The list of stuff I'd need to do to get waterfall back into a position where I can maintain it is generally just offset by the fact that the ecosystem is generally much happier supporting bungeecord rather than waterfall, and nobody was willing to aid, and so it's generally just a huge ass investment from me with 0 return
#waterfall-help
reef fulcrum
craggy trout
The list of stuff I'd need to do to get waterfall back into a position where I c...
Could i get a list?
reef fulcrum
I don't have one, it's all generaly just stuff that cropped up in my brain over the years
craggy trout
I would be happy to do it myself and my partner, in the way that would be accepted.
craggy trout
I don't have one, it's all generaly just stuff that cropped up in my brain over ...
If at any time you think of it, send me a message.
reef fulcrum
It's stuff like supporting adventure, wanting to clean up parts of the codebase, investigating certain things that have come and gone
things like migrating to gradle so that I'm not having to fight the broken ass system for lombok crud
craggy trout
I have a fork of Waterfall that includes many improvements such as the use of native java 17, native minimessage by default for all messages, better libraries such as Caffeine, dsutils and others.
craggy trout
things like migrating to gradle so that I'm not having to fight the broken ass s...
I doubt it, have it too.
reef fulcrum
Like, my health has slowly improvied since late 2019, and so, I can finally start and have been touching more stuff over the while, but, waterfall is just a large amount of investment in which I get nothing out of it, and i'm having 'minor' surgery next month so won't be around for 1.21
gradle is generally much nicer to deal with certain aspects, would still need to delombok, but afaik, getting the javadoc stuff working properly would be nicer
craggy trout
gradle is generally much nicer to deal with certain aspects, would still need to...
I think I can make a pull request by switching to gradle, it's not too much of a head breaker.
reef fulcrum
I really did want to just look at adding an intermediatary step to just delombok the source tree before we start patching stuff
Well, the project is in maintainance mode, I'd then need to fight TC
just not sure it's worth the effort
craggy trout
just not sure it's worth the effort
How ever, tag me if you see proper switching to gradle
ruby plover
Ok I think I have the switch from waterfall to velocity pretty well worked out but I have a question about a config. Have an image but since has all the backend server ip's prefer not sending it in a main chat. Can I DM anyone to verify or how should I do this. I want to be sure I have this set up correctly since my fallback plugin is no longer working and it relies on come of this. Thanks I have been working with the other plugin dev also on this too.
craggy trout
marble kelpBOT
pretty sure they already asked there
craggy trout
I don't think he will get any response here about velocity, either isn't the channel for it.
marble kelpBOT
?
tidal musk
Do i have to ip forward for every server that connected?
and how would i send the ip for the waterfall server to friends (where would i find the ip)
ruby plover
Just have to say I do not like Velocity I have had nothing but issues since installing it trying to get things to work. Waterfall was so much nicer
now I can not get fallbackvelocity to work anymore and players have to always join back at lobby and then go to server they want again if get disconnected. I tried velocityreconnect it worked a couple times and then stopped.
reef fulcrum
velocity reconnect just uses the initial server event, if that's failing I'd maybe wonder if you had some other plugin causing issues
ruby plover
tulip surgeBOT
message.txt by @ruby plover: https://pastes.dev/UbVwpvSQVF
tulip surgeBOT
message.txt by @ruby plover: https://pastes.dev/YTUdbo1jzo
ruby plover
ahh sorry Ill use that link next time... what is this example.com stuff
reef fulcrum
forced hosts, it's the exact same concept in waterfall
nothing changes there
if you're not using them, just remove them (make sure to leave the header)
otherwise, it's a hostname > try list override, so, if you connect with kingdom.example.com, you could make it go directly to the kingdom server instead of the gateway
ruby plover
what do they do? Guessing I am not using them if I dont know what there for though lol
ahh so if players come into the server lobby/gateway normally using enter.servername.com then I need to put that there
reef fulcrum
No
generally you just use the try list
forced hosts let you override the general try list
if you don't want to have special entry domains that take you to specific servers intead of the lobby, leave that section blank
ruby plover
ok think I get it now and do not want that
backends are secure can not get into them anyway
so I just use try and then velocityreconnect so they can reconnect to game where they left off if want?
reef fulcrum
yes
ruby plover
ok I will try that again without the forced though not sure they were my main issue
rigid root
So how come Waterfall isn't going to get updated anymore?
reef fulcrum
- I am the only person on the team who was working on it and I was generally dead for the most part of the past few years
- the compensation for working on it generally doesn't justify the effort that would be needed in order to bring it up to scratch
- regarding 1, 1.21 is slated for release around the same time as I'm having a corneal transplant
rigid root
I see. In that case, how easy would it be to convert to bungeecord when 1.21 drops?
reef fulcrum
generally should be trivial unless something you used relied on waterfall api
rigid root
I don't think that I relied on the waterfall api. Would the files be the same if I started a new bungeecord server?
reef fulcrum
I believe so
rigid root
ok
clear spoke
How to have 1.9.4 WaterFall ?
dry thunderBOT
broken lily
How to have 1.9.4 WaterFall ?
