jup
#waterfall-help
tidal musk
my hub servers
log_initial_handler_connections: false it is already set to false
so i dont know why it still does that
old stone
Hello,
I encountered something really weird. My Waterfall Proxy revived a few thousand exceptions which overloaded bungeecord to that point where we were no longer able to join.
All of them looked like this:
https://pastebin.com/raw/t5LraWan
Could this be some type of crasher? Or a (D)Dos Attack against Waterfall? Because all of the request were from different IP addresses.
If yes, is there a way to migrate this? Or is there anything I can to prevent Waterfall from crashing?
Thanks!
tidal musk
... important part of stack trace is missing
reef fulcrum
There have been some recent attacks against proxies, however mitigating is it not really super viable; There is a bunch of patches floating around to "fix" that, but the patch itself is broken and could be used to take down a proxy if not severly hamper performance in the correct environment
hard trench
Does anyone know anything about my issue above?
somber pier
How do I send someone to a server?
hard trench
/send (name) (server)
old stone
@tidal musk that is the entire stack I got from the console. After that it just repeats.
@reef fulcrum So its already known and being worked on? And are thes "fixes" something in Waterfall it self or a Plugin? If yes, could you maybe link me such a Plugin?
reef fulcrum
I have ideas on how to provide a way for plugins to be told when shit is going down, it's just not perfect
It's pretty hard to tell between shit going south for somebody and an actual attempt to impact the proxy itself, There was something called cprotector, but as I said, there "fix" is bad and can be used to cause issues for the proxy in other ways
Also, please don't ping
old stone
Oh ok sorry for the Ping then.
But I mean servers like hypixel mineplex and also some smaler server (100 to 500 users) are not affected by this. What are they doing different to prevent this? Custom Bungeecord? Own private Plugins?
reef fulcrum
They generally scale out their infrastructure, so asshats have a much wider surface to try and impact
tidal musk
noshit that they run custom solutions
hallow veldt
^
the only thing I know that hypixel uses thats public is citizens
but even then its most likely changed
tidal musk
everything bigger than 75 players/day use custom stuff, theoretically
reef fulcrum
The attack takes advantage of the most principle concept of servers:
They listen to network traffic
The only "fix" is to break the fact that it's listening to their traffic
tidal musk
perfect ddos mitigation is in fact pulling your plug yes 
there is no silver bullet, what might work for you, might not for me or other 50 server admins
old stone
dang, so I am pretty much screwed
stiff yarrow
is there any documentation on how to setup waterfall?
reef fulcrum
advice is to follow the bungee wiki, it is mostly self explanitory
stiff yarrow
ow ok
silk bronze
so where the wiki for the waterfall.yml? I don't want to ask for every configuration but they mess with a lot that will interfere with my authentication plugins
please provide me a link, i can't believe that you find nothing on the github about it
reef fulcrum
The documentation was on an old site that was lost
Well, what documentation there was; there is nothing in waterfalls config that's relevant to authentication, the async DNS option can have some issues for dns lookups in some janky networks
silk bronze
Is it impossible to rewrite a wiki for only 8 settings?
reef fulcrum
it's more my time is pretty limited and what time I do have I'd rather invest into programming stuff
if you have a question you're more than welcome to ask it
The papermc documentation is also open source, contributions welcome
silk bronze
So basically nobody besides you know what those options are for
I have multiple questions, because i actually want to inject the mojang authentication and have a setup where not all servers are only accessible over the hub server, so i have different input stream of players
so those 2 things bug me
game_version: '1.14.4' allow_empty_packets: false
reef fulcrum
Other people know what stuff is, just well, as you can tell, people love contributing towards documentation
game version sets what is displayed on the server list when there is a protocol version mispatch
empty_packets is basically a hackaround to prevent the proxy from booting clients when the servers sends an empty packet
silk bronze
as well as this option disable_modern_tab_limiter: true since it seems to be only there for compatibility while i will only run 1.14.4 and above
reef fulcrum
disables the tab limiter for 1.13+ due to changes in how tab completion works, the limiter caused issues
silk bronze
i see, that booting of clients maybe can become a problem for me
i will investigate this later on
so thanks for clearing out the most so far
i'm not a big friend of that many options in general if they are only fixes
reef fulcrum
That one ain't really a fix
That's basically an option to admit you have problems but don't wanna fix 'em
silk bronze
😄 alright
one other questions which might be a bit more important for the proxy:
network_compression_threshold: 128 is in general better right?
there are many topics around the forums about this option
reef fulcrum
That's subjective
silk bronze
my cpu is very powerful and the connection to US players is alright but could be better
reef fulcrum
Basically, it's a tradeoff between CPU and bandwidth
lower values = more CPU usage and less bandwidth usage, too high can impact those with slower networks, too low can impact those with slower computers
silk bronze
exactly because the other part needs to decompress it later on
alright i will investigate it, but as far as i noticed it will influence the ping shown in tablist
so basically the lower you set it the less filled the ping bar is
hard trench
If anyone here has used RedisBungee in the past, have you had a issue where the playercount placeholder takes more then a minute to update? been trying all sorts of fixes and changes, restarts over the past 3 days and no luck
Any tips are appreciated im down to try anything because im out of ideas
reef fulcrum
Given that it's a premium plugin, probably wanna speak to the dev of the plugin
hard trench
I guess i can give that a shot
old stone
mikroskeem I just had another look at the pastebin and you were right, I indeed forgot a bit of the stack trace. So here again the stack trace but now it should have everything in it. https://pastebin.com/raw/PPH5kpak
tidal musk
🎉
that does not tell me much though :-/
if i could replicate that and monitor what happens with proxy's memory usage then that'd be more useful i guess
old stone
If it happens again I will create a TCP Dump and check the Memory usage. Also I think I might have found someone using the exploit: https://youtu.be/_vNuZux_c4k?t=11 and they call it Null ping crasher
marble kelpBOT
(DiscordBot) NULL PING CRASHER - length 3m 58s - 28 likes, 7 dislikes (80.0%) - 198 views - mori0 тεαм ĦǺχÒЯ on 2019.09.22
hard trench
While i wait for the author of redisbungee to reply to my support ticket does anyone who has used it before known why placeholders (%redisbungee_total%) that display player count take more then a minute to update? I've tried so many things. The player count is the server list when not logged in updates instantly but ingame its so slow..
Im really in a dead end here ive been looking around for about 4 days now, on forums, discords and everything and the is nothing i could find that helps me
tidal musk
have you tried fetching the placeholder manually?
e.g /papi parse yourname %redisbungee_total%
full cypress
redisbungee_total is tallied via plugin channels so it should have at most a few seconds of delay
It was always relatively accurate last time I used it until I just directly started pinging in a multibungee setup
tidal musk
oh hmm
something smells bad here
i've seen plugin channels to fail sending data before
hard trench
Tried that just now, its also taking the same time to update
installed the Pinger placeholderapi expansion and ping the bungee and that updates instantly, its probably not the most efficient thing but im not sure
at least it works, so if i dont end up finding a fix for it then ill just use that
tidal musk
heady wave
anyone can help me with this?
[21:01:46 WARN]: Event PreLoginEvent(cancelled=false, cancelReasonComponents=null, connection=[/ip:53841|nickname] <-> InitialHandler) took 275ms to process!
i have some of this errors and sometimes i have dupe players in glist
old stone
mikroskeem I think I found something event better than the client which crashes the proxy. On this discord https://discord.gg/kV8agmw they clain to have a custom bungeecord version which prevents this exploit. I havent tried it yet and probably will not until I can be 100% sure its nothing shady.
reef fulcrum
Memory usage is irrelevant, basically; They're just sending invalid packets and the proxy does what it does, gets upset reading the data and throws an exception
There is no real fix for it outside of blackholing people who are sending that traffic
if that's the cprotector thing, their fix is stupid and doesn't actually fix it, it just adds a "firewall" to the proxy, and is implemented in a way that will cause performance issues down the line if you're under enough of an attack
old stone
Nah this is not the cprotector plugin. They forked bungee and changed a few lines to 'fix' it. But I cant test it because I dont have the crasher it self so I cant confirm any of this.
reef fulcrum
fixed_classes.rar seem familar?
cprotector stuff for bungee wasn't a plugin, but basically they messed with a few classes; and while yes, it's a hackaround for the situation, it's questionable code at best
old stone
Oh I didnt know that, I thought cprotector made a plugin instead of forking bungee. Well good to know.
unreal stag
@reef fulcrum My interest is piqued...
