use plugins
#waterfall-help
midnight latch
bungeecord.command.server permission
river panther
how can I either get the name of a players server on log out or get the players last server on log in?
marble kelpBOT
well waterfall isn't a normal bungeecord either, SeptyIsHere ;)
icanflyit: bungee stores that already if you disabled force_default_server
river panther
yes, I'm making a plugin that I want to intercept the log in event and store that server
but then send them to a different server first
@tidal musk do you have a permission manager installed?
Like luck perms?
What exactly is your problem?
gray haven
H'loo. So, my apologies if this is a dumb question, I've done a bit of searching and haven't found what I'm looking for, BUT.... Going through Waterfall to PaperMC 1.14.4.... it seems like it only works if PaperMC is set to offline mode, is that correct? And that my UUID will necessarily change because I'm going through Waterfall?
reef fulcrum
read the install guide, you need to setup ip forwarding and ensure that your stuff is secured proeprly
gray haven
The reason I ask is because I'm trying to figure out if I can continue to use the whitelisting that's built into the PaperMC server or if I'll need to get a whitelisting plugin for Waterfall.
Yeah, I've got it set to localhost internally....
reef fulcrum
enable bungee in spigot.yml, and ip_forward in the proxies config.yml
gray haven
Yeah, did all that.
In both spigot.yml AND in paper.yml because I'm not sure which one it's actually using.
I mean, in paper.yml it was already set to true.
reef fulcrum
if you do that properly, then the UUIDs won't change
gray haven
The IP definitely seems to be being forwarded now (though it wasn't at first, which was weird, don't think I changed anything)
Hmmmm.
reef fulcrum
as the proxy will pass through the IP and UUID
and, on that note, make sure that you didn't configure the proxy to be in offline mode
gray haven
No, proxy is not in offline mode.
That's interesting, the whitelist seems to have updated my UUID?
I guess it's working, then, maybe I don't need to worry about it.
reef fulcrum
the way the whitelist works is... odd
gray haven
But I can keep using my whitelist from PaperMC, I guess?
reef fulcrum
well, you might wanna consider using a bungee plugin instead, but, yea, should work for that server
next harbor
I have a question about the architecture of a network with waterfall.
How should I distribute chat messages on several servers? Should I send the chat messages via a plugin channel to the proxy, which then distributes the messages to the servers?
Would this slow down the proxy of large networks with many messages?
reef fulcrum
/ideally/ if you can do it on the proxy, do it on the proxy
caveats come into play with placeholders, but, I have doubts that it would really be a massive slowdown, not ideal, but, bleh
next harbor
So can I assume that the amount of data generated by chat messages is negligible compared to the data generated by movement, particel, etc.?
reef fulcrum
wouldn't be something of concern until you get uber large
but by then, you'd probs more be on multiple proxy instances, etc
next harbor
Hmm is there an estimation to much player you could solve it so that the chat messages go through a proxy?
1000 players or 5000 players or even more^^
reef fulcrum
the network concerns are basically none
You gotta remember that the server itself would be sending that info anyways
you're just adding extra processing into it, which shouldn't be that bad anyways
only real caveat is if you're planning to actually deal with servers hitting 1k players, you more wanna look towards proper networking solutions, bearing in mind that multi-proxy setups are somewhat neccesary there
next harbor
Ok, thank you for your help.
violet shard
nuvotifier make lag?
unreal stag
@violet shard https://github.com/NuVotifier/NuVotifier/wiki/Troubleshooting-Guide#nuvotifier-is-lagging-my-server
NuVotifier is incapable of causing lag when used by itself. NuVotifier's activities always take place off the server threads. Only when a vote is received and validated does NuVotifier take any sort of action that requires access to the server threads.
You filed one of the least useful bug reports I've ever seen in the bug tracker. I'm going to need a lot more information than "epoll make me server have 300% cpu usage pl0x halp"
violet shard
what information can i give from votifier plugin lol
only short config
and no logs
unreal stag
Start by responding to my comment and don't go off control panel "CPU usage" figures
next harbor
One more quick question. I've been researching how to set up multi proxy networks. Unfortunately I could not find any information about how the plugin channels behave.
Does a paper server send the message to every waterfall, do I understand this correctly?
reef fulcrum
no, it just sends it to one
which is where stuff like rabbit and redis come into play
next harbor
Well i read the tutorial code from bungeecord about the Pluginmessages, but there is no information how to select the proxy to send the information to.
Is a random proxy selected?
unreal stag
whatever proxy of the player you send it to gets the request
reef fulcrum
the "non-player" one just picks the first player online to use
next harbor
Oh i overread it, my fault. Sry
spiral crater
That's coming down from a plugin, phanatic, invalid json? o.O
@reef fulcrum yes a bungee plugin
reef fulcrum
bungee made some changes with hover events
Its's basically upset that a hover component isn't using json
that for a 1.7 client or something?
(or, just a general one?)
toxic sun
Heya, having players connect to server with forge mods seems fine but the moment they run a command via bungee they're kicked with the error NullPointerException : null @ net.md_5.bungee.util.chatcomponenttransformer44
Latest build of Travertine also
I also have the console error via proxy if you need it.
I also apologise in advance if i'm being a pain
It's probably something silly like startup parameters
reef fulcrum
update
toxic sun
Oh okay xD
Cool that fixed it xD, and here i was thinking i was on the latest version. Sorry for wasting your time.
marble kelpBOT
Project Travertine build #150: SUCCESS in 2 min 25 sec: https://papermc.io/ci/job/Travertine/150/
Project Waterfall build #361: SUCCESS in 5 min 2 sec: https://papermc.io/ci/job/Waterfall/361/
spiral crater
it was for anyone
reef fulcrum
might be fixed now
toxic sun
yeh its all good
thorny ridge
Is there any way to make the Bungee/Waterfall config leave comments alone? i.e not wipe them on start?
reef fulcrum
nope
thorny ridge
Is it possible to request that functionality for Waterfall?
[Meaning will it actually be considered]
reef fulcrum
issue is that we rely on saving for adding config options, etc
comments in yaml just don't work
thorny ridge
Fair enough, was worth a shot
reef fulcrum
velocity is literally a complete rewrite of a proxy
issue with waterfall is that we're based on bungee, so we get all the issues from there
Well, yes, because they rewrote the entire stack
Issue with waterfall is that we're based on bungee
We have o be weary about touching stuff, even deep, because then we fight against stupid plugins and such
So, if somebody could provide a useful means to reproduce this type of thing, it's like; I could look into it, spend a few dozen hours fixing it, and then be unable to keep the fix because it breaks 20 dozen common hackjobs people love to have on their proxy
the whole "just rewrite the crap" was like... the entire winning moving and pretty much essential to actually fix it
We know it exists
We just have literally 0 ways to reproduce it
lolno
Quite simply, I don't get paid for this shit
I have 0 interest investing cash into this
toxic sun
So not crazy sure if this is related to travertine or not since it's only started happening since this morning. When joining my modded server (Connecting using travertine build 150) i'm kicked with the error invalid payload register
Only with large modpacks
crisp thistle
are you using a normal forge server or is it a Bukkit+Forge server software?
toxic sun
Second one
crisp thistle
Mohist?
toxic sun
CatServer
crisp thistle
ah
toxic sun
Allows spigot plugins too
reef fulcrum
if that's the actual server doing the kick, glhf
crisp thistle
I had that issue with Mohist on 1.15.2 without any mods or plugins
reef fulcrum
CB has a limit on the # of plugin channels
if they kept that on catserv, that's just comical
crisp thistle
Maybe try with Magma
dusky current
Im trying too hookup a forge server to my waterfall? I am getting this error when I try and join https://pastebin.com/KJg5JewZ
Do I need to give my Waterfall more ram or somthing?