^
You won't find any support for outdated verions.
ruby plover
I am not sure where to go with this as I have tried reinstalling and reconfiguring both Discord and Venturachat and have tried many things. This use to work fine does Velocity not allow the bungee: true channels in venture chat to communicate with Discord? The channels I have the bungee: false seem to work fine but the ones that are set to true get these messages. https://pastes.dev/dVQzGdwNpO
or is this something to do with updating to Paper 1.20.4 on the backend servers?
reef fulcrum
velocity implements the bungee messaging channel
as for your error, that's coming from a placeholder, you'd wanna take it up with them
ruby plover
ok thanks
radiant hearth
Just to triple confirm
UpstreamBridge - read timed out = proxy lost connection to player's client?
reef fulcrum
We got no packets from the client in 30 seconds
vestal nymph
We got no packets from the client in 30 seconds
is it no packets or specifically no keep alives?
reef fulcrum
no packets
ruby plover
well I reinstalled discord placeholders and venturachat and have all chats working fine on all the servers except for this one guess I try and reinstall the Prison file or would that be overboard? I know hosting had me reinstall LuckPerms on all the servers when I kept having perm issues to be sure the hooked into Velocity correctly and that seemed to work for that.
pearl valve
How can i disable waterfall?
Because im having some issues and i want to fix it later but i cant connect to any of my servers
craggy trout
Maybe stopping the proxy?
pearl valve
i did
reef fulcrum
stop the proxy, disable bungee mode and re-enable online mode
reef fulcrum
This is the periodic reminder that in 2 weeks I'm going for a corneal transplant and won't be around. ~2 weeks also puts us into the typical time that mojang likes to do their releases for summer. (i.e. assume that I'm not going to be here for 1.21)
ruby plover
Thanks for the heads up and wish you the best with your transplant. Speedy and good recovery.
could you double check this for me to be sure I have things setup right. Some servers seem to fall back but the main server working as lobby just gets a protocol error and I wanted to be sure it is not something I have wrong.
and sometimes when I type in a command from another plugin it just disconnects me and gives the same message in picture image.
craggy trout
This is the periodic reminder that in 2 weeks I'm going for a corneal transplant...
I can pull request 1.21 stuff
reef fulcrum
I have no idea if anybody else will be willing to do the work to release it
winged pilot
who says 1.21 is even coming out in june tho
viscid silo
History
ruby plover
so my config in paste bin is correct? The pastebin sent above a few messages
winged pilot
that looks like a viaversion/backwards error
ruby plover
my connection loss?
hosting told me it was a know issue with paper 1.20.4 and would have to wait for 1.21 to come out. Said that is what the have been seeing mainly with servers getting this on a 1.20.4 server after I sent them my lastest log file a few days ago. Moved on to bigger bugs for now.
reef fulcrum
Network Protocol Error means that the client was upset with something
1.21 isn't magically going to fix your server sending bad data
uncut lynx
Does anyone know why this problem? Users who try to enter are kicked and have to try to enter several times to be able to connect....
uncut lynx
known bungee limitation
Without solution?
reef fulcrum
Correct
it's an inherient flaw with bungee in which a fix for would likely cause plugin compatability issues
solid dust
because waterfall is discontinued should i use bungeecord or waterfall
reef fulcrum
well, I'm likely not going to be here for 1.21, so, you probably wanna look towards bungeecord
solid dust
alright
tidal musk
can anyone help my
lost river
can anyone help my
normally some plugin messing with packets
tidal musk
normally some plugin messing with packets
this server is new, i have no plugins
lost river
this server is new, i have no plugins
no plugins on waterfall and no plugins on the backend server?
also if you start new, try to use velocity instead of waterfall
tidal musk
Waterfall: Bungeeguard, ViaVersion, ViaBackwards
Backend: Bungeeguard, ViaVersion, ViaBackwards
tidal musk
normally some plugin messing with packets
Now I get this message to
lost river
Now I get this message to
how is your plugin setup now?
tidal musk
how is your plugin setup now?
its only ViaVersion and ViaBackwards on the Backend
lost river
its only ViaVersion and ViaBackwards on the Backend
try removing them too
tidal musk
try removing them too
now it work, but i need this 2 plugins
lost river
now it work, but i need this 2 plugins
try it with only viaversion first
also what is the client and backend server version?
tidal musk
try it with only viaversion first
viabackwards make problems
lost river
do you use the correct viabackwards version for your your viaversion version?
lost river
oh, you are using /reload ? try restarting the server completely
tidal musk
oh, you are using /reload ? try restarting the server completely
same message
proud fiber
[15:49:29 WARN]: Event ConnectionInitEvent(remoteAddress=/XX.XX.XXX.XXX:, listener=ListenerInfo(socketAddress=/0.0.0.0:25565, motd=Another Bungee server, maxPlayers=100, tabListSize=60, serverPriority=[lobby], forceDefault=true, forcedHosts={pvp.md-5.net=pvp}, tabListType=GLOBAL_PING, setLocalAddress=true, pingPassthrough=false, queryPort=25565, queryEnabled=false, proxyProtocol=false), isCancelled=false) took 1,101ms to process!
[15:49:30 INFO]: [/XX.XX.XXX.XXX:] <-> InitialHandler has connected
[15:49:46 WARN]: Error occurred processing connection for Oxigate ConnectTimeoutException : connection timed out after 5000 ms: /XX.XX.XXX.XXX:25566
[15:49:46 INFO]: [Oxigate] disconnected with: Could not connect to a default or fallback server. Incorrectly configured address/port/firewall?
