Columnist Tré LaRosa writes about the vastly different experiences CF patients have, both in the diverse physical and emotional..."
(DiscordBot) https://www.cloudflare.com/products/cloudflare-spectrum/ -- Cloudflare Spectrum: "Increase TCP and UDP security and prevent DDoS attacks for gaming, mail, SSH, and other services with Cloudflare Spectrum."
either way, read about nginx's stream support & don't forget to enable proxy protocol support in both
.g nginx proxy protocol
(DiscordBot) https://docs.nginx.com/nginx/admin-guide/load-balancer/using-proxy-protocol/ -- Accepting the PROXY Protocol - NGINX Docs: "This article explains how to configure NGINX and NGINX Plus to accept the PROXY protocol, rewrite the IP address of a load balancer or proxy to the..."
.g nginx tcp stream module
(DiscordBot) http://nginx.org/en/docs/stream/ngx_stream_core_module.html -- Module ngx_stream_core_module: "stream tcp_nodelay variables_hash_bucket_size variables_hash_max_size. Embedded Variables. The ngx_stream_core_module module is available since..."
and here's the guiding light... whatever
yeah but i would rather not run nginx if at all possible
why so?
while you're at it, could also set up load balancing/redundancy, unless you have <10 players
:D
i mean ok, most of the people don't need that indeed...
well i have about 25 right now that my servers can handle
though ipv6 support in minecraft is quite shit anyway
iirc AAAA records aren't tried to be resolved at all
something something preferIPv4Stack=true
yes i know but i want it ready for if it is needed or changed to
(DiscordBot) https://discourse.codeemo.com/t/solved-cant-reach-server-with-ipv6/2223 -- [SOLVED] Can't reach server with IPv6 - Connectivity -...: "Dec 19, 2017 ... Hey guys, i installed MineOS on a little server. I can reach the WebUI via IPv4 and the minecraft server also via IPv4, both in..."
much useful
dig in mojira rather
world is not ready for ipv6 either
we'll first implement widespread ipv4 CGNAT than actually move on 
(which i can see already happening in some places)
you'll grow old before ipv6 gets usable everywhere, so tbh don't bother
eh not so much Europe ran out of allowed ipv4 addresses they can give out
(DiscordBot) https://en.wikipedia.org/wiki/Carrier-grade_NAT -- Carrier-grade NAT - Wikipedia: "Carrier-grade NAT (CGN), also known as large-scale NAT (LSN), is an approach to IPv4 ... the Impact of NAT444 on Network Applications". ^ Jump up to:..."
yes i know
thats a whole issue of itself
that does not make ipv6 to be implemented faster :p
still explain how it would work if both houses set up a service on port 80 completely broken
uh, they won't
ISP just won't forward inbound connections
another can of worms gone
dumbusers don't host any services 99.99% time either
thats an issue for home servers thus pushing everyone to company plans
that's the reality
and company plans cost 2x more then regular for no benefit it wont happen i can assure you that
"It has significant security, scalability, and reliability problems, by virtue of being stateful."
"It does not solve the IPv4 address exhaustion problem when a public IP address is needed, such as in web hosting."
like i said before...
any logical mainstream isp wont deploy it the back lash would be to much
yes i work for AT&T we already thought of it and vowed not to do it since it have flaws for mobile networks its fine since main servers have a cable or fiber wire coming in not 3g or 4g
and i have seen few broadband connections around with CGNAT already
they must not be as big as others
no but the isp and the government will
also inform me of how you might ban a certain house without banning the whole block?
explain the scenario
one idiotic kid manages to DoS my server and i'll ip ban him
correct?
and then his ip is shared by 3 other players
then bad luck, others get banned as well
yeah in the nat it would ban all the houses imagen if google did that???
bad luck
no but the isp and the government will
govt? they usually roll their own internet access, no?