Although Velocity wouldn't spam the exceptions that @old stone has mentioned, it would be vulnerable to the attendant risk of having to deal with many connections per second
reef fulcrum
I do gotta look into cleaning up some of the exceptions where possible
My overall idea is that it's not something which has a true fix, but I do wanna look into exposing some events so that it can be offloaded to another plugin
marble kelpBOT
Project Waterfall build #295: SUCCESS in 3 min 32 sec: https://papermc.io/ci/job/Waterfall/295/
merry minnow
I'm trying out the "Restricted" option for the first time, reloading with greload just to see if it actually works, but it doesn't. I've not given myself the bungeecord.server.[servername] permission but I can join just fine. Do I have to restart the proxy to get that to work?
reef fulcrum
Iirc yes, due to how serverinfo stuff works in bungee, the alternative is that we risk breaking plugins
merry minnow
Fair enough, time to make everybody angry because the server is offline for a whole 20 seconds 😄
Oh I see the issue. CMI's method of moving me to a different server seems to override the Restricted flag
tidal musk
what the heck is that dude trying to say who opened the issue recently? 
"old code" is unrelated to current codebase
also "new code" is not much different from "old code", besides varint reading happening inside while(true) & breaking if (in & 0x80) != 0x80; instead of happening inside do/while & running until (in & 0x80) == 0x80
unless i really overlook something
sand steppe
I don't see any potential dos attack with the current read method
reef fulcrum
empty packet has been fixed in waterfall for a while, we just had it hidden behind a configuration option for reasons
I still maintain that those cprotector patches are bad, and tux (original waterfall dev and velocity developer) can tell you how harmful those patches can be
sand steppe
can we ping em im interested 
reef fulcrum
Tldr; they just add the host address of anybody who ends up sending a bad packet to a list
sand steppe
local?
reef fulcrum
said list is checked in the decoder everytime a packet is sent, there is no TTL or anything on said list
They've also introduced some crash risks because they don't understand what they're doing; e.g. that list is is read/modified on multiple threads
reef fulcrum
I also don't think that they disambiguated between clients and servers, so potentially rip if your server sends a bad packet too
unreal stag
yes, a list
I think there would be no problems integrating the least objectionable bits into Velocity and Waterfall. Waterfall and Velocity are ignoring empty packets now, for instance.
having reviewed it, cprotector patches may fix a few DoS issues but open others related to the fact it leaks memory (no TTL) and is not thread safe as it should be
reef fulcrum
The disambiguation thing is irrelevant, I had a look, they basically only a certain phase of the connection
(That's what I get for tryna remember stuff after having just woken up \o/)
unreal stag
and of course, it does an O(n) operation every time a packet is unframed, so it'll kill your performance
something tells me the author only tested with maybe one or two clients and a few blocked IPs at best
heady wave
Today my server crashed with this error, i'll post it just for debug
https://pastebin.com/dfDS4wiP
heady wave
my terminal being stupid? WTF i had all terminals closed and server crashed by himself
Sorry for the wrong channel
reef fulcrum
Your server had an IO error reading from the terminal
heady wave
well effectively i have a problem with terminal, when i press backspace it deletes all written text, and i don't know why
tidal musk
said list is checked in the decoder everytime a packet is sent, there is no TTL or anything on said list
holy fuck 
muh "exploit fix"
I also don't think that they disambiguated between clients and servers, so potentially rip if your server sends a bad packet too
that's true
old stone
Whut, does anyone know why this could be happening?
Waterfall: https://dl.pluoi.com/5d8ba81585126.png
Paper: https://dl.pluoi.com/5d8ba8869e3db.png
Situation: I try to connect to a server called "knockbackpvp" which is in our waterfall config. But when I try to connect I get no message in my chat and the console just spams "connected" and after that "disconnected"
tidal musk
That means the player cant connect to that server, and the proxy disconnect the player from proxy, so far i know.
elder spear
does someone know if the latest waterfall update caused server-icon's to break/not load
pulsar fjord
there were reports of this as well with sponge. dont think its waterfall
unless they did the same update as sponge lol
elder spear
I dont have sponge
pulsar fjord
is this the issue you are having? https://github.com/SpongePowered/SpongeForge/issues/2887
pulsar fjord
atleast you have a starting point.
nocturne trench
Could it be that waterfall preloads all classes of a plugin?
Or in other words, how can I use my own class loader?
nocturne trench
Ignore me being to stupid to post in the right channel
manic agate
Any new yet on issue #419?
Sometimes spigot plugins don't like the quick reconnections on the hub server.
reef fulcrum
I mean, nobody replied to it, and I've been kinda out of it recently
manic agate
hmm ok. I can only hope for a fix. Thnx anyway.
junior talon
hey
i really need help
I am getting weird errors
2019-09-27 23:05:48,561 main WARN Advanced terminal features are not available in this environment
[23:05:48 INFO]: Using standard Java JCE cipher.
[23:05:48 INFO]: Using standard Java compressor.
[23:05:48 INFO]: Enabled Waterfall version git:Waterfall-Bootstrap:1.14-SNAPSHOT:c354548:295
[23:05:48 INFO]: Discovered module: ModuleSpec(name=cmd_alert, file=modules\cmd_alert.jar, provider=JenkinsModuleSource())
[23:05:48 INFO]: Discovered module: ModuleSpec(name=cmd_find, file=modules\cmd_find.jar, provider=JenkinsModuleSource())
[23:05:48 INFO]: Discovered module: ModuleSpec(name=cmd_list, file=modules\cmd_list.jar, provider=JenkinsModuleSource())
[23:05:48 INFO]: Discovered module: ModuleSpec(name=cmd_send, file=modules\cmd_send.jar, provider=JenkinsModuleSource())
[23:05:48 INFO]: Discovered module: ModuleSpec(name=cmd_server, file=modules\cmd_server.jar, provider=JenkinsModuleSource())
[23:05:48 INFO]: Discovered module: ModuleSpec(name=reconnect_yaml, file=modules\reconnect_yaml.jar, provider=JenkinsModuleSource())
[23:05:48 INFO]: Loaded plugin reconnect_yaml version git:reconnect_yaml:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[23:05:48 INFO]: Loaded plugin cmd_find version git:cmd_find:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[23:05:48 INFO]: Loaded plugin cmd_server version git:cmd_server:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[23:05:48 INFO]: Loaded plugin cmd_alert version git:cmd_alert:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[23:05:48 INFO]: Loaded plugin cmd_send version git:cmd_send:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[23:05:48 INFO]: Loaded plugin cmd_list version git:cmd_list:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[23:05:48 ERROR]: Exception in thread "main" java.lang.RuntimeException: Invalid configuration encountered - this is a configuration error and NOT a bug! Please attempt to fix the error or see https://www.spigotmc.org/ for help.
[23:05:48 ERROR]: at net.md_5.bungee.conf.YamlConfig.load(YamlConfig.java:80)
[23:05:48 ERROR]: at net.md_5.bungee.conf.YamlConfig.load(YamlConfig.java:64)
[23:05:48 ERROR]: at net.md_5.bungee.conf.Configuration.load(Configuration.java:75)
[23:05:48 ERROR]: at io.github.waterfallmc.waterfall.conf.WaterfallConfiguration.load(WaterfallConfiguration.java:49)
[23:05:48 ERROR]: at net.md_5.bungee.BungeeCord.start(BungeeCord.java:270)
[23:05:48 ERROR]: at net.md_5.bungee.BungeeCordLauncher.main(BungeeCordLauncher.java:62)
[23:05:48 ERROR]: at net.md_5.bungee.Bootst
Idk why
I have not touched my config
marble kelpBOT
you have an error in your waterfall.yml
marble kelpBOT
hard to say without knowing your full error log and config file
marble kelpBOT
yes
marble kelpBOT
the error is in teh waterfall.yml, you posted the bungee config.yml
nocturne trench
And line 38 you have one ' too much at the end
The config.yml is broken too
Or it might be just that one
marble kelpBOT
oh yeah, the config.yml is broken too
nocturne trench
Fix that and try again
junior talon
@nocturne trench i ment since it worked
junior talon
I know
I am trying to connect to pc
thanks
its all working now
players are connecting back
junior talon
what permission do you need for /send and /find?
junior talon
thanks!