[15:49:47 INFO]: [/XX.XX.XXX.XXX:51763|Oxigate] -> UpstreamBridge has disconnected
any help?
waterfall
but i have same error with velocity
reef fulcrum
you're connecting to the backend server using the public IP address which is not routable back to the machine
proud fiber
tulip surgeBOT
spigot.yml by @proud fiber: https://pastes.dev/w7b1rRPPnc
tulip surgeBOT
server.properties by @proud fiber: https://pastes.dev/MR52AsgTJm
tulip surgeBOT
config.yml by @proud fiber: https://pastes.dev/mZmrTlY5G3
reef fulcrum
you'd need to connect over a private network, which is going to depend on how you're running the server
proud fiber
there is a tutorial on it?
reef fulcrum
depends on if you're using a container/panel platform or not
if not, just use 127.0.0.1 for all of the backend servers
proud fiber
pterodactyl
reef fulcrum
then you would need to see their docs on how to set that up
proud fiber
if not, just use 127.0.0.1 for all of the backend servers
doesnt work
reef fulcrum
well, yea, cos ptero
reef fulcrum
I have no idea what you did
make sure that waterfall is still bound on 0.0.0.0
if you have any issues I'd suggest speaking to the ptero folk
proud fiber
if you have any issues I'd suggest speaking to the ptero folk
alr
grim walrus
if my main and proxy is on the same server do i use 127.0.0.1 or 0.0.0.0? because both didnt work and i get this
this is my like 10th time using waterfall in pterodactyl panel and now it just refuses to work in any way
firewall is allowing it
reef fulcrum
neither of those will work
you need to configure specifically for ptero, see the guide linked above
long jetty
Is waterfall no longer being updated? 🤔
reef fulcrum
I’m likely going to be in the hospital when 1.21 drops
split tiger
if my main and proxy is on the same server do i use 127.0.0.1 or 0.0.0.0? becaus...
you use your public ip for your proxy and pteros 172.18.01 for all of your backends
split tiger
I’m likely going to be in the hospital when 1.21 drops
same (edit: I forgot to uncheck the ping, sorry)
split tiger
Is waterfall no longer being updated? 🤔
Waterfall has reached end of life! It is no longer maintained or supported.
long jetty
`Waterfall has reached end of life! It is no longer maintained or supported.`
How so?
fair arch
No one will provide updates for it anymore
Whatever waterfall you have on the server should be considered as the last jar you will ever download
crystal lantern
craggy trout
`Waterfall has reached end of life! It is no longer maintained or supported.`
Just wait for upstream PR
Waterfall keeps updated by versions
hidden shore
How do I translate the messages?
craggy trout
What messages?
quick acorn
I’m likely going to be in the hospital when 1.21 drops
It comes out on June 13, will you be gone?
reef fulcrum
Yes
quick acorn
Good luck, everything will be alright!
mortal hull
container@pterodactyl~ java -version
openjdk version "21.0.3" 2024-04-16 LTS
OpenJDK Runtime Environment Temurin-21.0.3+9 (build 21.0.3+9-LTS)
OpenJDK 64-Bit Server VM Temurin-21.0.3+9 (build 21.0.3+9-LTS, mixed mode, sharing)
container@pterodactyl~ java -Xms128M -Xmx20000M -XX:+UseG1GC -XX:+ParallelRefProcEnabled -XX:MaxGCPauseMillis=200 -XX:+UnlockExperimentalVMOptions -XX:+DisableExplicitGC -XX:+AlwaysPreTouch -XX:G1NewSizePercent=30 -XX:G1MaxNewSizePercent=40 -XX:G1HeapRegionSize=8M -XX:G1ReservePercent=20 -XX:G1HeapWastePercent=5 -XX:G1MixedGCCountTarget=4 -XX:InitiatingHeapOccupancyPercent=15 -XX:G1MixedGCLiveThresholdPercent=90 -XX:G1RSetUpdatingPauseTimePercent=5 -XX:SurvivorRatio=32 -XX:+PerfDisableSharedMem -XX:MaxTenuringThreshold=1 -Dusing.aikars.flags=https://mcflags.emc.gs -Daikars.new.flags=true -jar server.jar --nogui
Failed to read all of the data from /home/container/libraries/org/xerial/sqlite-jdbc/3.42.0.1/sqlite-jdbc-3.42.0.1.jar
java.io.IOException: Input/output error
at java.base/sun.nio.ch.UnixFileDispatcherImpl.read0(Native Method)
at java.base/sun.nio.ch.UnixFileDispatcherImpl.read(UnixFileDispatcherImpl.java:51)
at java.base/sun.nio.ch.IOUtil.readIntoNativeBuffer(IOUtil.java:340)
at java.base/sun.nio.ch.IOUtil.read(IOUtil.java:306)
at java.base/sun.nio.ch.IOUtil.read(IOUtil.java:283)
at java.base/sun.nio.ch.FileChannelImpl.read(FileChannelImpl.java:234)
at java.base/sun.nio.ch.ChannelInputStream.read(ChannelInputStream.java:74)
at java.base/sun.nio.ch.ChannelInputStream.read(ChannelInputStream.java:103)
at io.papermc.paperclip.Util.readFully(Util.java:43)
at io.papermc.paperclip.Util.readBytes(Util.java:30)
at io.papermc.paperclip.Util.isFileValid(Util.java:79)
at io.papermc.paperclip.FileEntry.extractFile(FileEntry.java:66)
at io.papermc.paperclip.Paperclip.extractEntries(Paperclip.java:217)
at io.papermc.paperclip.Paperclip.extractFiles(Paperclip.java:188)
at io.papermc.paperclip.Paperclip.extractAndApplyPatches(Paperclip.java:153)
at io.papermc.paperclip.Paperclip.setupClasspath(Paperclip.java:72)
at io.papermc.paperclip.Paperclip.main(Paperclip.java:28)
at java.base/jdk.internal.reflect.DirectMethodHandleAccessor.invoke(DirectMethodHandleAccessor.java:103)
at java.base/java.lang.reflect.Method.invoke(Method.java:580)
at io.papermc.paperclip.Main.main(Main.java:26)
container@pterodactyl~ Server marked as offline...