dude i have to go i can sit here and try to show you how your thinking is wrong there is no "Bad luck"
i'm not thinking unironically that such solutions are the correct ways of solving the issue
i'm just speculating that world is kinda moving on to those shitty temporary solutions based on what i've seen throughout the years :p
dude shut it omg
indeed, isp-s should start pushing out ipv6 to everybody; ditch ipv4 and whatever the fuck. but how long has ipv6 been around already and nobody has really cared? :D
why should it get better soon?
it won't
it will slowly as they realize theyre out of IPs
which is still going to take a significant amount of time
like 30-40 years i bet
eh idk about that
larger ones can just buy them off of other companies
but thats going to get more and more expensive
they also already have if you look att pushed out country wide ipv6 duh
and who knows what the pace will be for that
depends on many factors in the end yea
i would say sooner
universities can start auctioning off blocks for extra $$$ too
a lot of old unis have massive blocks
i would give it 5-10 years max since most isp's are already moving to dule stacks
and then there's the DoD blocks, I dont imagine theyll ever give those up but there's a lot
USA
this year?
yes
fuck, Telia pushed ipv6 access out to consumers in 2013 i think
Learn what IPv6 is, how it differs from IPv4, which AT&T services support IPv6, and how to check if your Internet Services are IPv6 compatible.
in estonia atleast
yeah att should have sooner but the techs were all being trained like i was trained on ipv6 for there backend
alright, i'll lower my guess when ipv6 gets more widespread
i'll give it 10 years now lol
github more then likely already has ipv6 they just haven't given out the ip
good thank you lol
going back to the original topic
in listeners section
there's host
what happens if you'll set it to "[::]:25565" or something?
ill have to see in the morning as i have to go to work ill let you know in the am
👌
Is waterfall compatible with forge at all?
Yes
Cool
Can someone explain this?? Why does it claim its an ipv6 tcp protocol when its an ipv4 address it is listening on this is a responce from a lsof -i -P -n | grep LISTEN command
java 20198 john 38u IPv6 1820872 0t0 TCP 192.168.1.2:25571 (LISTEN)
as a matter of fact all the java applications that are listening have that as well?? says its ipv6 with an ipv4 address
Hello ! Latest build of Waterfall support 1.15 protocol ?
yes
latest = 303?
ok thank you, maybe you have to write somewhere that added 1.15 support
https://i.imgur.com/xaAE0ty.png what like, in commit messages?
I mean it's not in the main part of the message so easy to miss on the download page I guess
it's always worked like this tho
If you hover over the title it does show the full text
if someone got really brave they could make a dropdown or something
but web dev is spooky to all us backend people
Oh no, dropdowns
@heady wave gamerule reducedDebugInfo i think but not sure
iirc enabling reduced debug info also hides it, no idea if it still gets send though
afaik reduceddebuginfo doesn't stop sending anything
just sends a field in the login packet to ask the client "please hide info"
and the client does what it wants with that
ah yes, pebkac - said error was caused by outdated protocollib 😄
and i forgot to disable it
how would I get my server to display this instead of "Incompatible FML modded server"
https://i.hhservers.co.uk/uploads/5df23e58e6e692.34683905.png
hovering over that says "Bukkit modded server"
@lone void i use this plugin, and it worked so far 😛 https://www.spigotmc.org/resources/bungee-forgeicon.17695/
You're a life saver! Thank you so much 🙂 @uncut tinsel
np 😛
rip
back again to the issue I had (1.8.8 lag when connecting through proxy), debugger says also that its spamming packets:
[17:06:04] [Netty Epoll Server IO #1/DEBUG]: OUT: [PLAY:8] net.minecraft.server.v1_8_R3.PacketPlayOutPosition
1.8
maybe downgrading even further will solve the issue
Once again, you're literally telling us nothing new
The Q is "why is it sending those packets"
have you tried 1.8b yet
try the hackish method lol
just new Throwable().printStackTrace(); when said packet is constructed
aaand you'll very likely find out the origin
Hello,
Impossible to download latest waterfall
what's the url you're using?
https://papermc.io/api/v1/waterfall/1.15/303/download
changing 1.14 to 1.15 seems to fix it
Nice thank you
daily reminder that frontend dev succc
yes
Nice job, thank you 😄
Hi, I've a problem with my waterfall. It only shows the player count of the hub server and its motd, not the motd set for it in waterfall config but the actual server's motd
I have not altered the config in any way anytime recently prior to this incident
disable ping pass through in the config.yml
Will this not cause some new issues?