reef fulcrum
you probably have a mod sending data which waterfall doesn't understand, or something is mangling a packet
there is an option to disable entity metadata rewriting in waterfall.yml that might help
dusky current
I mean that might be another issue but I also had online mode toggled to true lmao
ok yeah Ill do that too
ok its working now but all my plugins on waterfall are freaking out
like it wont let me in
Only viabackwards actually
How do I stop Bungee plugins from attacking a server
lol
reef fulcrum
That's a via issue
Well, "issue"
more a limitation of.. well, versioning
you can probs disable those warnings
dusky current
ok then just gonna remove it from bungeecord and run it seperate
royal hawk
Remind me to implement proper forge 1.13+ support one of these days
strong mountain
is there a list of commands for waterfall somewhere? My players can't use /server after replacing with bungee
reef fulcrum
waterfall uses all the same commands as bungee
strong mountain
Hmm why would /server no longer work all of a sudden then?
strong mountain
Log on the hub server or waterfall server?
reef fulcrum
on the proxy, given that that is what has the server command
strong mountain
Waterfall logs are empty, command just isnt even in there & same for the hub server
so its not showing on the proxy at all
/bungee command returns "Server is running waterfall version 1.16 R2"
Fixed
Guess my perms plugin just wasn't loading
leaden igloo
What are the perms for like /server and suchh?
pale flax
hi, can I remove these commands from tab complete? All from proxy.
leaden igloo
Or is it the same perms as Bungee, too?
pale flax
pale flax
Must be incorrectly registered then.
eternal cairn
no, they work fine
eternal cairn
oh plugin ones yes
pale flax
yepp
pale flax
I'll just add it myself xD but thank you
eternal cairn
doesnt work on the bungee side tho
pale flax
This is the bungee plugin
eternal cairn
there one register the perms in code
pale flax
oh?
eternal cairn
yes
pale flax
I can do that too I guess
vital moon
for some reason when adding a updated server from 1.15.2 to 1.16.1 to waterfall, all items are missing
pale flax
Ok I got all but /bungee
@vital moon waterfall is just a proxy, it doesn't manage items if I'm correct..
Does /bungee have a special perm?
oof.
vital moon
seems like online mode has something to do about it
changing to server to offline gets rid of items in inventory
pale flax
I'm pretty sure they are different UUIDs with the same name? on/offline mode thing?
Can I not change this btw? Does it have to be /1?
sturdy cloak
Isn't it just in the config.yml? @pale flax
I may be wrong though
sturdy cloak
I'm trying to set up a firewall, but I can't figure out what my Bungee server IP is...
With IPWhitelist, the IP I had to whitelist was different from the IP my host told me the bungee was on and different from the IP that I can connect to. I can't connect to the IP I had to whitelist with IPWhitelist.
Would I have to use my connectable IP for the Bungee IP or the IP that IPWhitelist had me whitelist for the firewall?
crisp thistle
you have to use your connectable IP for the Bungee I guess
the IP you have in IPWhitelist is probably 127.0.0 1 which = localhost
so that would be why it couldn't work when you tried to connect to the server
and actually, maybe you can also use the 127.0.0.1 IP I guess if it's on the same machine
I'm not that experienced and others will probably have more details so I can't really confirm anything
sturdy cloak
It's on the same machine. It wasn't 127.0.0.1 but it was a different reserved IP address (source: https://en.wikipedia.org/wiki/Reserved_IP_addresses)
Interestingly with IPWhitelist I initially tried whitelisting the connectable IP but that did nothing, it had to be the reserved IP address
crisp thistle
you'll probably need someone else to answer that because I'm really not that experienced =/
sturdy cloak
alright, thanks for the help :)
granite latch
i have a problem here
not really a command block stuff question here, but i need this sorted out
How come my farmer villager isn't planting/harvesting any crops?
i have 1.16.1 paper server
like, every farmer isnt harvesting ore planting
eternal cairn
granite latch
sorry
tidal musk
i keep getting this error
tidal musk
marble kelpBOT
you have an error in your waterfall.yml
tidal musk
i didn't configed anything
but here
use_netty_dns_resolver: true
disable_modern_tab_limiter: true
log_initial_handler_connections: true
throttling:
tab_complete: 1000
game_version: ''
disable_entity_metadata_rewrite: false
tidal musk
can you name me the wrong config?
marble kelpBOT
?paste
Please paste all your code/errors/startup logs at http://hasteb.in/
ok actually it might also be an error in your config.yml
try pasting it in there and if there is an error fix that one: http://www.yamllint.com/
plucky imp
Does anybody have any ideas why my players cannot use /server <servername> on our waterfall network anymore? It is claiming they don't have permission, but they do and literally nothing has changed (as far as I know) in about a week on the server and now it all of a sudden says this?
marble kelpBOT
use lp verbose mode to see which perm it's checking I guess
plucky imp
Ok so I'm getting this when players use /server creative
[13:39:44] [LuckPerms Pool Thread #2/INFO]: [LP] VB > sapphirecrafty - bungeecord.command.server - undefined
reef fulcrum
it's saying that you've not given them the perm
it's undefined, bungee perms default to false
plucky imp
But I did add it at least I think I did
[13:46:46 INFO]: [LP] default already has bungeecord.command.server set in context global.
well, is that user in the default group?
If LP is saying something is undefined when it's not, you should go speak to them
plucky imp
I resolved it, it was a LuckPerms issue. Thanks for pointing me in the right direction though
tidal musk
should i set ip forwarding to true or false?
reef fulcrum
you probs want it set to true
tidal musk
what is saver
reef fulcrum
safer?
tidal musk
yes
reef fulcrum
Read the bungee install guide, specifically the firewall one
If you need to use one of the plugins, don't use any of the suggestions, but use bungeeguard
tidal musk
i know there is a bungee op bug you can give op if you can connect on special servers
ok
half carbon
hello, I am gettting an error from a plugin: https://pastebin.com/zpszZbwT
reef fulcrum
speak to the plugin author
junior sable
is there a way to download always the latest version of waterfall? Maybe .../waterfall/latest.jar for example
zinc sapphire
vocal lake
Why does ProxyPingEvent sometimes return a null response?
Like event.getResponse() is null
vocal lake
No idea
I got a NPE with that and after checking out some ping plugins, all of them check if the response is not null
reef fulcrum
it can't be set to null, the proxy will throw an exception if a null was actually in there
vocal lake
🤔
spiral crater
@reef fulcrum that issue is still happening
do i need to look into fixing the plugin or is that something that would be fixed on bungee side
patent lantern
don't ping server rules
proven crypt
Whats the API for waterfall downloads? lawl
eternal cairn
marble kelpBOT
Project Travertine build #151: SUCCESS in 1 min 37 sec: https://papermc.io/ci/job/Travertine/151/
Project Waterfall build #362: SUCCESS in 3 min 26 sec: https://papermc.io/ci/job/Waterfall/362/
royal hawk
@proven crypt I hope you’re not making an auto-updater
Wasn’t meant as mention. God damnit. Discord mobile ffs. Example given, wasn’t sure if it was gonna do it
unreal stag
odd musical choice @royal hawk
royal hawk
Was listening to radio at the time
Alas this stuff qualifies as summer music somehow
¯_(ツ)_/¯
shadow iris
pale flax
Hi, does anyone know a plugin which blocks the tab completion for /bungee and other proxy-only commands?
marble kelpBOT
not granting the permissions should make them not show up
tidal musk
This may be a dumb question. But if I'm running a paper server, would i just need to add the waterfall jar to the plugins folder to allow it to connect to another server using portals?