I still cant execute the command
It says I dont have permission to execute this command
Pls help
i cant execute those commands
bungee thinks im a default when i do /perms
how do I change that
And ranks wont show up on kitpvp
junior talon
Anyone know any free domain registers that support ports?
toxic temple
Any ideas what's causing this? https://pastebin.com/p5P85Tiu
marble kelpBOT
@volcano_monster you will only find free subdomains, top level domains usually aren't free when using major DNS services. e.g. check out https://freedns.afraid.org, they have lots of TLDs to choose from for a subdomain
shrewd bluff
Hey, is it possible to separate the nether world and the overworld in 2 different servers with waterfall/bungeecord ?
marble kelpBOT
if you have a plugin that implements that ans shares user data, sure
somber pier
I set the lobby as priority in config, but it just kicks the player from the server still. They don't get sent to lobby
marble kelpBOT
what does the log say?
somber pier
What log?
marble kelpBOT
from the waterfall server
somber pier
marble kelpBOT
and the lobby server doesn't say anything at all? can you share your bungee config.yml
strange brook
wrong channel
hard trench
oopsd
junior talon
Our perms wont work and I need a domain register anyone suggest any it has to allow me to have a port at the end of the IP
junior talon
Well we have to have :25566 at the end to join for us
reef fulcrum
So, you would just chuck that at the end of the domain name
junior talon
Ok
reef fulcrum
The only way to get around specifying the port is with SRV records, but those are a bit meh and don't always work
reef fulcrum
Yes, the srv record would provide the port
junior talon
ok
reef fulcrum
Only issue is that some ISPs have stupid issues with SRV records, so some people might not be able to connect to your server if you rely on that
junior talon
I cant connect to skyblock it says I am already connected to it when i am connected to lobby
rose badger
I'm looking to set up a bungeecord/waterfall server soon. Does waterfall work with paper? (I'm assuming it does)
And do I need to run an extra server just for bungeecord? Can I run it on the hub server or no?
marble kelpBOT
Bungeecord/Waterfall is a standalone proxy server, it's not a plugin
but you can run it on the same dedicated/virtual server as your minecraft servers if that's what you are asking
it doesn't really need much resources (if you don't put lots of plugins on it)
rose badger
Ok
Yeah I have 2 servers and I want to connect them together via a lobby
Wait what if my servers are on separate dedicated machines?
marble kelpBOT
would still work but you would have to setup your firewall properly so that only the proxy can connect to your paper servers and not someone else
rose badger
ok thank you
junior talon
skyblock and lobby are the same servers
i was on lobby and restarted it and i stayed on
and when i try go to skyblock it says i am already connected to it
so i try lobby and it says the same
its so wierd
please help
I just stopped skyblock and lobby and i still stayed on
this is so wierd
please help
marble kelpBOT
you'll have to show some log files of the servers and proxy
junior talon
How
I am so confused
when someone leaves it says they left twice
Ok let me get some
[22:11:03] [main/INFO]: Using standard Java JCE cipher.
[22:11:03] [main/INFO]: Using standard Java compressor.
[22:11:03] [main/INFO]: Enabled Waterfall version git:Waterfall-Bootstrap:1.14-SNAPSHOT:c354548:295
[22:11:03] [main/INFO]: Discovered module: ModuleSpec(name=cmd_alert, file=modules\cmd_alert.jar, provider=JenkinsModuleSource())
[22:11:03] [main/INFO]: Discovered module: ModuleSpec(name=cmd_find, file=modules\cmd_find.jar, provider=JenkinsModuleSource())
[22:11:03] [main/INFO]: Discovered module: ModuleSpec(name=cmd_list, file=modules\cmd_list.jar, provider=JenkinsModuleSource())
[22:11:03] [main/INFO]: Discovered module: ModuleSpec(name=cmd_send, file=modules\cmd_send.jar, provider=JenkinsModuleSource())
[22:11:03] [main/INFO]: Discovered module: ModuleSpec(name=cmd_server, file=modules\cmd_server.jar, provider=JenkinsModuleSource())
[22:11:03] [main/INFO]: Discovered module: ModuleSpec(name=reconnect_yaml, file=modules\reconnect_yaml.jar, provider=JenkinsModuleSource())
[22:11:04] [main/WARN]: Could not load plugin from file plugins\BungeeTabListPlus_SpongeBridge-2.8.11.jar
java.lang.NullPointerException: Plugin must have a plugin.yml or bungee.yml
at com.google.common.base.Preconditions.checkNotNull(Preconditions.java:787) ~[waterfall-295.jar:git:Waterfall-Bootstrap:1.14-SNAPSHOT:c354548:295]
at net.md_5.bungee.api.plugin.PluginManager.detectPlugins(PluginManager.java:401) ~[waterfall-295.jar:git:Waterfall-Bootstrap:1.14-SNAPSHOT:c354548:295]
at net.md_5.bungee.BungeeCord.start(BungeeCord.java:267) ~[waterfall-295.jar:git:Waterfall-Bootstrap:1.14-SNAPSHOT:c354548:295]
at net.md_5.bungee.BungeeCordLauncher.main(BungeeCordLauncher.java:62) ~[waterfall-295.jar:git:Waterfall-Bootstrap:1.14-SNAPSHOT:c354548:295]
at net.md_5.bungee.Bootstrap.main(Bootstrap.java:15) ~[waterfall-295.jar:git:Waterfall-Bootstrap:1.14-SNAPSHOT:c354548:295]
[22:11:04] [main/INFO]: Loaded plugin EnjinMinecraftPlugin version 3.0.1-bungee by www.Enjin.com
[22:11:04] [main/INFO]: Loaded plugin BungeeKickListener version 1.0 by ryans1230
[22:11:04] [main/INFO]: Loaded plugin reconnect_yaml version git:reconnect_yaml:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[22:11:04] [main/INFO]: Loaded plugin cmd_find version git:cmd_find:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[22:11:04] [main/INFO]: Loaded plugin cmd_server version git:cmd_server:1.14-SNAPSHOT:c354548:295 by WaterfallMC
[22:11:04] [main/WARN]: Error enabling plugin BungeeTabListPlus_BukkitBridge
java.lang.ClassNotFou
thats the bugee one
marble kelpBOT
ideally we need the log when the disconnct/server change occurs
also you might want to use a paste site instead of pasting full log files directly into chat
marble kelpBOT
why?
marble kelpBOT
lol no, just provide the log files, that's what they are for
toxic temple
Hi, looks like this error https://pastebin.com/p5P85Tiu
is related to this the terminal changes introduced in #291, any ideas?
tidal musk
smells like replacing proxy jar while proxy is running to me
can anybody help me? so I have a minecraft server and I can manage it with pteroodactyl. but how do I drag in the paper.jar file to the server and run it?
wrong channel my dude
-> #paper-help
marble kelpBOT
hm, anyone have any idea why the permissions in the config.yml wouldn't work?
wait, I forgot that bungeecord doesn't have a child permission concept
marble kelpBOT
waterfall doesn't have chunks?
rough warren
what
rose badger
So I need a separate ip address for when I run the waterfall jar?
marble kelpBOT
no, you only need a different port
and if you have everything on one server you usually run all the paper servers on localhost
rose badger
I have 2 servers on separate machines, I plan on setting up bungee on one of them
rose badger
Should I run waterfall in a loop script?
like example:
#/bin/sh
while true
do
java -Xmx512M -Xms512M -jar waterfall.jar
echo Restarting in 2
sleep 2
done
like idk if theres any scenario where it would need to reboot
unreal stag
Waterfall could crash, or you might need to update/reconfigure a plugin.
rose badger
ok I'll keep it in a loop
green rune
I might be retarded, server-icon.png goes into the same location as waterfall.jar right
nvm i got it, the image wasnt actually 64x64 like the author said -.-
junior talon
When I try connect to skyblock it says im already connected to it when im connected to lobby
Please help!
reef fulcrum
Double check your configuration/plugins, that's not really one of those things which is possible on the proxy side unless you've got a weird plugin or you're actually trying to connect to the server you're already on
reef fulcrum
Waterfall doesn't even have a /spawn command...
junior talon
Sorry wrong place
worthy gate
hi, will waterfall support the latest snapshot?
reef fulcrum
No
olive robin
I installed waterfall and get"patching vanilla"
tidal musk
lol
broken roost
Hello, In my WaterFall setup some of my players have an Issue that in the Hub server they have a ping of <50ms, but the issue is when they switch to another server, their ping is <500ms and eventually they get kicked out, what can be the cause of this?