That option basically passes the ping through to the server, disabling that will make the proxy deal with it
Hello there! My proxy crashed. I don't know why. What could it be? https://pastebin.com/sYg9R2b6
(It's Travertine, dunno if I can ask support here, since Travertine is only a fork of Waterfall)
Great! I saw on somewhere a WARN from KeepAlive. I can't paste it all since the file is more than 400mbs.
I saw on AAC that KeepAlive check should be disabled if servers are on proxy, but that check is removed on latest versions, so I think there isn't a problem 🤔
trying to figure out what version you're on
youre not on the one they posted in here right away are you?
Looks like somebody is intentionally sending you naughty packets
guess who's getting coal in their stocking
It could be that @robust meadow tbh. Some times my server goes lag soo much, it's really strange because now is the first time that the proxy crashed
But sometimes the server gets very laggy
Some of the ending server log: https://pastebin.com/1J2HTeNd - There was so many lag reported by users and they everyone got kicked by KeepAlive timeout
Looks like the proxy was just overworked
There are some recent attacks which are somewhat a pain in the ass to resolve, boils down to rewriting how the proxy deals with certain bad conditions
Oh, I understand
Well, I will try to help if there's some new strange errors, this happens eventually on servers, not on proxy 🤔
Also, thanks for response
@vital widget your proxy look under attack ?
Can you send more logs before crash ?
Isn't it on attack
Sure, wait me 10 minutes
I just can send you the file, can't edit because it lags so much my VSC
[19:04:32 WARN]: No client connected for pending server!
[19:04:32 ERROR]: [Player|/IP:52172] <-> ServerConnector [survival_1] - encountered exception: net.md_5.bungee.util.QuietException: Unexpected packet received during server login process!
0d0300
[19:04:32 INFO]: [Player|/IP:52172] <-> ServerConnector [survival_1] has disconnected
Sometimes I get this error. What is the problem?
read timeout is networking stuff
Is there a problem you can guess?
"your network sucks" is the only guess anybody could give from that
if it's only a singular person or so, it's probably their connection
All read timeout means is that a packet wasn't sent in so long
aha
Is waterfall recommended over bungeecord and for what reason? I assume I can use bungeecord with paper servers?
waterfall is to bungee what paper is to spigot
jw, but if your backend paper server has a player limit of 20, and a player connects via bungee wherein the player limit is say..100, does it just throw an error message?
assuming lets say you have 20 people already on the server and you try and connect
proxy just passes through to the server
if the server disconnects for full, that's what the client will get
okay thanks, just confirming
Have there been any thoughts of adding a password / secret authentication system to Waterfall like Velocity's modern player info forwarding?
Kinda, yes
Which is directly integrated into Paper 1.13+
It would be neat to see this ported to Waterfall
The idea has been rejected in 2015 for BungeeCord
I did consider adding velocity forwarding support, but; Wasn't too happy with the idea of having to modify bungees main config file
There is also bungeeguard too
But, especially if it's hitting sponge, it's worth having it native, just; On the long list of todos which I'll get to one day when I'm not dead
the idea was that I wanted it per-server
if they let you configure what forwarding system to use per server, yea
right
yeah well, would be nice to see it in waterfall & paper sooner or later
especially since it would solve issues for people that are incapable of setting up a proper firewall
per-server forwarding support is neat
no more tradeoffs with pre-1.13 servers and post
I'm having an issue with connecting to a new bungee we just rolled out on a new machine. For some reason the "encryption" portion of logging in takes upwards of 10 seconds sometimes... Anyone have any experience with this?
does the latest waterfall patch have 1.15 compability? Or am I dumb now? :3
what is waterfall like, is it like paper?