Basically I have 3 paper servers and I want to be able to bridge them together by using a portal or sign or command.
hearty coral
waterfall is a server
to do what you're describing, you will need a 4th server running waterfall
the players will connect to this server, and it will forward them to one of the 3 other servers
after which they can switch across those 3 servers as desired
tidal musk
i have a problem with waterfall permissions i gave myself admin rank in the config but i cant use commands like /server
Ok thank you. I'll look into that later. I'm sure it has all the available commands to make that happen. But essentially, once I make a waterfall server, I can create different portals within that server that can go to any paper server? And then would I be able to able to travel back to the waterfall server from a paper server?
marble kelpBOT
for portals you need a plugin adding those, by default only the /server command exists to switch servers
tidal musk
Ok that's not bad. Will a waterfall server accept spigot/bukkit plugins?
scenic pebble
reef fulcrum
it's a proxy, not a server
it's sole job is to bounce you between servers, it's not a server that runs server plugins
tidal musk
Ok I'll do some googling unless you can point me in the right direction :)
frosty finch
well players are connecting to your waterfall / bungeecord proxy, and that proxy is redirecting them automatically to the specified server
the proxy isnt a server with a world
tidal musk
Oh I seeeeee
frosty finch
its just a bridge
the player will only see a "world" switch
while they actually switch the server
tidal musk
That makes more sense. I was thinking of bungee being more of a lobby room/world type of thing. I've been running servers for years but have never messed with bungee. Typically if I wanted another world, I'd just create another world within the same server and restrict and grant access to specific plugins to be used on that world. But now I have 3 different servers to help load balancing with a completely different player demograph. But am just now trying to figure out how I can connect all 3 servers together to make them feel like they are different worlds within the same server and not have to constant log out and log back in with a different port number.
frosty finch
have you ever played on a large server like hypixel?
tidal musk
I dipped my foot in there once lol. But my daughter has played on them. I've seen what it looks like and the servers I've been building are more like hypixel just more streamlined and less chaotic.
But my end goal is to have many servers all connected together
frosty finch
well, hypixel is using bungeecord aswell
you connect to a bungeecord proxy and it will redirect you to the lobby, when you join a game it will redirect the player to the server where that game is running
waterfall is a better fork of bungeecord
just if you didnt know
but make sure to just let the proxy be able to talk with the outside world
otherwise other people can just put your server in their bungeecord proxy, join and do whatever they want, giving them op, grief etc.
tidal musk
I want to keep this going. Keep talking, but I'll brb.
Hello on my waterfall, i am an warning strange :
Query - Incorrect magic!: /96.22.38.113:53079
with +200 times this warning to 5 seconds.
What is it ?
Can we talk about that last part real quick. Where you said to let just the proxy speak to the outside world and if not, then others can take it over. I understand how proxies work and port forwarding but that concept is confusing me lol. So if I don't set it up correctly somohow, someone could just take over the server with their own proxy?
frosty finch
yep
when you enable bungeecord on your paper servers, every bungeecord will be able to connect to your servers, so you need to make sure your servers behind the proxy wont accept any other connections
the most simple thing would be (if everything is on the same physical machine), to change the ip of each server in the server.properties to 127.0.0.1 and put that ip into the bungeecord config
dusty kindle
so super dumb question, possibly
can I chain two waterfalls together
as-in, have one waterfall pointing to another waterfall as the server
reef fulcrum
no
dusty kindle
gotcha, thanks for the quick reply
frosty finch
but you can use multiple waterfalls for the same servers
heady sedge
Just any player?
I imagine there are loads that make player join messages network wide but not sure about only to staff members
hearty coral
.g bungee join messages spigotmc
marble kelpBOT
(DiscordBot) https://www.spigotmc.org/resources/globalbell-join-leave-switch-messages-for-bungeecord.10239/ -- GlobalBell : join/leave/switch messages for BungeeCord |...: "All players can now see when another player join/leave/switch server on the network ♢ A lot of configuration ♢ Welcome message ♢ Aliases for ..."
rain loom
what is the best cloud system for waterfall that can automatically create paper servers? im thinking of using timocloud but i dont think there is much documentation on that, + i want to be able to define a custom choosing strategy for choosing free servers based on statistics of the player (aka skill-based and ping-based matchmaking)
median isle
I dont have an issue, more of a question just to clarify something. Theres guides (for bungeecord) which tells you to secure your servers if you run them on the same machine using IPtables. Now I have only just moved onto waterfall so I actually dont have much experience with it. But am i correct to assume that seeing as its essentially a proxy, the players are all connecting THROUGH the proxy when playing on the servers rather than just being bounced to them. I wanted to clarify as I will be securing my server completely soon but I wasnt sure if i needed to block the ports for every server aside from the waterfall proxy.
rain loom
@median isle im pretty sure the answer is yeah, just make sure your waterfall proxy is the only thing that can connect to those servers
prisma whale
@median isle yeah or you can look into bungeeguard if you cant
meager citrus
Hi, could I somehow disable these messages? ServerConnector [Survival] has disconnected disable logging
violet shard
what is normally should i allocate ram for 200 player and 5 plugin only?
noble kestrel
hey guys,
since yesterday it not possible to connect to the server. i didnt changed anything, just restarted the whole server.. so anybody knows whats happend?
Bungee config:
ebon kernel
@noble kestrel does your console output any thing on join? if can i see it?
noble kestrel
16:48:45 [INFO] [/213.239.204.169:38307] <-> InitialHandler has pinged
16:48:48 [INFO] [BigLife5] disconnected with: Could not connect to a default or fallback server, please try again later: io.netty.channel.ConnectTimeoutException
16:48:48 [INFO] [BigLife5] -> UpstreamBridge has disconnected
just this
reef fulcrum
The connection timed out
as it says
something is wrong between the proxy to the backend server
ebon kernel
The server your trying to connect to (TeamBlock) is down
or timed out
try restarting that
noble kestrel
but why? i didnt changed anything and it worked for months
ebon kernel
is the TeamBlock server down?
noble kestrel
no
noble kestrel
strange thing too: if i restart the bungee, it takes a long time if it restarts. since yesterday
ebon kernel
probably just because its trying to ping servers that are not working/timed out
noble kestrel
16:51:53 [INFO] [BigLife5] disconnected with: Could not connect to a default or fallback server, please try again later: io.netty.channel.ConnectTimeoutException
16:51:53 [INFO] [BigLife5] -> UpstreamBridge has disconnected
16:51:55 [INFO] [/95.90.223.202:13791] <-> InitialHandler has connected
16:52:01 [INFO] [BigLife5] disconnected with: Could not connect to a default or fallback server, please try again later: io.netty.channel.ConnectTimeoutException
16:52:01 [INFO] [BigLife5] -> UpstreamBridge has disconnected
ebon kernel
As i have already said TeamBlock is either timing you out or down
noble kestrel
the console says its on
reef fulcrum
connection timeout means that the server basically didn't even say "am here" back
either the server is down or misconfigured, or, there is a networking issue between the proxy and the server
noble kestrel
i basicly restarted the whole server yesterday (on time 30 days). after that restart, the game servers died
reef fulcrum
So, maybe you didn't change nothing
Maybe the reboot caused something to change
We don't know your setup so we really can't say
reef fulcrum
Well, yes, but that basically tells us nothing about your actual setup
best guess is that you're using ptero
noble kestrel
yes
reef fulcrum
do the servers respond to commands? tried checking that the interface is properly setup? etc
noble kestrel
the server respond to commands yes
rose kite
hello
I want to switch to waterfall which antibot do you prefer?