(I posted this in the wrong channels two times lol)
reef fulcrum
depends on the exact kick, but generally network/plugin related
broken roost
[18:35:04 INFO]: [Player] disconnected with: IOException : An existing connection was forcibly closed by the remote host @ sun.nio.ch.SocketDispatcher:-2
Sometimes is this message, sometimes is a "Timed Out"
[18:32:44 ERROR]: [Player|/Player_IP:28259] <-> ServerConnector [Hub] - encountered exception: net.md_5.bungee.util.QuietException: Unexpected packet received during server login process!
0d0100
And now this message that I didn't notice before
chrome flame
Hello, i wrote premium logging on my server but i want to keep cracked uuid, so i wrote java if (pl.premium.contains(e.getConnection().getName().toLowerCase())) { e.getConnection().setOnlineMode(true); e.getConnection().setUniqueId(UUID.fromString(String.valueOf(UUID.nameUUIDFromBytes(("OfflinePlayer:" + e.getConnection().getName()).getBytes(Charsets.UTF_8))))); }
to change uuid to cracked but it prints error: java.lang.IllegalStateException: Can only set uuid when online mode is false at com.google.common.base.Preconditions.checkState(Preconditions.java:444) at net.md_5.bungee.connection.InitialHandler.setUniqueId(InitialHandler.java:619) at pl.acmc.bungee.Join.onLoginEvent(Join.java:111)
do you know how to bypass it ?
strange brook
lmao
tidal musk
Having an issue with Travertine, probably because I'm dumb, but I'm running a 1.14.4 client, and the server behind travertine is 1.12.2, and I keep getting an Kicked whilst connecting to Lobby: Outdated server! I'm still on 1.12.2
Shouldn't I be able to connect to that server?
marble kelpBOT
what plugin did you install to translate the protocol?
rapid dock
WTF
naive pecan
Hi, we've got a 1.14 waterfall proxy setup and we we're wondering how to tell it to allow 1.8 .ect clients through as we have viaversion .ect installed on the servers behind Waterfall, any suggestions?
strange brook
viaversion doesn't do backwards compat
naive pecan
Nvm got it working ❤
tidal musk
waterfall supports 1.8-1.14.4....
rose badger
If bungee/waterfall is on the same machine as the spigot backends, all I have to do security wise is make their ips be 127.0.0.1? I don't get how that protects them
hallow veldt
coz its can't be accessed outside your localhost
Only accessed thru bungee and thats what u want
rose badger
Ok so they need separate ports
I should tell bungee to run on 25565 correct? and the servers on slight variations
hallow veldt
yup!
rose badger
awesome thank you
rose badger
What do i set network-compression-threshold to on the spigot servers?
halcyon oar
-1 i think the guide says that.. https://www.spigotmc.org/threads/guide-server-optimization⚡.283181/
prisma whale
*if its in the same datacenter as the proxy
hallow veldt
I used to always have it on 512 but I recently changed it back to the MC default and it seems to be a lil better on my users with bad internet
rose badger
Is there a way to run a command on 2 separate servers connected via waterfall at the same time?
Like I want to give a player a rank on both servers, I only have to type it once instead of going into both servers individually
marble kelpBOT
I mean, in part sounds like you need a proper permissions setup
But there are plugins to execute commands on multiple servers at once, no idea what I'd recommend there however, overall as a solution it is a pretty crappy one
:shrug:
rose badger
The permissions are generally separate for the servers but there is one rank I manually add to players when they do something. I guess it isn't a huge deal
left basin
i kind of need help to make a waterfall server
prisma whale
Are there any recommendations for starting flags for waterfall? or are they the same as the ones for paper?
lime breach
What's a normal CPU usage for waterfall
it's very high i don't know if that's normal
prisma whale
Mine sits at 0% with 1 player
lime breach
If someone can help me. Does a timings for Waterfall exists? I really Need to find out why Is CPU so high
reef fulcrum
No, use a profiler
lime breach
But im not on a VPS hosting. Thats Multicraft
I know like NetBeans profiler JProfiler but i dont know if i can do that from a Multicraft host
reef fulcrum
Nope, being on a host means that your options for debugging stuff like that is severely limited
spiral coral
Anyone know any good queue plugin for 1.14.4?
tidal musk
another anarchy server, huh?
lime breach
I don't even understand how the queue plugins that 2b2t has works. Is it a plug-in or they modified their server
tidal musk
either they modified their server or implemented a server from scratch
if latter then it can be easily embedded into plugin
for example if it's based on MCProtocolLib (that's what i've used before to create limbo servers) then it's pretty easy
tidal musk
For me, creating a limbo server using MCProtocolLib was far easier and the final code was pretty small.
tidal musk
yeah, i used to embed TyphoonLimbo inside a java plugin and launch it as subprocess
but as typhoon ended up being quite buggy then i dropped that
lime breach
I think that maybe could be also achievable with just a plug-in . Although it may be a little complex
tidal musk
rendering a world to the player is complex, yes
with mcprotocollib atleast
i haven't managed to do that yet (nor had any time to do that)
I personally haven't even tried to do that as I was fine with having a limbo.
glossy yacht
Hi.
There is a very high chance I will be buying 4 servers tomorrow, 3 servers and 1 waterfall.
Can someone tell me how to link them all together, and @ me after?
I have no idea what I'm doing and could really use help with this.
strange brook
"I have no idea what I'm doing but imma drop money on a shitload of hardware. WCGW?"
unreal stag
but what if it's shared hosting?!?
gray fossil
@glossy yacht maybe i
junior talon
whats the perm node to connect to a server also ranks wont work on bungee luck perms
Please I need help with it
rain pier
How often should i restart my proxy?
(i am hosting my proxy on a vps)
marble kelpBOT
I mean, I've never restarted a proxy for any reason other than updates, etc
rain pier
It's alright then but i am having a small problem
Like today a bot wave hit my server and i had no problems what so ever except
everything connected to the proxy had like a crazy ping spike
Keep in mind bots didn't connect to the server at all, sentrybot blocked them
botsentry*
marble kelpBOT
The proxy still has to deal with the process of opening the connection, etc
rain pier
Is there anyway i can combat this?
marble kelpBOT
Load balancing
rain pier
I'll do some research about this, don't want to waste anybody's time lol
marble kelpBOT
Basically, hosting multiple proxies, so you don't have just one instance taking the brunt of the attack
rain pier
I understand, thank you very much
glossy yacht
Hai, I can't use any admin commands for my proxy like /alert
but I have this set up:
raimuakuna:
- admin```pls help me
It says "You do not have permission to execute this command!"
h
glossy yacht
I can't even use /server
glossy yacht
----- haha yoink I got it
weary sapphire
When there is a 1.8.8 player with forge in my server and someone with 1.7.10 joins the server, he crashes. I'm using latest version of Travertine and 1.7.10 paperspigot
tidal musk
can't tell much without server, proxy and client logs
reef fulcrum
protocol hacks are unsupported, adding forge to the mix only makes it much more "fun"
round ocean
1.8.8 is dead, 1.7.10 is a rise of dead.
tidal musk
lets go back to beta
trail portal
whats the diffrence between paper/waterfall?
reef fulcrum
paper is a server, waterfall is a proxy (Bungeecord fork)
tidal musk
search for the messages file inside the .jar file
inner wagon
Players keep getting kicked from the server back to the hub randomly, and get the bungee message: 'The server you were previously on went down, you were moved to the fallback server.'
The server doesnt shut down or anything.. any idea how to fix?
reef fulcrum
Check your log files, there is generally a message why somewhere
inner wagon
reef fulcrum
something sent a bad packet
inner wagon
that something is bungee plugin or spigot?
reef fulcrum
spigot
reef fulcrum
Define "can't use"
reef fulcrum
Do you have a permission plugin installed?
reef fulcrum
if you have a permission plugin installed on the proxy, that file is ignored
glossy yacht
oh.