Nah, it lets you have multiple servers on one IP.
Hey guys so I am not sure if you can help me but I am makinga bungeecord plugin and for some reason I'm unable to use the BungeeCord Plugin Messaging Channel. Nothing happens when I try to send data and retrive it on the other end. Someone said there is a different naming convention in 1.13+ (pluginname:channelname) is that true?
That's true, yes. I don't think the first thing necessarily needs to be the plugin name, but that's recommended.
Works now, thx! 😉
Hey! I was wondering if there's a n updated set of JVM arguments for running Waterfall in an optimized way. From a long while back, I have the following kind of arguments and was wondering if they still apply or if I should tweak them:
java -Xms2G -Xmx2G -XX:+UseG1GC -XX:+UnlockExperimentalVMOptions -XX:MaxGCPauseMillis=100 -XX:+DisableExplicitGC -XX:TargetSurvivorRatio=90 -XX:G1NewSizePercent=30 -XX:G1MaxNewSizePercent=60 -XX:G1MixedGCLiveThresholdPercent=35 -XX:InitiatingHeapOccupancyPercent=30 -XX:+AlwaysPreTouch -XX:+ParallelRefProcEnabled -Dusing.aikars.flags=mcflags.emc.gs -jar Waterfall-303.jar
you don't need that much ram for the proxy
Most memory is off-heap, so unless you're running a few hundred players, you can easily 1/2 the ram allocation there
256 or 512mb should be perfect
I know, but I'm having a bit of a memory leak problem with Denizen where sometimes it screw up if I allocate less than 2G
denizen is a plugin meant to run on backend server though?
denizen has nothing to do with the proxy
It does if I have DepenizenBungee running :P
(inb4 flooding plugin messages)
.g depenizenbungee
(DiscordBot) https://github.com/DenizenScript/DepenizenBungee/issues -- Issues · DenizenScript/DepenizenBungee · GitHub: "A plugin for Bungee proxy servers to relay messages sent through the Bukkit plugin, Depenizen! - DenizenScript/DepenizenBungee."
Only real concern for memory and plugins is plugin messages
I've seen them use the wrong channel basically, and the proxy just queues them up
Either way, I just wanted to know if the rest of the flags look okay - I got them like a few years back from the Waterfall IRC and I'm not sure if they're still applicable
They're applicable
The allocation pattern that mc has is pretty similar to that of the proxies, just; On a smaller scale; G1 also helps with the whole "keeping it running for daaaays"
OK, thank you. As long as those keep the proxy running relatively smoothly
I mostly had an irrational concern that maybe something changed that would affect long-term performance or something
(There are some tweaks in those flags that are different from the traditional Aikar flags, mostly NewSizePercent?30 and MaxNewSizePercent=60 and the addition of -XX:InitiatingHeapOccupancyPercent=30 but I don't think that should affect things terribly much, hopefully)
new size flags basically lets you balance how memory is being allocated on the heap
larger new size = smaller old gen
So basically I can just leave them like that and forget about them for the near future, then?
Basically, I don't suggest playing with the flags too much unless you have a reason to
Thank you. I don't plan to, I just wanted to get some reassurance that I'm not screwing the server up from the start :P
hello i have a problem with waterfall. https://paste.playspark.de/tITL19PSls
does not appear to be waterfall's problem
something related to cloudnet, try to contact with them instead
Project Waterfall build #304: SUCCESS in 3 min 9 sec: https://papermc.io/ci/job/Waterfall/304/
Project Waterfall build #305: SUCCESS in 2 min 57 sec: https://papermc.io/ci/job/Waterfall/305/
(Wanted to test if disabling the 2nd run of tests would offer any gains.... 12 seconds!)
Hey is Waterfall a drop in and replace solution to BungeeCord or do I need to do something to make it work?
drop in
Means you don't have Maven installed on your PC.