meager citrus
Is it possible somehow? https://discordapp.com/channels/289587909051416579/555462214606389275/731321424266133515
unreal stag
@neon palm No proxy is compatible with Forge post-1.12.2 due to changes in the handshake protocol. Just use Fabric.
gray skiff
hey guys. Please advise plugin for bungee portals for 1.16.1
tidal musk
hey can somebody help me please? i have a problem with the bungee perms
i gave myself admin but i have no perms
listeners:
- query_port: 25577
motd: '&1Another Bungee server'
tab_list: GLOBAL_PING
query_enabled: false
proxy_protocol: false
forced_hosts:
pvp.md-5.net: pvp
ping_passthrough: false
priorities:- lobby
bind_local_address: true
host: 0.0.0.0:25565
max_players: 1
tab_size: 60
force_default_server: true
remote_ping_cache: -1
network_compression_threshold: 256
permissions:
default: null
admin: - bungeecord.command.alert
- bungeecord.command.end
- bungeecord.command.ip
- bungeecord.command.reload
- bungeecord.command.server
log_pings: true
connection_throttle_limit: 3
server_connect_timeout: 5000
timeout: 30000
stats: 87b0b3b0-9ee5-470c-99a8-450fc009a0ef
player_limit: 1
ip_forward: true
groups:
Skygamer_777: - admin
remote_ping_timeout: 5000
connection_throttle: 4000
log_commands: false
prevent_proxy_connections: false
online_mode: true
forge_support: true
disabled_commands:
- lobby
- disabledcommandhere
servers:
lobby:
motd: '&1Just another Waterfall - Forced Host'
address: localhost:25575
restricted: false
oneblock:
motd: '&1Just another Waterfall - Forced Host'
address: localhost:25545
restricted: false
reef fulcrum
- Don't flood the chat, use a damned paste site
tidal musk
ok
reef fulcrum
- Did you have a permission plugin on the proxy? are you sure that you have the correct perms? admins don't have all perms, etc
reef fulcrum
the bungeecord wiki has a list of commands and their perms, etc
empty path
oo shoot wrong server
reef fulcrum
- that mod is doing an illegal cast
meager citrus
log_initial_handler_connections: false
log_pings: false
Also tried:
https://www.spigotmc.org/resources/console-spam-fix-bungee.65471/
https://www.spigotmc.org/resources/quietcord.12940/
and nothing works... so how can I stop it finally? 😅
meager citrus
and this should happen by normal on waterfall? it has to be annoying for people, no?
why any of mentioned settings from config.yml and waterfall.yml don't work for this?
gloomy sundial
is there a way to list all available commands from console like how u can do ? with a backend server
timber stag
What's the difference between waterfall's
log_initial_handler_connections: true
and bungee's
log_pings: true
reef fulcrum
ones for pinging, the others for the initial handler log messages
timber stag
so which woukd [ip] <-> InitialHandler has pinged fall under?
timber stag
but it also says pings,
reef fulcrum
That would fall under pings
timber stag
ah okay
so setting waterfall's to false would also disable stuff like <-> InitialHandler - read timed out and connected and all that
reef fulcrum
no
waterfalls would disable the initialhandler has connected messages
errors, etc, are all logged
timber stag
ah, no matter what
gloomy sundial
i dont even see the initial handler connections option in the config 🤔
reef fulcrum
that's in waterfall.yml
gloomy sundial
ah
reef fulcrum
we had an option for both of those because it was super trivial to spam
gloomy sundial
is "game_version" the thing you see in the server list when you're using an outdated client
reef fulcrum
My irk with adding options for disabling logs is that 99% of the people asking for them to be removed are the people who expect us to solve all their other issues for them, so, removing logging to have to deal with further crap because they disable logging is very... er...
That's the version that's displayed
gloomy sundial
ah
reef fulcrum
otherwise you get the long list of supported versions
gloomy sundial
yeah i've got another plugin that overrides that anyways lol
icy needle
Hey guys, I'm trying to set up a waterfall server but stuck on this error, does anyone know what I'm doing wrong, or of a guide I could follow? https://gyazo.com/d30f259ab8caa21b52d5cc3bbb4a87ae
(this happens after I run the Waterfall jar)
reef fulcrum
i mean, it thinks your plugin directory isn't a directory, so...
icy needle
nevermind lol, running it as superuser fixed it
somber geode
Our waterfall just ran out of memory, this isn't normal right? 🤔
We are 2 versions behind, but seeing the commits none of them mention memory leak of any kind. Could it be a plugin that's causing this?
kind karma
@somber geode not waterfall's issue, some plugin is probably doing something wrong
somber geode
I see, I have atleast one suspect
reef fulcrum
Biggest one is that BungeeServerInfo stores a "queue" of packets, iirc, probs some plugin tryna shoot messages over the network doing it badly
woven stone
How much ram is recommended for Waterfall on a server with ~30 people? (without any plugins installed)
reef fulcrum
no
woven stone
Is it normal for Waterfall to go up to 60% cpu usage if a player changes a server than drops back down to 10%?
young jasper
How do I pull commits into Paper-Server
I do git pull upstream master but it says up to date even though it clearly isn't
reef fulcrum
looks at the channel name
odd hatch
We run a network that can get a few hundred joins within a couple of minutes and have problems with Waterfall choking the CPU and making login times suck
Is this my infra or is there something I can do to halp
unreal stag
@odd hatch Care to share details of your infrastructure?
odd hatch
We run a Waterfall instance on it's own hardware as the only publicly accessible server to the network. Then, Waterfall distributes users into 1 of 4 hubs located on another machine in the same datacenter
unreal stag
What hardware do you use?
Because if it's some crappy i3 from 2011 or a Pentium 4 I'm not going to be surprised
unreal stag
PhoenixNAP 
Yeah, we use them too.
odd hatch
Just pretend that my coworker selected Ubuntu for the OS. We've reinstalled it already :|
atomic grail
the people installed it as centos even tho we chose ubuntu
it's running on ubuntu just isnt updated on the panel
odd hatch
^ Said coworker
unreal stag
Classic PhoenixNAP incompetency.
odd hatch
We're at just about 3GB with 30ish players, which doesn't feel.. right
https://u.dejay.dev/😧🚲😳.png And then this
unreal stag
They fucked up our Ubuntu install too. We got an outdated version of the OS (16.04) and they gave us a really tiny root partition. Thankfully, they did configure it as LVM so we could online resize... but it was a terrible ordeal.
They'll get the OS install right if you use CentOS I guess.
odd hatch
I make it clear to install "Latest Ubuntu (20.04)"
unreal stag
And it was kinda scary doing the online resize over SSH
Already, enough complaints about PhoenixNAP.
How many players are we talking?
Usually BungeeCord-based stuff doesn't respond well to sudden player count spikes
Your best strategy would be to switch to Velocity. I'm biased but also not even joking, if you want to improve BungeeCord scalability you have to rewrite the entire thing.
You could run multiple proxies (on different servers).
odd hatch
Well, my guy is a fairly large youtuber who likes running spontaneous events
@atomic grail How many users was the last event
He didn't reply, 130
reef fulcrum
Part of my worry is around the netty dns stuff
odd hatch
So that's at least 70 people connecting as soon as the announcement drops
And then trickle down
reef fulcrum
There was some bug report that I spotted a while bak where in some cases, the netty DNS stuff ends up causing a network lookup in the event loop or something dum from what I recall
unreal stag
Yeah, I gave up on Netty DNS. In Velocity 1.1.0, I just do a blocking lookup on a separate thread. That works fine for the most part.
Performance is still really good afterwards, AsyncHttpClient is probably caching it for a bit afterwards
odd hatch
I might run a Velocity experiment to see how it goes...