LOL IM SO DUMB
so can I just remove the bungee one?
or am I required to keep them on Spigot & bungee?
glossy yacht
my perms plugin is LuckPerms
reef fulcrum
if you're wondering if you should have LP on the proxy, that is the better way, a proper permission plugin works better than the config stuff
glossy yacht
do you know if I need them installed on the proxy & spigot
or can I just use it on Spigot
reef fulcrum
That's basically up to you, given that you're running multiple servers, it often makes sense to share permissions across the board so you're not having to reconfigure several things
glossy yacht
ye, I already use MySQL on Spigot to sync the 3
I just want to know if I should or should not
All, including the proxy, are connected to LP's MySQL
reef fulcrum
Basically, I recommend "yes", but you don't have to
inner wagon
Any idea on how to allow only internal traffic for bungee? I tried
iptables -I INPUT ! -s $BUNGEE_IP -p tcp --dport $SERVER_PORT -j DROP
and it did block the outside connections for the sub servers ports (25566, 25567, etc) but it also blocked the internal connection so i cant join the servers through bungee as well
prisma whale
Lookup bungee cord setup. Pretty much the same thing
keen shoal
so its basically same setup`?
keen shoal
nvm have found out
weary sapphire
@reef fulcrum are protocol hacks are unsupported? WTF your team made the project with protocol hack 1.7.10 - 1.14.4
strange brook
- no need to ping him
- protocol hacks are things like viaversion which allow clients and servers of different types to connect
the proxy supporting 1.7.10 means that is supports a 1.7.10 client connecting to a 1.7.10 server
that's what proxies do - they pass through traffic
changing the packets between MC protocol versions is an entirely separate issue and up to plugins like VV/Rewind/Backwards etc
tidal musk
tada
marble kelpBOT
Project Travertine build #109: SUCCESS in 1 min 36 sec: https://papermc.io/ci/job/Travertine/109/
limber jacinth
I NEED HELP!
Whenever i (the owner/has op) joins my waterfall connection that is on 1.14.4, everyone gets kicked
tidal musk
can't help without logs and more information
we aren't wizards who can remotely debug your stuff via tele-whatever ;)
weary sapphire
wizjany if a client 1.7.10 try to join in a 1.7.10 server if crashes
Also without plugins
tidal musk
logs...?
lone void
getting blvxr has disconnected, reason: Invalid payload REGISTER! when connecting using DDSS modpack client but doesn't happen when using Omnifactory modded client
Modded Client > Waterfall proxy > Modded Server
I've seen multiple server networks with this set up so surely there is a way around this payload issue?
https://gist.github.com/craving-death/986690f29a12eda4ed4447ec3d324f5f
full cypress
Cause I believe most of them user Tavertine. lol
@lone void are you connecting to a paper server with the modded client?
if so read here for your issue. https://github.com/PaperMC/Paper/issues/2063
sour mortar
Can this be removed? http://prntscr.com/pi1z8e and how?
reef fulcrum
config.yml
r3xet "When a 1.8.8 player..." "in a 1.7.10 server, he crashes"; Can you actually replicate that issue without having 1.8.8 players in there? if no, that's not something on our end
molten glade
what is waterfall
reef fulcrum
I don't care about video, I want logs
weary sapphire
[22:10:25 INFO]: R3XET lost connection: Internal Exception: net.minecraft.util.io.netty.handler.codec.EncoderException: java.lang.RuntimeException: An internal error occured.
Log
weary sapphire
It's the full log of the error
reef fulcrum
As I said, enable debug mode...
weary sapphire
i have did it
reef fulcrum
And when you do, and reproduce it, provide the full log file
reef fulcrum
the full log
weary sapphire
I encounter this error also without plugins
reef fulcrum
That's not what I asked
reef fulcrum
It makes it much easier to look at stuff when you have the full picture
if you're going to make this much more of a chore than it needs to be, you're on your own
my head is killing me
reef fulcrum
Are you actually running waterfall? what version?
coarse gale
why do you need the information that is required for you determine the error in order to help me
junior talon
I need help with adding new worlds and changing some servers to different versions
Anyone?
tidal musk
wrong channel my dude
prisma whale
I read it though, stop being dumb and give the log if you want help
coarse gale
you only provided partial logs, not the entire thing
how are we supposed to get the full context and figure out what's happening?
toxic quiver
is it possible to have a paper 1.14 + 1.13 server respectively along with a FTB server running side by side and waterfall them all together?
reef fulcrum
Using waterfall or potentially velocity is your best bet
Proxies and forge tend to be a bit of a icky topic, as for how it all actually plays in terms of players being able to move over to the modded server is also fun, as some mod packs won't connect to vanilla at all, and others will blow up in odd/spectular ways sometimes
toxic quiver
i see
it would be RLcraft modpack
if yo uhave any experience with that particular one
reef fulcrum
No experience with it
lean zenith
"player disconnected with: Kicked whilst connecting to Hub: Outdated server! I'm still on 1.13.2"
Anyway to actually fix this, using viaversion with its addons
It only occurs if my proxy suddenly crashes (rare chance) and im needing to stop the proxy and the hubs completely, and restart them back up.
tropic isle
at net.md_5.bungee.entitymap.EntityMap.rewriteMetaVarInt(EntityMap.java:287) ~[bungeecord.jar:git:Travertine-Bootstrap:1.14-SNAPSHOT:99bd8a8:109]
at net.md_5.bungee.entitymap.EntityMap.rewriteMetaVarInt(EntityMap.java:139) ~[bungeecord.jar:git:Travertine-Bootstrap:1.14-SNAPSHOT:99bd8a8:109]
at net.md_5.bungee.entitymap.EntityMap_1_12_1.rewriteClientbound(EntityMap_1_12_1.java:151) ~[bungeecord.jar:git:Travertine-Bootstrap:1.14-SNAPSHOT:99bd8a8:109]```
Still receiving this error with our SkyFactory4 server even after setting `disable_entity_metadata_rewrite=true` in the waterfall config?@lean zenith Why do you need to stop the hub? isn't restarting the proxy enough?
also that message is probably just from the fact that the server isn't fully started yet
paper wont allow connections until after it's done and that message is used when disconnecting connection attempts before that iirc
tropic isle
Ah, I think ViaVersion causes issues with the metadata rewrite fix
I still had it running but after me removing it, it seems to work okay as of so far
austere solstice
Whats that? http://prntscr.com/piqry2
inner wagon
[Server] ERROR POI data mismatch: never registered at BlockPosition{x=977, y=77, z=1025}
Whats that?
subtle cedar
Hello, im trying to use waterfall and im not quite sure how to set it up properly, im pretty sure i have done everything right but nothing seems to be working properly, any help?
lean zenith
@acoustic adder Thats whats weird, I've thought restarting the hubs is the way to fix, but its either the proxy or the plugin causing the issue.
I just have to be lucky for the proxy to start up first then proceed to start up my hub, or my hub to start up first, then the proxy to avoid the issue.
tidal musk
does waterfall have a built in /announce ?
lean zenith
I does /alert
tidal musk
Can someone try doing /announce on their server and tell me if it works please
marble kelpBOT
with which plugin?
tidal musk
Waterfall/travertine
marble kelpBOT
that's not a plugin
tidal musk
i know. i was wondering if it had an announce feature in it @acoustic adder
marble kelpBOT
the command is /alert
native panther
Is there a 1.12 Forge mod version of IPWhitelist? My google-fu power has only lead me to a 1.8 edition.
hard trench
Use a firewall its safer
native panther
Unfortunately it's a shared hosting environment.
marble kelpBOT
good luck with that
native panther
Make BungeeGuard into a mod 😅
I'll hop into Sponge and see if I can get any better suggestions.
gloomy sundial
is there a perm or plugin with a way to make it so certain ranks can bypass the player limit
loud pivot
essentials.joinfullserver
gloomy sundial
ah k thx
and then just set the player limit in the paper server with everything else to unlimited in the waterfall one?
glossy yacht
How can I get my server to display the 2 people currently online?
(pls ping me with an answer)
reef fulcrum
You'd need to disable passthrough in the config, or use a plugin
tidal musk
use ServerListPlus for example
hexed gale
if i were to move to waterfall from normal bungee would i have any issues with my bungee plugins?
reef fulcrum
Bungee plugins should work fine
Ofc, you should generally test stuff before pushing it to production
hexed gale
of course
cold notch
what are causing this ? out of memory?
https://pastebin.com/6EKFJr42
reef fulcrum
Your OS ran out of memory
cold notch
@reef fulcrum i have 16GB 10GB for server and 1GB for bungee
so is the OS or bungee?
tidal musk
the OS
reef fulcrum
Might actually just be direct memory specifically, Might need to increase that, but some level of o.O as to how you'd actually manage to pull that one off
cold notch
seem like that msg started after
-> InitialHandler - bad packet ID, are mods in use!? Did not read all bytes from packet class net.md_5.bungee.protocol.packet.Handshake 0 Protocol HANDSHAKE Direction TO_SERVER
reef fulcrum
.g enable netty leak detector
marble kelpBOT
(DiscordBot) https://netty.io/4.0/api/io/netty/util/ResourceLeakDetector.Level.html -- ResourceLeakDetector.Level (Netty API Reference...: "Enables paranoid resource leak detection which reports where the leaked object was accessed recently, at the cost of the highest possible overhead..."
reef fulcrum
urgh, there is a system property to enable the leak detector
Something is either leaking or those clients are somehow sending enough data to flood the buffer
(Either that or you have too many worker threads, does bungee let you configure that?)
cold notch
idk but it been running fine fore like 10+days. and recently this morning (10) InitialHandler - bad packet ID start spamming as in the picture around 30 InitialHandler - bad packet ID
and then the error that i send https://pastebin.com/6EKFJr42 start spamming
because overnight there was no player online so bungee is prob using like 128 or so 5% ram
so ithink they flooded the buffer or smt but is there a way to fix/block or improv? idk haha
agile lotus
Hello,
Since we went on Waterfall we have no more "server-icon"!