[Pterodactyl Daemon]: ---------- Detected server process in a crashed state! ----------
[Pterodactyl Daemon]: Exit code: 1
[Pterodactyl Daemon]: Out of memory: false
[Pterodactyl Daemon]: Aborting automatic restart, last crash occurred less than 60 seconds ago.```
I'm getting this error on startup any ideasi use paper not waterfall
twin solstice
i use paper not waterfall
then use #paper-help
mortal hull
i didnt see it lol until i sent it here too xd
agile lotus
I have a question.
Is BungeeCord really worse than Waterfall in terms of performance?
Do we lose possibilities by switching from Waterfall to BungeeCord?
marble kelpBOT
it has some issues, yeah. they tend to not matter for small servers though
agile lotus
it has some issues, yeah. they tend to not matter for small servers though
You talk to me ?
marble kelpBOT
yes? you were the only one to talk for a couple hours
agile lotus
Okay, so I didn't understand what you wanted to tell me.
Sorry, I'm French and I didn't really understand the meaning of your sentence.
Do you mean that if I report a small problem they don't care?
marble kelpBOT
no
agile lotus
Ok so I don't understand
marble kelpBOT
which part of it?
agile lotus
they tend to not matter for small servers though
marble kelpBOT
if you are a small server then you will most likely not run into the issues that waterfall fixes
marble kelpBOT
100 players at the same time might already big enough to experience issues
agile lotus
60 ? 🥲
ruby plover
Ok I have been trying but I can not figure out why reconnect velocity will not actually reconnect players to the last server they were on the velocity.reconnect perms is in luckperm also
Maybe the plugin just does not actually work anymore. Also so with Velocity you can fallback to a different server unless it is the 1st server on the list/lobby and you can not actually get players to be able to reconnect to the last server they disconnected from.
Guess maybe I just need to go to bungee if waterfall is no longer going to be around and supported I have been trying to get just this part figured out for a long time with Velocity and had it all working fine before. I am sure I am missing something maybe or like said it just does not work.
I really wish I could get some help here eventually as I do not even know what to send or where to look anymore. I know I sent my whole config file a while back asking if it looked correct but never heard anything back from anyone. We will try to hang it out for a while but it is annoying to be sent back to lobby server everytime relog into game.
I know there was a surgery but hopefully after some recover time if no one else can help would be wonderful 🙂
craggy trout
Ok I have been trying but I can not figure out why reconnect velocity will not a...
You're definitly getting helped on #velocity-help
agile lotus
Use Waterfall, is updated.
Yes but we are not sure that it will be updated for 1.20.1
craggy trout
It supports versions to 1.20.6
agile lotus
It supports versions to 1.20.6
Yes but I'm talking about that because I'm planning to upgrade my server to 1.21
agile lotus
Yes, he is having surgery, we hope that everything goes well and that he will be back on his feet quickly.
ruby plover
Ooo thanks that channel for some reason was not in my channel list as an option so I have been talking here since I use to have waterfall and I so loved it.
I hope for the best for him also and speedy recovery
serene falcon
@craggy trout please do not directly mentioning staff members unless they are actively involved in the discussion
Do note that Waterfall is reaching EOL and any update is not guaranteed.
craggy trout
# Do note that Waterfall is reaching EOL and any update is not guaranteed.
ht...
I will personally make some changes so waterfall can keep up a bit.
serene falcon
You are welcome to do anything you wish with your own fork. It has nothing to do with waterfall
serene falcon
I believe that if you re-read the announcement linked above, you will fully understand the meaning of EOL - it means that it will no longer getting support nor accepting contribution. Our proxy project is now focused on Velocity
vagrant sequoia
travertine 4 life
frank coral
ouha
tidal musk
ye
reef fulcrum
Is there anything in the logs? Only real difference between waterfall and bungee there is that waterfall will use alternative DNS servers by default unless you disable async dns in waterfall.yml
spare torrent
Ok let me chexk
There is ill paste it in a minute at pastebin
Could i pm it to you i see there is lot of other peoples ips here
reef fulcrum
Yea, you could try disabling that setting, looks like the dns servers that netty uses for dns by default might be having issues
reef fulcrum
It was added before I started working with waterfall and was never really benchmarked that I'm aware of, but, I'd imagine that where it works it's going to be better due to how it worked
full cypress
Note, I've had some issues, haven't gotten around to test and see if they were related to the async netty DNS resolver
tacit garden
is there anyway to change what dns servers it uses?