Then it's probably not defined in your Environmental Variables.
more like Environmental Warmings
I'm looking for a way to have a seperate public server ( no whitelist ) connected to the rest of the servers, so people connecting from the whitelisted environment can connect to the public server but not the other way around
but that means that everyone that is whitelisted needs to be defined in waterfall ?
i mean, LP at least is context-aware if you install it on the proxy
so you can just give permissions to anyone who is on a whitelisted server, or just negate it from those on the public server
ah 🙂
i only use perms on a roleplay server, other then that i can't be bothered 🙂
if people can't behave then i just remove m from the whitelist
but i'll have a look at a permission plugin 🙂 thanks for the 2 cents 🙂
p e r m i s s i o n s e x
^^^^
permission sex
👀
is there any public trusted travertine fork that cancels the book exploit of jessica?
Does the Latest Waterfall support 1.15 yet?
yes
Hey guys, My logs is constantly filled with
[20:48:23 INFO]: [/54.85.222.81:57572] <-> InitialHandler is pinging
same IP but different ports. Any idea what it is? I tried an nslookup and got
Name: ec2-54-85-222-81.compute-1.amazonaws.com Address: 54.85.222.81
This is on my bc only.
turn off ping logging
someone's simply pinging you, or attempting to spam with ping packets
no, waterfall doesn't support rcon
And there is not something for rcon?
Not that I'm aware of
shouldn't be hard to implement rcon support using a plugin lol
@tidal musk why not implement it into Waterfall ?
if somebody wants to submit a PR, I don't exactly have anything against the idea
If someone can implement it or create a waterfall plugin for that, I really need it
@little monolith https://www.spigotmc.org/resources/bungeecord-rcon.5817/
It don't work
I already tried
the plugin are not working on 1.15 Waterfall
it does a NoClassDefFoundError into logs
so
Strange because it works on my server (1.15 waterfall)
inb4 didn't read the description...
does any one have a solution for cross server ./warps (1.14)? I'm new to the bungee/waterfall world and not finding many existing resources for this!
NoClassDefFoundError
mfw
also 1-star review with the same error
reading is always overrated
guys i need help
.ask
If you have a question, please just ask it. Don't look for staff or topic experts. Don't ask to ask or ask if people are awake or available. Just ask the question to the channel straight out, and wait patiently for a reply.
post forwarding is not working and is at the same time its tough to explain but here goes port 25572 is my waterfall server i cant connect to it but any other port can the server is listening so im completely lost another port on the same server is also running on 25571 and it can connect just fine just not waterfall or bungee i have tried another port and it wont work on that either? could it need root perms. i should also mention from a local ip and 25572 i can connect
then why cant i connect???
kinda srv but yeah
enable NAT loopback
its enabled
if i use the external ip it wont connect but if i use local ip it will i have tried 2 different routers with the same config still nothing
have you asked someone else to test the connection for you?
e.g neighbour/friend or something
no wanna try
do you want to try? idk i dont have friends lol
no, i don't want to try
perhaps try online pinging tool then
.g dinnerbone server status
(DiscordBot) https://dinnerbone.com/minecraft/tools/status/ -- Server Status Query - Minecraft - Dinnerblog: "Minecraft Server Status Checker. Want to test if your server is running, or stalk somebody elses? Enter the address below and we'll find out..."
um thats not loading
then use some different site with similar purpose
you'll find that yourself very likely
then your port forwarding must be broken
on 2 sperate routers very very unlikely
no idea honestly
i mean, the implication is that your setup is broken, not the router
port forwarding is a thing what you should get right on your own, we don't really know how it works on various hardware
also worth checking whether your ISP is blocking incoming connections on their end as well
its working for every port except any port running waterfall though
some ISPs do that by default (for example mine)
not mine i have looked into that
wait, so you're saying that if you bind waterfall on 25573 then it's accessible externally?
it is not accessible externally
then you're still stuck at poorly done port forwarding i guess
or computer's firewall perhaps?