How to have our server icon?
Thank you in advance.
tidal musk
huh
that logic isn't changed in waterfall
so if it worked before with bungeecord then it continues to work with waterfall
agile lotus
Okey I wonder if the dns redirect change can also affect the "server-icon" because we have changed host.
reef fulcrum
No, you just need to throw the server icon in whatever is resolving the ping
if you have ping-passthrough, on the server; if not, in the proxy
cold notch
can i use this for bungee too?
-XX:+UseG1GC -XX:+UnlockExperimentalVMOptions -XX:MaxGCPauseMillis=100 -XX:+DisableExplicitGC -XX:TargetSurvivorRatio=90 -XX:G1NewSizePercent=50 -XX:G1MaxNewSizePercent=80 -XX:G1MixedGCLiveThresholdPercent=35 -XX:+AlwaysPreTouch -XX:+ParallelRefProcEnabled -Dusing.aikars.flags=mcflags.emc.gs
marble kelpBOT
Why on earth you want to give Waterfall 2 GB of ram?
How many users do you actually have?
tidal musk
64gb of wam to proxy or bust
cold notch
i was testing smt but i give it 512 or 1gb
screen -S BC java -Xmx1024M -Xms1024M -XX:+UseG1GC -XX:+UnlockExperimentalVMOptions -XX:MaxGCPauseMillis=100 -XX:+DisableExplicitGC -XX:TargetSurvivorRatio=90 -XX:G1NewSizePercent=50 -XX:G1MaxNewSizePercent=80 -XX:G1MixedGCLiveThresholdPercent=35 -XX:+AlwaysPreTouch -XX:+ParallelRefProcEnabled -Dusing.aikars.flags=mcflags.emc.gs -jar BungeeCord_1.3.1_20191013.jar
this work for bungee or no?
marble kelpBOT
I don't think server flags will be of any real benefit for the proxy
cold notch
i dunno which flag that allocate all 1gb for the proxy because when i oonly use screen -S BC java -Xmx1024M -Xms1024M -jar BungeeCord_1.3.1_20191013.jar
it allocate 1gb yes but when looking in taskmng it say it use around 128MB but when i use that flag it use all 1GB so i wanna know which flag give the jvm the full ram
marble kelpBOT
-Xms instructs the jvm to allocate the provided amount on startup
oblique adder
should i be worried? https://hastebin.com/exixoxaboj.sql
tidal musk
No
Or at least I don't think so
I don't know much about Java but I think it says that it can't translate the domain into an IP address; Try changing your DNS settings.
oblique adder
Alright so should I be setting ip forwarding to true?
marble kelpBOT
that has nothing to do with the ip-forwarding setting but yes, you should always have that on true
austere solstice
hm i get that team already exist error in bungee but i dont know how to fix it ._.
Team team = b.getScoreboard().getTeam("xy");
if (team == null) {
team = b.getScoreboard().registerNewTeam("xy");
team.setPrefix(prefix[1].split(":")[0]);
team.setColor(ChatColor.valueOf(prefix[1].split(":")[1]));
}
team.addEntry(players.getName());
What is wrong with that check?
reef fulcrum
Servers can't share teamnames, basically
austere solstice
all my servers has that method for setting the team
reef fulcrum
Bungee basically has a bug
If the first server you're connected to has a scoreboard with the same name as the new server, it will get upset
I have ideas on how to address it, but it will create some other issues, which is kinda bleh
austere solstice
and how can i avoid that?
reef fulcrum
Make sure that your servers use different team names
austere solstice
not possible to remove the scoreboard from the proxied player?
at serverswitchevent
reef fulcrum
Yea, you could do that
austere solstice
Is ServerSwitchEvent the first called event if the player successfully changed?
reef fulcrum
Wait, might be too late
Erm
Basically, you need to clear the entries out before the new server starts sending team data
austere solstice
Which event would you suggest?
reef fulcrum
ServerConnectEvent, I wanna say
Or maybe that's too early
hmph
my brain is just rip right now, stupid meds I'm on \o/
austere solstice
okay i will try with server connect event
ProxiedPlayer p = e.getPlayer();
p.getScoreboard().clear();
Is that right? xD Didnt saw how to set a emtpy scoreboard
reef fulcrum
No idea if that will clean them on the client or not, would fix it bungee side assuming that isn't a copy
austere solstice
That fixed not the error xD
aiaiai
Sometimes i can join but the most time the error comes
i will simply try to add the servername to the team name
ah now the names are to long
Hm, now i have different names and the error is still there
oh why is that happening 
(I dont have that problem with all my servers, just two)
Have no idea 😦
Would it not be fine if you add a check for illegalargumentexception and let happen nothing?
reef fulcrum
No
We're trying to avoid a state desync
Bungee just does it horribly
Issue is that fixes around that involve behavioral changes somewhere
austere solstice
But why do i have it on just 2 of 8 server, its the same plugin everywhere 😂
reef fulcrum
Wonder if there is some level of race condition on it 🤷
austere solstice
I will open an issue to md5, hopefully he solve it 🙄 Thanks to you 🙂
reef fulcrum
There is already an issue open on that one
Also, make sure that you're clearing up old scoreboard teams and stuff
austere solstice
I deleted the world and after the first join i have that error 😄
marble kelpBOT
Optimized & Recommended JVM Startup Flags: https://mcflags.emc.gs
oblique adder
should i be using these for waterfall
prisma whale
I do, except for the XMS, XMX ones, I don't think its necessary though since its just a proxy server
for the XMS XMX I just use 1G, but you can do with much less really
oblique adder
I do 512 and thats honestly wayyy more
oblique adder
How does one go about having the max players of a bungee show 1 above the current total players on
Ive seen it done before and want to emulate that but cant find a config for it.
marble kelpBOT
above? like when you hover over it?
oblique adder
Like if there is 20 players online the max players on the server list will be 21, if there is 80 then the max would be 81. Know any way to accomplish this
marble kelpBOT
Just listen to the ping event and adjust the result accordingly
marble kelpBOT
@Chasewhip if you didn't want to program something yourself and are still searching, here is a very simple plugin that I made a while ago that can do that: https://www.spigotmc.org/resources/dynamicslots.40598/
tiny bear
How do I set it to where players when they connect they are force to join a certain server
Because as of now, when players connect to the server. They will connect to the waterfall and get sent to server 3 instead of server 1
upbeat ember
forced_hosts in config.yml?
reef fulcrum
Players will join the first priority, assuming you have force default server true
reef fulcrum
If you want them to join the lobby, yes
tiny bear
force_default_server: false
Welp
I think this should have been turned to true
But wait, in the force_hosts do you have to specify a port?
reef fulcrum
It just matches the hostname that the client used to connect
tiny bear
so leaving it as ipaddress.org: lobby good?
reef fulcrum
99% of the time you don't even need to touch the forced host section
(Majority of servers do not use that feature)
serene tide
There's a config setting in the paper.yml which can disable them
reef fulcrum
disabled by default in the latest builds of paper
172 is not the latest build of paper...
prisma whale
?? Just replace the paper jar with the latest
strange brook
waterfall
prisma whale
I mean is that not the process? I run waterfall as well. It's that easy for me.
reef fulcrum
looks at the topic of the last discussion in here
strange brook
yea, was just commenting that he was asking about paper in the waterfall channel
tidal musk
but they're the same thing????
tidal musk
cc tux
prisma whale
shame it wasn't something book related, since a book has multiple papers
upbeat wave
sweet jewel
Is there any documentation available ? I'm REALLY new to this domain and I have no clue
I have 2 spongeforge servers that I'm trying to link with waterfall
and it's necessary because we want to recrut builders in our server and this functionnality seems to be precious (separation)
sweet jewel
Should I enable ip-forwarding everywhere or it's just a option ?