reef fulcrum
I wanna say that there was a system propery for it or something
unique linden
why would there even be a config for that
dns servers are your OS's responsibility
the only thing related to DNS I'm aware of is using the netty async dns resolvers
reef fulcrum
it is, but it doesn't respect the systems settings, it's using googles DNS servers
I have no idea why but some people seem to have intermittent issues with their DNS servers, however
reef fulcrum
Actually, it tries to grab the systems ones and then falls back to just using googles, wonder if some hosts have horrible dns servers which are breaking o.O
tacit garden
ok well i use cloudflare
for the dns servers so
hopefully that is working and not grabbing googles
silk gust
What does happen if netty reaches the Google DNS code and the localhost is ipv6 but the ipv6 Internet isn't accessible? 🤔
unique linden
pass flag to say prefer ipv4
oblique adder
you should use cloudflares for default
spiral crater
would travertine support go here too
reef fulcrum
Yup
spiral crater
ok
marble kelpBOT
boops the channel
haughty nimbus
unboop
strong mantle
reef fulcrum
Needs to be disabled in waterfall.yml and config.yml, bungee added it after us, I maybe shoulda required it to be true in both configs, but I'll probably just drop our side of that config
jaunty crag
what "thing" is it?
ah log_initial_handler_connections?
And then so which one is it in config.yml?
marble kelpBOT
log_pings or something like that
jaunty crag
ah
tidal musk
Have anyone EchoPet Plugin for 1.12.2 ? When i use this its give me io.netty.handler.codec.DecoderException: java.lang.IndexOutOfBoundsException on Waterfall ?
tacit garden
java.lang.NullPointerException: null
at net.md_5.bungee.BungeeServerInfo$1.operationComplete(BungeeServerInfo.java:145) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at net.md_5.bungee.BungeeServerInfo$1.operationComplete(BungeeServerInfo.java:139) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.util.concurrent.DefaultPromise.notifyListener0(DefaultPromise.java:511) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.util.concurrent.DefaultPromise.notifyListenersNow(DefaultPromise.java:485) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.util.concurrent.DefaultPromise.access$000(DefaultPromise.java:33) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.util.concurrent.DefaultPromise$1.run(DefaultPromise.java:435) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.util.concurrent.AbstractEventExecutor.safeExecute(AbstractEventExecutor.java:163) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:404) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.channel.epoll.EpollEventLoop.run(EpollEventLoop.java:335) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at io.netty.util.concurrent.SingleThreadEventExecutor$5.run(SingleThreadEventExecutor.java:909) ~[Waterfall.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:aad4b90:260]
at java.lang.Thread.run(Thread.java:748) [?:1.8.0_191]```What happened here?
tidal musk
NullPointerException happened
haughty nimbus
This is not necessarily waterfall question probably more of multithreading question.
I have a regular HashMap that stores servername as a key and Queue(LinkedList) that stores players.
Players are added when they switch a specific server and I have a async task that polls the queue.
Are there any concurrency issues I might run into?
sand steppe
yes but they will but subtle
haughty nimbus
how should I deal with this
full cypress
@haughty nimbus is there a reason you can't poll the players on the server from the ServerInfo objects?
after ServerSwitchEvent or ServerConnectEvent the users are put into a list attached to that ServerInfo object which you can get with getProxy().getServerInfo(ServerName)
and waterfall has a convenient method to get all the servers objects as a copy to avoid concurrency problems.
haughty nimbus
basically I'm making a queue plugin. Once they join they are added to the Map<String, Queue<ProxiedPlayer>>
then there is a task that executes every x units of time. In that task y amount of players are removed from the queue and sent to the specific server.
I can't really see a way how I could use that.
I guess problem is with Queue itself
full cypress
Ah, what you're doing is fine then.
What you need is locking to make it multithread safe.
but you should be safe for the most part.
haughty nimbus
yeah that's the problem
sand steppe
Use a ConcurrentHashMap and a ConcurrentLinkedQueue
How are you initing the map values as well?
(if it's all at startup you could drop the concurrent map entirely)
haughty nimbus
They are initialized only at startup. Then I just operate with Queues themselves
well ofc I use map to retrieve them but that's all
sand steppe
Then use CLQ instead
leave the map as it is and just use CLQ (provided you never modify the map)
haughty nimbus
Alright thanks
sand steppe
That'll work because CLQ is MT-Safe and since you never modify the map you don't need MT-Safety guarantees on it
Also make sure to correctly publish the map
haughty nimbus
what do you mean by publish?
sand steppe
Correct publishing is where you ensure the value can be read by other threads after it is written to a field
Final fields provide that guarantee
haughty nimbus
oh okay
sand steppe
(provided your mods to the value are made before writing to the final field)
You might not have to worry if you start your handler threads after initing the map
as that also provides synchronization
haughty nimbus
Well first it's populated and then those tasks are scheduled
haughty nimbus
can you recommend some good resources about MT ?
sand steppe
Most are doc'd in chpt 17 of the jls
Others are provided in documentation in atomic types
(ie atomicinteger) and more doc is provided in the java.util.concurrent package
haughty nimbus
Okay Thanks for help!
marble kelpBOT
Wanter to my server into a friends Mojang MC realm. Is that possible. Never used Mojang.