yeah right just like how the earth is flat i have been doing this for 5 years i know how it works okay don't say its poorly done because its not just like my cable managing skills being on point
don't know honestly
you can check whether waterfall is actually binding on all interfaces (or your ethernet/wifi/whatever interface)
make sure it binds on ipv4 as well then
java 32766 john 213u IPv6 273808 0t0 TCP 192.168.1.2:25572 (LISTEN)
it never has and im not sure how to make it bind on ipv6 it has that on all my java applications runnning
honestly i'm out of ideas now
kk
according to this github issue for another issue it is a java issue ill have to try setting -Djava.net.preferIPv4Stack=true
https://github.com/elastic/elasticsearch/issues/16994
that doesn't work either
Port forwarding has literally nothing to do with the service in play, port forwarding on general is fun due to routers being 10/10, had a Netgear router that every once in a while needed a settings wipe to work...
Hey guys, yesterday i've got an bots attack whick it was 300 packets for each second - I have the AntiBotDeluxe plugin modify by my friend, so from the 10000 bots, only 5 got in the global server - but my bungeecord OS stuck, crashed and i can't do anything for about 3 minutes, what should i do to block VPN attacks or packet sending from users, that i don't want to be on my server?
I've got poor OS for bungeecord, tbh it's about 1,7GB RAM and 1 CPU, should i try to upgrade them? I mean i use about 1GB to bungeecord
I can send you a log file but it's about 20 MB of text
I've got only Proxy, i mean if you click on the server, the bungeecord verifity you, if you on whitelist (because you logged before verifited well) bungge proxy you to global server
As i can see in a loggs of my OS - the proccesor while botts attacking was 100%, so it's problably why crash
Anyway i've got a small website on my OS, so 1,7GB i think it's enough for Bungeecord proxy and a webiste
Is there anyway i can make it so if one of my server closes or restarts all players get put into another server / lobby instead of being kicked?
does the latest Waterfall and Travertine support 1.15?
yes
@daring moss theres a plugin called slingshot it works pretty good for that
Its on spigot
Hello
I would like to find documentation to be able to properly configure the file waterfall.yml, mainly the line game_version:''
I assume it's a setting for which version of the customers will be able to connect it to the server as well as the information that will be redistributed to external APIs such as voting sites.
Or a link to a tuto that explains all the intricacies of the different settings available as well as their function.
My guess is that you can just write anything in there though I could be wrong
yea you can put whatever and that's what shows up in the client's server list
Project Travertine build #114: SUCCESS in 1 min 17 sec: https://papermc.io/ci/job/Travertine/114/
you realize the backend needs to support it too right
like, the proxy will let the connection through, but the backend server doesn't recognize the version and dcs the player
@daring moss There's a plugin called Slingshot that can do that for you
ElFly already told me that but thank you
recommending non-open-source plugins
Does Travertine not support bungee commands anymore? It doesnt recognize my /server command.
Well I know its a fork of bungee, but on the newest version, I am no longer able to use any bungee commands.
I cannot do /server [server name]... It says that it cannot recognize the command
proxy log? server log?
Has some PI on there. Let me know when you are ready for me to send it.
What is this https://prnt.sc/qc8dh7 is that fixed?
I made sure I removed my modules before updating. Let me see
Hmmm odd. I guess I needed the remove another file as well.
1.15.1 protocol support on travertine but not in waterfall ?
That's what the changelogs say
cool
Rip, still having the potion effect issue with Travertine on only one of my 1.7.10 modded servers. Not sure what mod it is because there is no good error message
Travertine has no chance of seeing what mod it is
I don't really expect it to either, but I don't get detailed errors on either client, backend, or bungee.
(remove this msg if needed) I know that Hexacord had this problem before and then some how fixed it.
if hexacord no longer has this problem but did, they dropped a patch which fixes an issue with forge
I can send the version of the 1.14.4 version that is currently working for me if you want and see what they may have done. Thats if you guys are wanting to work on 1.7.10 mod/forge support.
ez thing to do is go check the entity effect packet class
i have been considering making it tryna just store the bytes after the effect, should in theory work assuming that the mods weren't too stupid with the protocol
Not entirely sure where that would remotely be on either projects.