full cypress
Absolutely enable it.
reef fulcrum
ip forwarding is enabled on the proxy, that passes UUID+IPs through to the backend server properly
Overall advice is to follow the bungeecord installation steps
full cypress
on your waterfall servers, set ip-forward: true
on your other servers, set online mode to false and enable bungeecord: true in spigot.yml
reef fulcrum
For spongeforge, they have their own setting for bungee mode in one of their config files
full cypress
Then whatever that specific setting is. I have never used SpongeForge. Last platform I used was Cauldron.
sweet jewel
I want them to be still able to connect directly on my sponge server instead of the bungeecord ip
it's enabled for me
got this : "If you wish to use IP forwarding, please enable it in your BungeeCord config as well !"
in config of waterfall, online-mode and ip-forward are true, but in my server, offline-mode is true and ip-forwarding=true and module bungeecord=true are enabled in sponge config
reef fulcrum
You basically can't do that
The proxy needs to handle authentication, the only way around that would involve modifying the server itself to be able to be in online mode but also allow the proxy to handle the authentication and bypass the servers mechanism, which creates its own challenges
marble kelpBOT
why do you want that in the first place? just let them join through waterfall?
tacit garden
ERROR StatusLogger Log4j2 could not find a logging implementation. Please add log4j-core to the classpath. Using SimpleLogger to log to the console...
any idea?
error explains itself but is this a waterfall issue
looks like it was a bad upload 🤷
dont understand how tcp uploads work if this occurs
winged orchid
Latest version of jessica client can crash any bungeecord/waterfall server
Specifically, its " .figure2 " command
Even with packet limiters on the spigot servers, the bungeecord server just instantly crashes.
Tried it on multiple networks, including mine, and it works flawlessly
prisma whale
any? like 1.14.4?
winged orchid
It justs restarts
Nothing at all shows up in the logs
Anybody that is on the same bungee just crashes with "the server you were previously on went down"
I use OpenJ9 11 64bit
marble kelpBOT
so something most people don't use 
winged orchid
Yes indeed
I'll try with java 8
But I do not think the java version is related to this crash
Only crashes with jessica client's .figure2
Has to do with shulker box packets
When I blocked them from spigot, it did not crash anymore
Seems like jessica sends an item packet containing a shulker box with strange books, with too many pages or pages too long
marble kelpBOT
probably just too much data
winged orchid
Shouldn't it be there something to prevent stuff like this in waterfall?
Because the spigot server is fine
only bungee crashes
Only players on the same bungeecord server as the player that did the ".figure2" stuff crash
Players on other bungeecords but that are on the same spigot server are fine.
They crash because that specific bungeecord server restarts
I opened a ticket anyways
reef fulcrum
I'm currently kinda ill, but like, nothing really useful there
(Can't say I'm super interested in tryna find this fabled client)
winged orchid
I can give it to you
I just don't want to post it here
Because people may want to use it in a bad way
reef fulcrum
i guess, no idea what the best way to send it is
winged orchid
dms
signal prism
Hi, I recently got an attack on my server, which sent my server offline for a few minutes, this was the error that I received, I was wondering, if there is a way to figure out how to fix it... https://pastebin.com/kWjWzvsR
full cypress
That's not an attack.
Thats a bad plugin.
Unless your server just received over 1,000,000 connections in like less than 20 seconds, that shouldn't happen.
signal prism
He never gave me this error, anyway how do I understand what is the plugin?
This is a part of the log https://pastebin.com/JdX9chTu
full cypress
So "Too many open files" is exactly what it says. On most linux systems as root a single process can open 500,000 files at the same time. for unprivileged users this is ~4096. Network connections, like a player connecting, also open a file. So you'd have to either watch what files are being opened in your server directory with inotifywait, or check the number of open sockets to your server. You'd see this as inithandlers in your log file
Ya, someone hates you.
signal prism
can it be a mysql request?
full cypress
Nah, this is a malicious player connecting from several IP's within a single minute.
signal prism
ok, so there are no solutions?
full cypress
So the only real way to fight this is with more bungee's and a round robin DNS.
Or to take it one step further, a connection throttle needs to be configured inside bungee or through a plugin to stop more than 3 connection attempts a minute.
from the same IP
marble kelpBOT
or go a step further: use fail2ban to blacklist the ips in your firewall xD
full cypress
^ fail2ban might actually catch that. lmfao
marble kelpBOT
a nice sideeffect: it gives these idiots the illusion that they actually managed to take the server down, lol
full cypress
Dude. I'm impressed, I wrote a plugin for that. that checked ip flooding and username flooding
eg 1 username to 5 ip's within set timeframe = block
or 5 usernames to 1 ip within set timeframe = block
marble kelpBOT
well a plugin could potentially be better suited if they connect from different IPs, that fail2ban config only really limits connection attempts from a single IP
but seeing as you want know the name when the first connection is made (which can already be used for ddosing) a filter by ip is usually quite usefull
in *wont
-in
can't type today, lol
reef fulcrum
Open handles cont to sockets too
Basically, something opened so many connections that it hit your resource limit
signal prism
so, do you suggest to ban the ips dangerous from iptables?
reef fulcrum
Basically, yes; Using your firewall to limit this kinda stuffis the best option, as well as ensuring that your resource limits are configured for your needs
signal prism
thanks for help
steep tundra
How can you make Waterfall compatible with 1.7 and 1.8 only?
marble kelpBOT
use travertine and then limit the allowed protocols
steep tundra
Is there a config option for it?
tidal musk
?
haughty nimbus
is this fixed on waterfall?
https://github.com/SpigotMC/BungeeCord/pull/1557
tidal musk
using travertine, when a 1.7 player is on the server and a 1.8 player joins, it kicks the 1.7 player. i was told i have to change some of the connection handlers or something to fix this. anyone have any ideas?
Ping me with answer please 🙂
lone void
does waterfall automatically put players from a server back to the lobby or do i need a plugin for that? - if so which?
*Ping if response pls ❤
tidal musk
explain your question
sour mortar
Why do I and my staff get kick?
.... [11:27:11 WARN]: [/IP:PORT|Kapakulia] <-> DownstreamBridge <-> [Factions] - bad packet ID, are mods in use!? Empty minecraft packet!
.... [11:27:11 INFO]: [Kapakulia] disconnected with: The server you were previously on went down, you have been connected to a fallback server
tidal musk
update waterfall
tidal musk
using travertine, when a 1.7 player is on the server and a 1.8 player joins, it kicks the 1.7 player. i was told i have to change some of the connection handlers or something to fix this. anyone have any ideas?
Ping me with answer please 🙂
tidal musk
who told you and what
we can't tell you answer if we don't know what we should change exactly either
reef fulcrum
If you can prove that it is infact travertine causing the issue, we can look into it; But like, without any information, and the whole fact that protocol hacks aren't supported; 🤷
tidal musk
Is it travertine. It says [Proxy] lost connection. Works with every other bungeecord
@tidal musk @reef fulcrum
reef fulcrum
Please don't ping me
That also just says that the connection was closed
We'd need to know why it was closed, e.g. client or server logs detailing the issue
tidal musk
don't ping me either please
I can provide logs
And I apologize
then do so
heres the travertine log: https://pastebin.com/u1neN49x
this is the kick message: https://gyazo.com/3b3768d9bfa7de1ccc01a732675ecf44
reef fulcrum
10:06 PM] electroniccat: That also just says that the connection was closed
[10:07 PM] electroniccat: We'd need to know why it was closed, e.g. client or server logs detailing the issue
also, it has to do with forge. not vanilla
electroniccat can you guide me where to find that info please
reef fulcrum
Considering that the server was the one that booted the connection, probably there
tidal musk
it wasnt the server... one of the accounts was on 1.7 (lunar client) on "Practice" and the 1.8 forge client connected to "Hub" and it kicked the 1.7 player.
its not only me thats having this issue there are several more people with the same thing
reef fulcrum
"Lost connection to server"
That means that the connection was closed between the proxy and the server
Nothing more, nothing less; Beyond that, we literally go back to the "We'd need to know why it was closed"
tidal musk
any ideas on how to find that issue... i dont know why it was closed because i didnt develop this stuff man ://
im trying my best to provide as many details as possible
reef fulcrum
You've yet to provide server logs
tidal musk
i will do that now.
here is the "Practice" log, this is the one which the player gets kicked from https://pastebin.com/Uc5YnnTy
reef fulcrum
[18:52:57] [Server thread/INFO]: SachaPvP lost connection: Internal Exception: net.minecraft.util.io.netty.handler.codec.EncoderException: java.lang.RuntimeException: An internal error occured.