Want to...jezzz
marble kelpBOT
Realms is basically a private platform, waterfall can't be used with it
tacit garden
Should I be deleting the modules folder when updating waterfall?
reef fulcrum
No, waterfall will automatically update those
tidal musk
What would be the best thing to do against ping spam? Some kids are spamming with diff ips making it unjoinable, we’ve tweaked with iptables to limit the amount of requests per second and connection throtle but its still goes unjoinable
marble kelpBOT
did you try disabling the logging of the intial handler connections yet?
or maybe even the server list ping logging, iirc one of them was related to the performance impact such pings have
low glen
Yes we tried ^ didnt work, same issue as danny
Anyone know a fix? players literally cannot join the network due to this ping spam, all different IPs.
wild trellis
is it just pings or actual players joining?
reef fulcrum
I mean, that's happening in waterfall code, but it shouldn't bem seems like something is screwing with netty
tacit garden
yeah I can't figure it out
Im going to restore my config
@reef fulcrum What did query_port: 25577 do again? I forgot to add it to ufw
reef fulcrum
sets the port for the query protocol
tacit garden
tcp or udp?
reef fulcrum
udp
tacit garden
should I enable the query?
reef fulcrum
Well, I mean, do you need it?
If you're not using it/have no clue what it is, you probably don't
tacit garden
Unsure if I do to be honest
cool ok
well it has been a few minutes so far and no disconnect
just from resetting the config
might have been something weird going on since the config was generated a year ago with bungeecord
reef fulcrum
Doesn't really look like it would be a config issue
Something was janky with what the proxy expected from netty there
tacit garden
weird, I really dont know the error is gone and so is the disconnect issue
I rebooted it like 10 times previously and it still happened as i removed plugins
maybe it was because I had a duplicate entry in the config
two servers on the same port listed but only one server was active
tthats the only thing I can think of
re-enabled the plugins now
With all the plugins enabled I am still getting this error but a single one: https://hastebin.com/pezakipuvu.bash
but the player is no longer disconnecting
now to find the plugin...
long ridge
what's the difference between log_pings in config.yml and log_server_list_ping in waterfall.yml?
reef fulcrum
Ours was their first and then spigot added theres, and I'll be dropping waterfalls one soon™
long ridge
👍
reef fulcrum
looking back I did a dumb and made it an OR between spigots and ours, shoulda made it an and, and then dropped it down the line
velvet junco
idk if this is me or just a plugin that is causing this issue but I keep getting spammed with this error. Anyone able to assist me with this issue.
cedar quiver
https://i.imgur.com/aj5JPdM.png - this is for setting a version number that appears in server list, right?
reef fulcrum
No, it's a string
pretty sure it was, should be shown in the server list when somebody has the wrong version
velvet junco
Quick thing I get disconnection messages when a player disconnects from a modded server it shows up in my Hub (Paper Server) that they disconnected. What causes that issue
So when they leave the server they are one it comes up on Hub that they left the server.
tacit garden
I am using the NoProxy plugin. It seems to cause these on login [21:04:04 WARN]: Event PreLoginEvent(cancelled=false, cancelReasonComponents=null, connection=[/1.1.1.1:53961|Player] <-> InitialHandler) took 323,019,517ns to process!
What does this mean so that I can report it?
took a long time to connect?
tidal musk
Oh interesting, we have been having similar issues with our onlinemode paper server ever since i started a seperate waterfall server. Maybe it could be related. @tacit garden what was ur disconnect fix? Did u just remove query port?
tacit garden
My disconnect fix from yesterday was backing up my old bungee config.yml then regenerating a new one, and setting it up again. and yes I set query to false
No more disconnect issues after that
tidal musk
Ah ok il give that a try, cheers.
tacit garden
👍🏿
vivid hollow
Apologies if this is a stupid question, but is there a set of docs for Waterfall? I went to the ReadTheDocs site and it looks like its just for Paper itself
velvet junco
I have that issue as well " took 323,019,517ns to process!" when first starting up waterfall. but only happens once or a few times then goes away.
marble kelpBOT
Got a private server I want to connect with a Mojang realm using Waterfall. That possible?
I know dick about Mojang. Freind has the realm there.
reef fulcrum
Nope
reef fulcrum
You can't connect to a mojang realm using waterfall, it's basically an entirely locked off platform
marble kelpBOT
what a hunk of shit they are
vagrant sequoia
Seems sensible to me.
reef fulcrum
I mean, you pay for the convenience, Bungee also adds several concerns, which mojang 100% do not want to introduce to their network
marble kelpBOT
If I was retarded I would agree
what could they introduce at layer 7? Yeah maybe you trash your own world. I am assuming they have every realm in a isolated QEMU/kvm instance.
reef fulcrum
For bungee support, you'd 1) Be leaving mojangs systems 2) Disabling authentication on your realm instance in place of an unprotected connection 3) Mojang would specifically need to implement support for bungee logins
marble kelpBOT
You are probably right. It is profoundly amazing how stupid people can be.
reef fulcrum
Can you transfer worlds from realms?