They don't deal with that packet
If you don't mind modifying your own jars, you should be able to easily yank that patch out
I don't know how to do jack when it comes to this. I wish I knew java and such. Ill see if I can get someone to do it.
Project Travertine build #115: SUCCESS in 1 min 43 sec: https://papermc.io/ci/job/Travertine/115/
Hi sorry but can someone help me, i am just being atacked, the console is spammed with this https://pastebin.com/N8X5mMHs
lates.log of waterfall is 500mb heavy, there is a lot, and players can't connect
Project Waterfall build #306: SUCCESS in 3 min 39 sec: https://papermc.io/ci/job/Waterfall/306/
does someone know that it could be?
Configure your firewall, basically
The proxy needs some work to try to cut down on the damage those people can do
Just, not really top priority with my health
Sorry for asking i know you can be very busy but what kind of firewall configuration should i need, at least give me some reference word so i can see on google
And thanks
You basically either want connection limits or start blacklisting bad networks
hem is a bungeecord/waterfall atack with some packets or something like that, thousands of them
Millions* 120 Million packets i count the last 5 minutes in the logs
i know, i thought fail2ban was to prevent brute force attacks for ssh
or other login pass stuff
fail2ban is a pretty generic tool
You can hook it up to anything with the right output/config
Ok thanks i will do some research
Is there a way to manage this exception with a plugin? https://pastebin.com/N8X5mMHs, i can program something to block those ip's but i don't know how to get it
That's part of what I've been tryna expose
Already added an event for booting connections super early in the connection process
Might be able to throw an exception handler into netty for it or something
Not too skilled with netty, but pretty sure it would pass up the pipeline
Ok i will do some testing, I'm sorry for asking several things, this doesn't happen to me in years and I had +1000 connected players that cause uproar in the forums just when I was about to go to my brother's birthday
Only a last question, is this the event? "ConnectionInitEvent"
if that was the newly added one, yea
I erm... had a lombok stroke
:^)
thanks for the commit
lol
too complex changes for git to understand
Project Waterfall build #307: SUCCESS in 3 min 5 sec: https://papermc.io/ci/job/Waterfall/307/
inb4 something is simply two or three lines off
Project Travertine build #116: SUCCESS in 1 min 25 sec: https://papermc.io/ci/job/Travertine/116/
wait TFO patch was pretty invasive compared to those oneliners what i do
nothing to wonder then
time to test
do you like, clear that packet count regularly?
yes every 10 seconds
if the cound reach 500 the "under atack" set to true
count*
the "attempToBan" adds the ip to an ipsed readed by iptables from linux
ipset*
the patch worked well but still the erros ar shown in the console, but thanks for all
hi can I have help configuring the water.yml file?
i tried to find documentation on these settings but i cant find it
ping me if u can help/link me to some documentation
Project Travertine build #117: SUCCESS in 1 min 25 sec: https://papermc.io/ci/job/Travertine/117/
Project Waterfall build #308: SUCCESS in 3 min 8 sec: https://papermc.io/ci/job/Waterfall/308/
hey sorry for the noob question I have a paper plugin and i want to send a player from server A to server B, how do i do this?
.g bungee plugin messaging channel
(DiscordBot) https://www.spigotmc.org/wiki/bukkit-bungee-plugin-messaging-channel/ -- Bukkit & Bungee Plugin Messaging Channel | SpigotMC - High...: "Dec 21, 2018 ... This page explains the Bukkit and BungeeCord plugin messaging channel. The channel allows plugins to communicate with the..."
The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy.
ok i think i got this
thx
so i dont really need to import any dependency right, just send the packet thing to "bungeecord"
yup
should i be using this for sending data between servers aswell? bc it requires at least one online player
Basically, if you need to be able to send messages between servers and can't garauntee that there will be a player on them, you either gotta rethink your design so that the plugin on the proxy is dealing with the logic so that that is no longer a requirement, or; use a messaging service or something similar