You'd need to enable debug in server.properties
reef fulcrum
No, that would hopefully give you more info on what the issue is
But, at this point, it's basically not our issue
tidal musk
but it works fine without travertine...
reef fulcrum
You mean, a direct connection to the server works fine in these cases?
reef fulcrum
as in, connecting directly to the server without a proxy at all
You said that it worked fine? Why do you need to test it?
tidal musk
i dont know if its possible though due to it being on offline mode
i meant that it worked fine without travertine with a different bungeecord jar
reef fulcrum
You'd need to disable bungee mode to be able to test that properly and pull it from the network or test it
What different bungeecord jar?
tidal musk
a default 1.8 bungeecord jar
reef fulcrum
Did bungeecord 1.8 support 1.7? I honestly don't remember that far back
tidal musk
yes it did
reef fulcrum
Basically, at this point, you'd need to work out what it's doing that is causing the issue; Potentially worth disabling the entity metadata remapping in the waterfall config, but given in mind that the server software you're using is something that we can't even look at the code for and I'm far too ill to waste my time tryna see if I can replicate it somehow, your support here is going to be limited
tidal musk
alright ill see what i can find.im really sorry to hear your ill, i hope you feel better soon!
lone void
How do I get rid of the red X and "incompatible FML modded server" (i run a modded network)
[ping if u respond pls ❤]
tidal musk
not enough information...
what mods does client have & what mods does the server have?
fair halo
Hello, Good evening, I discovered a problem with Waterfall on the last update, From the 9th server the waterfall refuses to link the servers, Yet the ports are well released which is very weird, I tried with the same configuration on Bungeecord and I had no problem, I wanted to let you know so you can fix that.
fair halo
In fact, when I do /server spleef to connect me to the spleef it sends me back to the lobby server or it tells me I'm already connected to this server off that it's wrong
tidal musk
logs?
fair halo
no sorry :/
Then no erraters were generated when I tried to connect to the spleef server
reef fulcrum
Shamefully, our magical ball kinda smashed and no longer works, so we'd need actual info/logs showing what's going on; I find it pretty hard to see that it would tell you that you're already connected to a server when you're not unless there is a plugin interferring or something was misconfigured
manic agate
Can anyone advise a good anti bot plugin? I've found a few like L2S antibot, Bot-Sentry AntiBotDeluxe. I don't see much difference at first sight. It's for a WF setup in online mode. Attacks are probably strange packets/pings or something like that.
rapid dock
hey,after the #291 update,the Windows Console can not appear the color text because of log4j update,how can fix it?
reef fulcrum
Update from what?
There was a recent fix to windows being stupid with log4j which was confirmed to fix some issues
Oh, I thought you was talking about paper, I'll try and pull the fix for that over
marble kelpBOT
Project Travertine build #110: SUCCESS in 1 min 39 sec: https://papermc.io/ci/job/Travertine/110/
Project Waterfall build #296: SUCCESS in 3 min 55 sec: https://papermc.io/ci/job/Waterfall/296/
rapid dock
The new release works!Really thanks to fix it! @reef fulcrum
tidal musk
electroniccat
remember that issue we were discussing yesterday
i found the fix
this is the code that needs to be implented
You have to make 2 changes:
UserConnection.java -> https://pastebin.com/Uh0yhnmJ
ServerConnectior.java -> https://pastebin.com/4hEduV3N
(into travertine)
reef fulcrum
You gotta highlight the changes, and explain why
Oh, your fix is literally to strip out the FML support
tidal musk
well without it players using multiple versions cant connect...
reef fulcrum
I don't care specifically about oddball versions
tidal musk
well now that you understand my issue better, is there a better way to fix this?
reef fulcrum
No, because I don't see the actual exception being caused on the server to actually see what is going on
tidal musk
its a problem with forge
reef fulcrum
I can't even begin to try looking into the issue without knowing the actual exception of what is going on
tidal musk
wdym by actual exception
reef fulcrum
The exception that is causing the server to boot the connection
tidal musk
its whatever is in those two classes that i sent you
reef fulcrum
That's not the exception
Something on the server is handling the data being sent to it improperly
tidal musk
thats forge
reef fulcrum
I need to know what the exception on the server is
Then we can actually look into fixing it and avoid stripping FML support from IP forwarding
tidal musk
cause this issue doesnt occur without forge
reef fulcrum
Your "fix" literally involves removing a feature from the server
Basically, if you can't provide the underlying exception from the server, I can't do anything about it
I'm not going to start pulling features from the proxy for oddball setups
tidal musk
can you tell me where i can go around finding it?
reef fulcrum
Because you're using unsupported server software, I can't
reef fulcrum
Enabling debug mode is often enough, sometimes it requires modifying the server to actually display the exception
mSpigot is 1) Not open source 2) Is stupidly old 3) is entirely unsupported by any community
reef fulcrum
Okay, "is based on a stupidly old version of the server software"
tidal musk
ugh
i just dont know how to tell you its not the server spigots
its travertine causing the issue
reef fulcrum
You're missing the entire point
tidal musk
i could use any spigot and have the same issue
reef fulcrum
The server is blowing up because travertine is sending some data to it
I need to know what the server is blowing up on so it can actually be looked into
reef fulcrum
No
tidal musk
anyway you could make me a version with my fix just for me?
reef fulcrum
Basically, I do not have the health/patience to try to reproduce this issue on my own, if you or somebody can provide info on what is actually blowing up on the server due to the data being sent to the server, we can look into it; But I'm not going to just pull features from the proxy all because some unsupported setup is having issues
tidal musk
i understand
reef fulcrum
Are you using modded servers on your network?
tidal musk
if you mean using plugins by modded, yes
reef fulcrum
plugins are not mods
tidal musk
then no
reef fulcrum
Basically, if you're not actually running a forge server, try disabling forge support in config.yml
reef fulcrum
Oh, that won't actually work, bungee still deals with the handshake data
marble kelpBOT
why not just make it yourself if you already have the code?
tidal musk
i have no idea how to go about it
also, here is the issue with debug on in console: https://pastebin.com/xrNj5m9K
actually im not sure if thats related to travertine
marble kelpBOT
just follow the readme of the repository to get the dev environment setup, edit the classes and the build your version
tidal musk
alright
reef fulcrum
> [20:01:36 WARN]: at me.allen.ziggurat.impl.v1_7.b.write(b.java)
reef fulcrum
That's part of what is leading up to the exception
reef fulcrum
an NPE is being thrown due to an improper named entity packet
I couldn't say exactly what is wrong with the packet as I don't have source for that version of the software
reef fulcrum
Not really my issue
tidal musk
you're basically on your own with this lol
lone void
@tidal musk i run multiple modpacks
And the default server is a vanilla HUB
On RLCraft, no red X and incompatible FML modded server
On Direwolf20:
responding to this btw: https://ptb.discordapp.com/channels/289587909051416579/555462214606389275/637975672534597632
but if anyone can help me get rid of that red X and incompatible message that would b gr8
marble kelpBOT
I doubt there is a good way around that besides making a compatible server the default one and passing the ping through to it (which might make the other modpack incompatible again)
reef fulcrum
^
That's handled on the client
You'd basically need to trick the client into thinking it can connect somehow, which as noted, will probably cause issues elsewhere
marble kelpBOT
the best solution is probably to just use different sub domains to connect to different default servers that support the mod pack, then if someone connects redirect to the hub
inner wagon
is it possible to have a bungeecord 1.8.8 with 1.14.4 spigot?
together with viaversion on the bungee side?
tidal musk
bungeecord 1.8.8?
you do realize that bungeecord and its (up to date) forks generally support from 1.8.x up to 1.14.x?
and viaversion doesn't work like you think it does - it does not add support for older protocols, it does the opposite; making newer clients able to connect to older server
marble kelpBOT
there is ViaBackwards though which in general is even more broken than ViaVersion
rough warren
just use protocol support to allow older clients in 1.14.4 servers
bleak nymph
Hi i got this Error how do i fix it? https://pastebin.com/7CEfauS8
reef fulcrum
your config file is bad
bleak nymph
how do i fix it
reef fulcrum
Well, you busted the config somehow, shamefully we don't have a shiny magic ball to see what you busted
bleak nymph
I fixed it