I'd imagine that if you can grab the would, would be pretty trivial to spin up a small server to replace it, I have no idea what mojang does with that stuff, but I doubt that those instances are super powerful in the first place
tidal musk
you can sort of use bungee with realms
if you spoke to realms api
you can get the backing ip address
and if you modified bungeecord to handle online mode
but is it really worth it?
dusk abyss
I think realms runs on EC2, they can't be very powerful for that price
Unless it's Azure now because microsoft
marble kelpBOT
do....do you need power from realms though
vivid hollow
Is there a set of docs for Waterfall? I went to the ReadTheDocs site and it looks like its just for Paper itself
reef fulcrum
Currently, not really, what are you looking for?
vivid hollow
Mostly base-configuration stuff. Formatting for the locations.yml and what-not
reef fulcrum
locations.yml is a data file for the reconnection module
vivid hollow
I apologize, I am not intimately familiar with BungeeCord. Just kind of playing around with tying individual servers I have together
vagrant sequoia
which is a Bungee thing, isn't it?
I don't think there are any particularly interesting end-user stuff for Waterfall, just plug and play as you would Bungee, could be wrong though.
reef fulcrum
The config for bungee.yml is on spigots website, for waterfall it's mostly self explanatory, but feel free to ask anything, some stuff isn't so much
vivid hollow
ahh, I wasn't looking at the config.yml 😃
so it would follow the same configuration principals as bungee for that file then?
reef fulcrum
Yup
reef fulcrum
We avoid modifying bungees config, it would just complicate/break stuff
vivid hollow
haha
config.yml is where I define all my servers, and that would be about it I take it?
reef fulcrum
Yup
vagrant sequoia
Most of the benefits you get from using Waterfall are always on, automatically.
reef fulcrum
the waterfall configs defaults should be fine as-is
charred steppe
Hello! I've been trying to look for what some configuration of waterfall.yml does, but my Google searching didn't yield anything useful. Does anyone know a page that tells what each configuration does?
vagrant sequoia
I don't think there is one, but it should all be pretty straightforward. The default configuration should be fine for 95% of use-cases.
charred steppe
There's an option for game_version and I'm wondering what that means or what it effects. Also, not sure what disable_entity_metadata_rewrite does either ;-;
reef fulcrum
game version controls the version shown in the server list when a client ping returns an unsupported version, entity metadata rewriting is a mechanism in bungee that makes the jump between servers transparent, disabling that uses a more modern mechanism of actually sending new login packets to the client to allow it to reset its own state, it makes it much more comptable with mods as we don't need to support every metadata type that mods might add, but it also has the potential to break other things
glossy yacht
marble kelpBOT
no
marble kelpBOT
Project Waterfall build #265: SUCCESS in 4 min 37 sec: https://papermc.io/ci/job/Waterfall/265/
marble kelpBOT
Project Travertine build #90: SUCCESS in 1 min 9 sec: https://papermc.io/ci/job/Travertine/90/
Project Waterfall build #266: SUCCESS in 4 min 23 sec: https://papermc.io/ci/job/Waterfall/266/
tidal musk
it seems im unable to use any sort of plugin designed to send players to servers with no error messages just none of them seem to work? Ive tried multiple things from using plugins to run commands bungeeportals etc but with no progress
reef fulcrum
Plugins can't run commands, they need to send messages to the proxy for it (e.g. making a plugin run /server or /send will have no effect)
tidal musk
yeah but where bungeeportals would normally work in the past im not really getting any results
ive also tried other plugins designed for the same thing but all the same
reef fulcrum
ah, apparently bungeeportals is broken now due to 1.13
tidal musk
im running 1.12.2
tidal musk
1.13 snapshot
reef fulcrum
are you sure that the server names match up with those in your config.yml?
tidal musk
yup
the /server command works and shows the name and i can connect via that but it wont work on anything else
not sure what the issue is as i dont get any errors
reef fulcrum
I'd need to setup a server to test it and stuff, and right now that's likely not happening, it's always worked for me in the past though, so :/
tidal musk
i think i found the solution im an idiot i forgot to add the plugin to the bungee proxy
marble kelpBOT
Project Waterfall build #267: SUCCESS in 4 min 8 sec: https://papermc.io/ci/job/Waterfall/267/
marble kelpBOT
test
spiral crater
anyone else have a bug where 1.13 tab completes dont work on the initial server you connect to
you need to switch servers to get it to work, happens pretty consistently for me just need to go to hub and back to get tab complete to work
im on travertine 87
reef fulcrum
That's up to your plugins
(Those aren't waterfall messages)
disable ping passthrough in the config
No idea what your voting player list derives info from, but if it's relying on the response from the server for that, I don't think that bungee filled in the send player list, not too sure
marble kelpBOT
Should do, there was some plugin that did all of that
reef fulcrum
Use a plugin, basically
Not too sure, really
Bungee has never supported showing the players there
shadow walrus
Hello, so I have an issue: I have a vanilla lobby and modded servers, I am using the newest Waterfall and I configured the config properly, however whenever I try to connect to the modded from the lobby it just says "Connected to XX server" and then instantly after it says "Disconnected from XX server" with no visual changes in game or in console from either servers besides the waterfall one.
marble kelpBOT
Can't really say without any form of logs, if the connection was closed it was closed by something for a reason
reef fulcrum
There's nothing useful there, we'd need to see a message indicating to why the client was disconnected
you might need to enable some debug mode on forge, not toyed with it in a good while
reef fulcrum
On the server most likely
You could also try flipping the entity metadata setting in waterfall.yml
shadow walrus
I did, but what does it do?