Did Velocity ever get 1.16 client support
odd hatch
Trying to figure out what plugins I have to rewrite lol
atomic grail
hey tux, if you or anyone still around, does velocity have multi proxy support?
quartz vale
HEEEL! P... network-compression-threshold sat to -1 on waterfall but! 64 or 512 on the servers! I read 64 is better then on spigot 512 to reduce lag! WTH! I want to give ppl. form us a better connection.
unreal stag
You need to set the network compression threshold on the proxy to the default (256) to avoid issues with some clients. For the backend servers you should set it to -1 as having it enabled just burns CPU for no good reason
quartz vale
Thank you so mutch... I was totaly confused!
unreal stag
If your proxy is starved for CPU, then bumping the compression threshold to 512 or 1024 may reduce load on the proxy somewhat
quartz vale
But will it help ppl from overseas to set the threshold up? I had some problems but i have geo blocked China and conf faild2block because i was totally under "attack" ...
reef fulcrum
lower threshold = smaller packets = more CPU burning
cloud torrent
Thank you
quartz vale
I haven´t paid attention to this section so thanks to you too 😄
cloud torrent
Damnn. Yeah that helps a lot!
quartz vale
It's like the tps command in papermc! 😄
quartz vale
Scot, my server is virtual on a proxmox with 4 cores... i think that the Load average is for the whole server cpu in my case because my server is almost idle..
marble kelpBOT
Please paste large logs to a pastebin: https://paste.gg
tidal musk
I use the following (https://paste.gg/p/anonymous/d2b53a9ad88d4a9eb9fcd64b22778027), to broadcast to all servers. It works most of the times but I have scenarios where it wont broadcast at all. It seems like in some cases where a player joins a empty server (I broadcast a "X moved to Y" message) and I use my function it wont broadcast to any servers at all.
Is there something sketchy I'm doing or is sockets not reliable?
errant pivot
players and me getting kicked for running commands
from more than 5 plugins so far
kind karma
lower threshold = smaller packets = more CPU burning
It isburningbecause writing proxy in JAVA was VERY BAD idea
(even Velocity went on the same road, lol)
torn river
Is there a way to fix this? They're the exact same modpack im on.
unreal stag
@kind karma Well, actually it's more complicated than "Java slow af". The real answer is "JNI method calls are fucking expensive". I optimized this in Velocity - stopped using zlib and started using libdeflate. If you spend less time doing native crap it should in theory pay for the cost of dropping from compiled Java code to native code.
kind karma
I know)
unreal stag
So libdeflate will let me (de)compress in one shot, it doesn't require streaming (in fact you can't stream with it!), and I can do just one JNI instead instead of at least 2, sometimes up to 5-10
kind karma
We wrote our own Rust library for ciphering, it spends like 10 ns on each iteration
and JVM is like: okay I'll just slow this down to like 10 milliseconds per iteration
that's why we lost hope on JVM and writing proxy with Rust
unreal stag
I've been thinking about combining the compression and encryption into one
But I've also been thinking about "fuck it, I'll rewrite it in Rust and find a way to have Java compatibility hooks"
kind karma
One game on Hypixel's murder mystery is 0.6 seconds of CPU time. (with Rust)
that's how terrible JVM is.
Yeah, we are doing exactly what you was thinking about
.We are leaving JNI callbacks.
marble kelpBOT
(DiscordBot) The geocode API is off in the Google Developers Console.
kind karma
oops
unreal stag
Well, at Hypixel scale, you have money and resources to throw at just getting rid of Java
kind karma
nah I'm just testing this on Hypixel
And as far as I know hypixel was using BungeeCord
unreal stag
Testing in production? 
kind karma
don't know how it is now though
frosty finch
somehow it must be tested
kind karma
Nope, there are quiet a few things left.
We have an ability to mod the client, so I'm making it now so I can change compression threshold during PLAY, for example.
We was actually thinking about modifying JVM itself
but like
fuck it
unreal stag
C++ gives you permanent brain damage, I can't blame you for not wanting to touch HotSpot lol
kind karma
well it's not CPP
it's actually terrible hotspot sources
and there are so many things you need to do to compile it
so we just gave up on that
frosty finch
what about openJ9 ?
kind karma
OpenJ9 sucks
unreal stag
I doubt OpenJ9 is going to be significantly better, just saying.
kind karma
OpenJ9 beats HotSpot only in GC.
U can try Zing tho
We did not try, but don't think that it will improve something
unreal stag
Basically, JNI is expensive. You need to avoid it as much as you can.
kind karma
You can't
sadly
i think that most time is spent on actually preparing for native method call
like pushing stack frame, safepoint, injecting stuff like JNIEnv etc
(JavaCritical does not help)
unreal stag
Unfortunately, I had to make trade-offs when making Velocity, and I made the trade-off of focusing on what the community expects, and on my own familiarity.
If performance was my utmost goal, I'd probably write it in a native language.
kind karma
Well, there are some things you can do :)
unreal stag
Safe to say that Rust or Go would have much better performance than Java
kind karma
Go is like Java
it does have GC too.
Well, there are some things you can do :)
Like unpacking only N amount of bytes before actually doing stuff
Unpack 5 bytes, read packet id, Velocity does not use it - pass through
That's what we are doing rn.
unreal stag
That only helps for decompression. It's not very helpful if we have to compress packets, which tends to be the common case.
kind karma
You don't need to compress packet again :)
It's already compressed
you pass through compressed buffer
unreal stag
Also wouldn't help me now, I'm doing libdeflate, so we decompress the entire buffer anyway
kind karma
that's sad.
The disadvantage is that all backend servers must have the same threshold as the proxy.
But I think everyone can deal with that
unreal stag
The idea is interesting, but in practice, the approach I'm taking is pretty amenable to the "just throw more cores at it" solution
kind karma
bruh
while we are speaking about proxy stuff
I've noticed that BungeeCord reads unused server side packets
That might be optimized a bit too.
Chat, for example, is registered for both Client and Server, however, only upstream handles it.
unreal stag
At least Velocity optimizes that.
unreal stag
Though it's possible libdeflate optimizes compression to the point where JNI overhead definitely takes up most of the room.
From some experimentation, I've seen a 2.5x improvement on both amd64 (Zen 2) and aarch64 (AWS Graviton2).
unreal stag
This overhead can be significant if a native function is simple and is called frequently. JDK has a private API called Critical Natives to reduce overhead of calling functions that do not require much of JNI functionality.
unreal stag
I wasn't even aware Critical Natives even existed.
kind karma
Yeah, it was first time introduced to speed up ciphering
But it does not help
Nothing changes, lmao
unreal stag
The only thing I can think of is to reimplement AES CFB-8 over ByteBuf directly in Java.
Or swallow your pride, and use Cipher and heap buffers.
unreal stag
I know, it just makes a byte array copy.
unreal stag
I've read JDK code
timber root
Hello, I am just now switching to Pterodactyl panel. I saw where there was an Egg for Paper and that is working great. I was wondering if there's a script somewhere that works for Waterfall? I tried duplicating the paper one and renaming everything to waterfall, but that didn't work
unreal stag
Search for a Waterfall egg
frosty finch
there is a bungeecord egg
you can just replace the bungeecord.jar with the waterfall.jar
and rename it
unreal stag
Basically, unless we can instrinctify native ciphers in the JVM, it doesn't matter.
kind karma
or just switch to Rust!
unreal stag
Yeah, but that requires a rewrite of everything to Rust.
For previous reasons, I can't really do that
The "array access" benchmark is kinda a joke. Just writing the relevant code directly in Java would seem likely to be faster...
Also, the sum loop would be a field day for autovectorization optimizations.
I might look at JavaCritical anyway, maybe you missed something. Or I need to dig more deeply into HotSpot code.
kind karma
may be.
unreal stag
My understanding of what happens on JavaCritical:
- Create a C-style stack frame
- Freezes GC, could do a garbage collection.
- Move arguments to proper register or stack locations according to ABI.
- Calls JavaCritical function.
- Check if safepoint is needed.
- Unwrap object result and reset JNI handles block.
- Remove the stack frame and return result to client.
So it does reduce the steps by about half, in theory.
The calling convention is very C-esque too
It's also mentioned that the method only gets called from compiled code (C1? C2?)
Hopefully it's both C1 and C2.
LWJGL developer says JavaCritical is snake oil https://github.com/jnr/jnr-ffi/issues/86#issuecomment-250325189
In fact, we tested functions that do absolutely nothing and there was no real difference between critical and standard JNI.
that's pretty damning
timber root
I tried taking the BungeeCord egg and modifying everything to match Waterfall. It says this when I try to start the server No server egg configuration could be located; aborting startup.
Is there some place I can get a Waterfall egg? I tried searching for it but couldn't find it
unreal stag
But there is https://github.com/jnr/jnr-ffi/issues/86#issuecomment-252527201, showing that improvements can be made to get it to near bare-metal costs.
odd hatch
@timber root It's not that hard
Just reset back to the BungeeCord Egg
Then after your first startup, replace the server.jar with the Waterfall jar from source.
Writing the actual egg if you need to deploy Waterfall is different, but I think at that scale you'd have less issues/know how to say that's what you want :P
timber root
We're trying to make it where it automatically uses the latest build of Waterfall
unreal stag
kind karma
@unreal stag C2 only.
odd hatch
unreal stag
So yeah, looks like JavaCritical either doesn't help or is a regression
odd hatch
Then upload that json file.
timber root
It still says
No server egg configuration could be located; aborting startup.
odd hatch
Make sure the egg is selected in here. https://u.dejay.dev/🥔🍢🥠.png
I think at this point you'd benefit from the Pterodactyl discord rather than here :P
odd hatch
Try manage, rebuild server container
then start again, otherwise idk what to tell you but try the pterodactyl cord
unreal stag
@kind karma I did some more Googling and found some pretty interesting things about JNI critical methods.
JNI critical methods were introduced to optimize cryptography... on Solaris: https://bugs.java.com/bugdatabase/view_bug.do?bug_id=7013347
Then the Solaris port got deprecated, and there was an issue to remove JNI critical methods: https://bugs.java.com/bugdatabase/view_bug.do?bug_id=8233343
Well, looks like LWJGL uses JNI critical methods. But at best, you maybe get a 5% performance improvement for a function that does basically nothing.
timber root
Yeah I've tried both that^ and reinstalling
unreal stag
LWJGL probably pays the cost as it's the kind of thing where you need low latency...
timber root
I'll go to their discord
unreal stag
You can make it faster if you're willing to play fast and loose.
tidal musk
I have connected all my servers to bungee. When i try to connect to one of them (not connecting to the proxy) is says "Ip forwarding is enabled but spoofed data can't be decoded or missing"
unreal stag
Project Panama is supposed to usher that future in
kind karma
nah
unreal stag
This article gave me a crazy, stupid idea
kind karma
yes?
unreal stag
What if I could batch a bunch of writes together and do just one JNI call?
unreal stag
I'm crazy, I will find a way lol
signal trout
I'm a bit new to bungee stuff and I am wondering what plugins are updated that let you send players between servers from the "client" servers, I've looked around quite a few forums and found some super outdated stuff but surely there must be some working plugin that everyone uses?
timber root
If you're a fan of signs, there's this: https://www.spigotmc.org/resources/bungeesigns.6563/
I'm sure there's other plugins for npcs or portals though
signal trout
Is there something that lets you use like a console command on the client server to send someone to another server?
unreal stag
@kind karma working on it now
PoC shows promise. Now I've managed to halve the number of JNI calls per packet, which is pretty sick.
Now I want to try Tuinity as it has flush consolidation.
unreal stag
The results are in. It does seem it helps, but not by as much as I'd like. It does however reduce JNI encryption calls by half in the Velocity case, which probably makes it worth it.
Switching servers seems to get the most benefit from my brief testing.
reef fulcrum
just wait till you see all the scoreboard packets
unreal stag
Yeah. Main catch is that Velocity encodes the length buffer in a slightly unusual way, just to avoid a memory copy.
unreal stag
@reef fulcrum I own a copy of FeatherBoard. Challenge accepted
reef fulcrum
there was a PR for it, but, in the midsts of my death and well, still kinda dead
unreal stag
Figured it out. The FlowControlHandler was basically making Velocity flush consolidation toothless.
Now I get to come up with an alternative
kind karma
what's the purpose of flow control handler
I don't get it
alright I see for what it was used but still don't get it
why is it used in velocity
unreal stag
@kind karma ServerConnectedEvent
I came up with an alternative, that only holds messages if autoRead is off, but releases them on read() again.
Working great with Tuinity, and even with Paper's flush consolidation.
frosty finch
btw thanks for reccomending zulu, works quiet nice
unreal stag
It’s working but I need to do some extensive sanity checking.
unreal needle
I am wondering if anyone has had this issue before and might know what needs to be done to fix.
I am apparently stuck on using waterfall build 357. If I go higher or lower in build versions, my players get stuck on a loading terrain screen when they try to change servers. They are able to connect to the hub server just fine, but when they try to connect to a different server it gets that problem. I had tried updating my servers to 1.16 and it fixes the terrain issue, but we are being forced to wait on a couple plugins to update before we can commit to 1.16. So we are currentlyrunning on 1.15.2.
cursive flicker
@unreal needle I had a similar issue with an outdated forge client version, updating that fixed it.
marble kelpBOT
Project Waterfall build #363: FAILURE in 1 min 1 sec: https://papermc.io/ci/job/Waterfall/363/
Project Travertine build #152: FAILURE in 17 sec: https://papermc.io/ci/job/Travertine/152/
Yippee, build fixed!
Project Travertine build #153: FIXED in 1 min 33 sec: https://papermc.io/ci/job/Travertine/153/
Yippee, build fixed!
Project Waterfall build #364: FIXED in 3 min 33 sec: https://papermc.io/ci/job/Waterfall/364/
unreal stag
For PacketCompressor, the answer is "md_5 allocates too small buffers", "md_5 is doing an extra JNI call", and "zlib is flat-out slow"
If you optimize all three, compression still eats CPU time, but as optimized as you can make it.
The other factor is that it does a memory copy if it decides the packet shouldn't be compressed.
The JNI boundary is expensive to cross, and md_5 hasn't really done a good job of reducing those boundaries.
Heck, there's even areas of improvement in Velocity, some of which I'm working on right now. But Velocity's JNI boundary crossing for compression is very well-optimized.
quartz vale
Is there an easy fix if i wish to have more than one bungee/watherfall to the same servers? RedisBungee is not maintained anymore and the problem is that players can't see each others in the tab menu 🙂
reef fulcrum
no
it's for online mode, that option lets you break the bungee mode == online mode assertion on the proxy
mental eagle
encountered exception: net.md_5.bungee.util.QuietException: Server is online mode!
I have minecraft.net account why it says that
mental eagle
thx i opened waterfall servers long ago and i forgot them 🤔
honest spruce
Hi, can you see, why my waterfall display only the players connected in the hub server ?
green palm
since that's the case you probably forgot ip-forwarding too, don't forget that
reef fulcrum
ping passthrough does that
honest spruce
So, i have to set it false
tidal musk
is this 1.16? https://papermc.io/ci/job/Waterfall/364/
green palm
yep
tidal musk
alright thx
candid panther
Hello ! I have some issues with a player who can't join my bungee network, i don't know if it cause of bungee or other stuff, but can you help me ?
Sometimes Disconnected
Sometimes An existing connection was forcibly closed by the remote host
keen patio
formal cloak
so i have a problem with bungeecord and paper 1.16.1
Logs:
Bungee https://pastebin.com/UYvYURs5
My server https://pastebin.com/XYNV8Paw
ancient atlas
If you are using paper try using waterfall instead of bungeecord
wispy fable
Hi! Does anyone know why this is happening when i try to join the bungee hub? I think i have ip forwarding turned on in the bungee config as well
kind karma
You need to enable bungeecord mode in paper.yml/spigot.yml as well
formal cloak
Its Always in spigot.yml even tho your on paper
kind karma
Paper uses it's own option for something else.
rough lintel
do i have to download bungee cord or can i just use waterfall?
reef fulcrum
waterfall is a fork of bungee, meaning that're essentially the same thing
unborn arrow
So, does the MOTD only update on a restart and not a greload?
unborn arrow
and its supposed to just pull the config.yml from the waterfall directory right? cause my MOTD refuses to cooperate at this point 😂
reef fulcrum
yes
unborn arrow
whelp, i guess imma just restart the network during low count. The MOTD for every server listed in the config.yml is correct, but clients still load the previous one. Thanks anyways zzzCat
reef fulcrum
wait, you're using forced hosts?
reloading the server list is kinda shaky and often doesn't work properly, it gets reaaaally weird; Need to look into it but dead + low priority
reef fulcrum
the listener list should always reload
/greload literally throws out the existing listeners
unborn arrow
well, see how it says 'This shit work?'
here its like this after a greload
we're restarting network now
reef fulcrum
best guess would be check your plugins
unborn arrow
a restart cleared it up
tidal musk
how to make it so that when a server shuts down it sends everyone to the hub?
agile sigil
You icon is 64x64 and in png format ?
reef fulcrum
known issue
quaint bolt
So, we're getting attacked:
[18:35:38 WARN]: Query - Incorrect magic!: /65.xx.xx.xxx:123
Questions:
- How tf can they specify their outgoing port, that's usually not possible without a custom driver, right?
- What query/magic is this about?
- Could this cause the whole server to be come irresponsible? Shouldn't it only affect a few threads/cpu or is this about the linux limit of pending connections?
- Couldn't I safely write a patch to blocklist those IPs directly for at least until the restart?
- Would
connection_throttleandconnection_throttle_limithelp in any case?
unreal stag
- UDP contains little verification, and you can send raw IP packets. In fact, you can easily spoof source IPs and ports.
- It's to prevent UDP amplification attacks
- UDP doesn't have connections, though it's possible it can burn up CPU time on one core due to how UDP is handled by the kernel (it's handled like a single connection)
- You could also just disable query, then you get full protection
- No
quaint bolt
Oh, didn't know it's UDP, that changes quite a few things. I guess it just used up all our bandwidth or exceeded some queues then.
I guess blacklisting queries after the first false attempt won't help much as it's on a too high level already and I can only prevent reading the message but the OS and JVM needs to handle the packets already?
We're just implementing geoblocking now, as the attacks came from russia, china, etc which is not our playerbase
amber root
Is there any way to have like a backup proxy should the main one go down so players just get moved over to the other proxy?
frosty finch
you can use multiple proxys for your servers, yes
but there is no "get moved over" function
except if they reconnect manually
frosty finch
no there is no such function possible
celest kindle
@agile sigil yep, but its only in 1.16 + , other versions work ok
delicate phoenix
There are pluins
Witch ones send the pleayer to an other server if the one falls they are on.
just replace the jar?
@tidal musk yes, should work
It shoud be enough.
silent hedge
A lot of fixes, performance improvements, better forge support
cursive flicker
Is it possible to connect to waterfall with a paper hub and move to a forge server on the network?
peak knoll
is there a way to reload waterfall's config
without restarting it and kicking players
cursive flicker
Technically I think there is a bungeecord command /greload but it is very buggy.
cursive flicker
Often crashes or crashes servers, doesn't remove servers if they were previously configured, etc.
raven pawn
@peak knoll yeah that doesn't happen too often
hidden star
reef fulcrum
java.lang.IllegalStateException: Cancelled ServerConnectEvent with no server or disconnect.
I erm.... I mean, it says?
A plugin cancelled a connection event and didn't send them anywhere else
tidal musk
My bungeecord keeps dying
lofty basalt
is bungee/waterfall considered safe once i have it all set up, or do i need additional plugins to prevent players from connecting to the servers in offline mode?
reef fulcrum
use firewall or localhost binding if they're all on the same box, or, use bungeeguard
lofty basalt
so i dont have to have the ports of the offline servers open to the firewall?
and localhost binding is just done in the server.properties, right?
reef fulcrum
you just need them accessible by the proxy, not the whole world
yes, just bind the servers on 127.0.0.1
lofty basalt
thats really smart!
so since its on the same machine its fairly simple to be locked nice and tight 😄
thanks for your help!
sullen flume
twin solstice
what
unreal stag
probably was looking for a wiki or docs of some sort
icy schooner
got this can someone help please ? https://pastebin.com/KcWME3Hg
marble kelpBOT
Project Waterfall build #365: SUCCESS in 3 min 11 sec: https://papermc.io/ci/job/Waterfall/365/
quaint bolt
btw looks like we got an NTP Amplification Attack, so processing speed wouldn't have helped much, we've now just iptabled port 123, which was the reason for this strange port
frosty finch
[12:19:19 ERROR]: at net.md_5.bungee.conf.YamlConfig.getListeners(YamlConfig.java:245)
[12:19:19 ERROR]: at net.md_5.bungee.conf.Configuration.load(Configuration.java:92)
[12:19:19 ERROR]: at io.github.waterfallmc.waterfall.conf.WaterfallConfiguration.load(WaterfallConfiguration.java:49)
[12:19:19 ERROR]: at net.md_5.bungee.BungeeCord.start(BungeeCord.java:278)
[12:19:19 ERROR]: at net.md_5.bungee.BungeeCordLauncher.main(BungeeCordLauncher.java:62)
[12:19:19 ERROR]: at net.md_5.bungee.Bootstrap.main(Bootstrap.java:15)```
latest waterfall seems brokenyour config is busted
tidal musk
hey
how would i make my waterfall version thingy look like this server https://cdn.discordapp.com/attachments/722198099132678148/733711393441972244/sexy.png
is it like a plugin
or do i have to edit the source
or config
reef fulcrum
plugin
slate moat
Is bungeecord plugins compatible with waterfall?
reef fulcrum
yes
slate moat
Thank you
wispy fable
Anyone know why i cant use /server on a bungee server as an op with * perms?
Is it because i dont have those perms in the permission plugin of bungee?
winged cape
what does waterfall do and is it better than paper
reef fulcrum
waterfall is a fork of bungee, paper is a fork of spigot
winged cape
what does that mean
reef fulcrum
essentially, it's nothing to do with paper
winged cape
what does waterfall do different to paper
winged cape
so like hypixel?
reef fulcrum
essentially, yes
winged cape
can you run multiple servers on the same world so that it's faster
reef fulcrum
you can't share worlds across servers
keen thistle
hello i keep receiving this error every time someone tries to join on the proxy error : io.netty.handler.codec.DecoderException: javax.net.ssl.SSLHandshakeException: PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to requested target. could someone please help.
reef fulcrum
That generally means that your java install is busted
either that or your network is jacked up
keen thistle
ok thanks
warm violet
reef fulcrum
fork of bungeecord
keen thistle
hello when i load up minecraft it says my server is out of date
i installed the latest version of water fall
keen thistle
ok
keen thistle
it still says that the server is out of date
fallow flume
Hi I’m thinking of syncing 2 survival servers together. So setting resource worlds to a separate server.
Does Waterfall support this (with player data syncing/inventory/other plugin data etc)
cursive flicker
That requires a plugin
delicate phoenix
Hello! My players can't login in to the server:
REMPIZSOLTI lost connection: Internal Exception: io.netty.channel.unix.Errors$NativeIoException: syscall:read(..) failed: Connection reset by peer
What can cuse this issue?
reef fulcrum
Connection reset by peer
something else on the network closed the connection in an unexpected manner
delicate phoenix
Hm, that's strange...
I've only got my bungee, 5 servers and a website
The servers are running as localhost
reef fulcrum
No
as in, "something between your server and the client"
All the server knows is that the connection was reset unexpectedly, meaning that something closed it
soft forge
Hello! Im using Waterfall on my server and I notice this option on the config.yml
remote_ping_cache: -1 What does it does?
There is no information about it on the bungeecord guides, so I guess it is something about waterfall?
reef fulcrum
it's not actually used
🤷♂️
okay, maybe it is; but, basically, it just caches ping requests
faint yarrow
I am having this issue right now: https://pastebin.pl/view/a3b76f89
Anytime someone tries to join my survival server (which I updated to 1.16) it throws that error on the proxy. It connects them and then you are immediately disconnected and sent back to the hub. Anything helps, thanks
frosty finch
update your proxy
faint yarrow
I have
reef fulcrum
it's an issue with multiple servers storing the same team names across them
server A has a team, player jumps to server b, server b has a team with the same name = boom
faint yarrow
I am not familiar with team names
reef fulcrum
It's a long standing bungeecord issue; Best fix I got so far basically involves causing other issues, soooo.... #rip
scoreboard teams
generally used by plugins for the scoreboards in the sidebar, or tab names, etc
faint yarrow
yeah I tried disabling all of those still no good
reef fulcrum
some plugins are really bad and store their team names on the main scoreboard, which means that they'll be in scoreboard.dat in the world folder
faint yarrow
Would clearing that give me any different results?
reef fulcrum
if the "bad" data is in there, yea
faint yarrow
Yeah that did fix it thank you
reef fulcrum
at the end of the day, the ultimate issue is that server A and B tried to define the same team name, some plugins really have no consideration to bungee, which is a pita because bungee has no considerations to scoreboard plugins
It's literally fun
faint yarrow
Yeah it's not too fun haha, I can join no one else can
But you recognized the issue and helped point me in the right direction so thank you
faint yarrow
I still can't find a solution for this problem I have been trying for hours straight now, I have removed every placeholder, tab and scoreboard plugin and still no luck I cannot find a solution for this.
Is it something with my proxy? Or the hub and the survival server? Hub is running on 1.8 whilst survival is running 1.16
reef fulcrum
all I can tell you is that you have duplicate teams
This is generally down to plugins on both servers not doing stuff to avoid this type of collision
flint shell
Hi, when my players die, they were kicked out of the server with the following message
the server you were previously on went down, you have been connected to a fallback server
not sure what happened
can anyone shed a light
cursive flicker
Any logs in paper console or waterfall console?
cursive flicker
Upload them to pastebin/gist/hastebin
faint yarrow
@reef fulcrum thank you I was able to finally find the issue, had a plugin messing it up, thank you so much
flint shell
okay
cursive flicker
I don't know if I can help, probably not, but no one else will be able to unless you do that 🙂
cursive flicker
You're running ResourcepacksPlugins?
Looks like it may be an error/incompatibility with that to me.
cursive flicker
I have a test server I run at home with all the same plugins etc. So I can test removing / adding things easily.
reef fulcrum
needs updating
cursive flicker
@flint shell see above.
flint shell
that's an excellent idea!
and turns out that bungeecord resource pack plugin is the source of the problem
cursive flicker
Yep, now you know where to look which is at the very least useful
flint shell
yea, i really need to get into the habit of looking at the console first
i'm having trouble setting up resource pack for all of my servers, do you have any ideas?
cursive flicker
Not me sorry, I've never used one
Usually plugins have a guide how to set them up on bungee
Or if you search "pluginx bungee" you can find something.
cursive flicker
No problem, someone may be able to help you here, but generally speaking you have to find out about the specific plugin so better to get specific support, if that makes sense?
flint shell
yep! thank you
tender axle
Is this "error" ok?
[07:41:19 INFO]: Failed to download: SSLHandshakeException : Remote host closed connection during handshake @ sun.security.ssl.SSLSocketImpl:994```am i doing something wrong?
trying to setup a spongeforge 1.12.2 server connected to waterfall
but it just keeps getting stuck at Logging in...
doesn't seem to be the firewall either
forge_support is on in the bungee config
i also have entity metadata rewriting disabled which might cause issues
bungee can definitely reach the server
since i use a motd passthrough plugin
spongeforge has bungeecord ipforwarding on & is in offline mode
my vanilla servers connected to the same proxy work fine, everything is on the same server aswell
tender axle
@twin solstice try this https://www.reddit.com/r/admincraft/comments/eiw2uc/waterfall_wont_connect_to_spongeforge_server/
marble kelpBOT
(DiscordBot) Waterfall won't connect to SpongeForge server : admincraft - 4 comments, 19 points - Ecrfour, 6mo ago
twin solstice
fucking bruh moment if that's it
tender axle
lol lmk if it works
twin solstice
though i only have spark + phosphor + sponge
tender axle
its modded minecraft. if you told me the solution to my problem was to touch my nose with my knee, i would believe you.
twin solstice
nvm apparently you have to enable bungeecord as a module in sponge's config too
got em
marble kelpBOT
Project Waterfall build #366: SUCCESS in 3 min 21 sec: https://papermc.io/ci/job/Waterfall/366/
tidal musk
What is more stable, Waterfall or PaperMC?
reef fulcrum
waterfall and paper are two different things
eternal cairn
which kiwi is best, apples or pears?
tidal musk
since i wanna make a plugin
reef fulcrum
They're both "stable", if you ignore vanilla being derp
making a plugin is irrelevant to stability
You throw the plugin where it makes sense
tidal musk
ill choose spigot since someone had a paper server and he changed to spigot and it was more stable with very many plugins
green palm
yeah I'm somewhat confused how he jumped from "What is more stable, Waterfall or Paper" to "I'll choose spigot"
tidal musk
ill use PaperMC with paperlib rather then
green palm
So you know, Paper is a typical server you're used to with blocks, zombies, creepers, mobs, worlds, etc.
Waterfall has none of those. It's a proxy.
tidal musk
oh
green palm
bungeecord is also a proxy
waterfall is a fork of bungeecord
spigot is not a proxy but a typical server
paper is a fork of spigot
tidal musk
i know theyre forks
tidal musk
if i overwrite /tell with Waterfall, will both the server and the proxy react on that command?
tidal musk
only proxy will
reef fulcrum
literally nothing to do with waterfall
wispy fable
Anyone know what could have caused this massive change in the bandwidth?
thats bandwidth for my entire dedicated server which hosts only a bungee w/ a few servers
frosty finch
mass teleportation onto the same location and keeping the player there
a server crash
players quitting the server
steel orchid
violet radish
I use pterodactyl with waterfall and my papermc server, but it doesnt work like it did before, can it be because of ptero and how do I fix that?
I get also Could not connect to a default or fallback server, please try again later: io.netty.channel.AbstractChannel$AnnotatedConnectException
mighty anchor
just coming in here to ask a quick question, does waterfall suppoer new versions of forges (after they changed there packet system)