Also, I turned on the debug mode for forge and nothing pops up when I try to connect via waterfall
reef fulcrum
It basically stops rewriting entity IDs by sending a login packet to the client, allowing its id to reset; fixes a lot of issues as mods don't expect the client and server entity IDs to be out of sync
shadow walrus
https://pastebin.com/PH9wKnb0
Here is the config for my waterfall
reef fulcrum
If there are no logs, not too sure, something caused the connection to close, that's all we can really say from that
shadow walrus
I only changed the ipforwarding
silk gust
hmmmm waterfall seems to send legacy:fml|hs in minecraft:register even if it is in 1.13, it is making my fabric client be disconnected
tacit garden
Any way to fix something like this [01:50:17 WARN]: Event PreLoginEvent(cancelled=false, cancelReasonComponents=null, connection=[/1.1.1.1:52236|TeaMp0isoN] <-> InitialHandler) took 2,953,243,220ns to process! Sometimes happens on login takes them a long time.
reef fulcrum
That's plugin related
unique linden
that value should be converted to ms....
ns just seems fakely large
that was 3 seconds
reef fulcrum
BUT SCARY
marble kelpBOT
Project Waterfall build #268: SUCCESS in 4 min 4 sec: https://papermc.io/ci/job/Waterfall/268/
tacit garden
Can you convert to ms
reef fulcrum
I'm actually about to push that, as I agree there
tacit garden
thanks
I need to go over java and learn how to make my own waterfall plugins
I am tired of dealing with plugins acting funny
The error above is so hard to figure out which one caused that then when you report it to the dev they have no idea what it is or how to fix it
tacit garden
ill look into that i forgot about profilers
sand steppe
do you enjoy the code style electronic
reef fulcrum
if timings ever gets to the point that we could add support for timings, I'll consider it
tacit garden
that would be nice
plugins causing issues with paper sucks but they are easy to fix usually
waterfall errors not so much and usually cost you a few players from disconnects
when i say waterfall errors i mean plugins running on the waterfall instance
otherwise waterfall is solid
marble kelpBOT
Project Waterfall build #269: SUCCESS in 4 min 4 sec: https://papermc.io/ci/job/Waterfall/269/
tacit garden
👍
marble kelpBOT
Project Travertine build #91: SUCCESS in 58 sec: https://papermc.io/ci/job/Travertine/91/
tacit garden
Oh one other thing dont know if you can answer but why would a plugin cause the player to take a while to login randomly and even just disconnect after it times out?
some sort of weird player handling?
i dont know much about how waterfall works besides it is a proxy
reef fulcrum
networking stuff?
e.g. plugins using mysql or making external requests might "have a moment"
tacit garden
yeah you know what
you are right
mysql
I tried out litebans that is what must have caused it
it hangs the server until the connection connects
even though it is a 5ms response time between the mysql server and client for some reason litebans takes a few seconds
must have been a reconnect or something
unique linden
timings in waterfall makes no sense
theres nothing relative to time towards
theres no main thread, no tick rate, it would just act as a random profiler at that point lol
tacit garden
it would be nice to have something besides installing a profiler on the server
unique linden
if you want to port the code behind TimingHandler to act as a report of 'these tasks take on average x time to run', sure go for it
but it wouldnt make sense for my viewer, theres no 'start of the stack' to relate to
cant build a tree
reef fulcrum
I was somewhat hoping that regional timings would add that, wasn't too sure what your intents for that was
unique linden
regional timings purely identifies world context so you can see region X is using 80% of that BlockPhysicsEvent timing
reef fulcrum
Hm, I'm thinking it makes sense to just move that event code around invokation of the event too, but I'd need to deal with intents too
unique linden
to show you WHERE
tacit garden
would it be possible to log the response time of a plugin (not sure what I am saying here), and where warnings came from what caused them (which plugin)
this is out of my range but maybe youll understand
reef fulcrum
The former is what we're basically talking about
reef fulcrum
Yea, my intents was to basically use timings as a profiler, but looking at it that wouldn't actually fit too well with how timings works, was kinda hoping future changes would make it more viable, but I forgot the overall concept of timings kinda breaks here
reef fulcrum
You'd not think that english was my main language at times
tacit garden
i feel the same way
random topic here but would it be possible to somehow implement the ability to hand over a player to another waterfall instance
tacit garden
dang that would be really cool
tiny monolith
onyx condor
hey umm, for a creative server, what would u suggest? a plot plugin? or somethign different?
tidal musk
-> #paper-help
frail girder
Hello.
Sorry may be wrong channel..
I want transfer message info from Spigot server to BungeeCord (I use fork WaterFall but code from Bungee).
From "plugin messaging" thread I found what I need:
p.sendPluginMessage(Main.getPlugin(Main.class), "BungeeCord", b.toByteArray());
But if server without players.. I not understand how to send message.
Could I get any help with how send message from server to BungeeCord without players?
zinc sapphire
bungee message channels operate via player connections
reef fulcrum
You can't basically, because ^
zinc sapphire
no players no message channels
reef fulcrum
If you need to send messages to servers without players, you'd need to look into creating your own socket connection or using a messaging broker like rabbitmq/redis
frail girder
Are there any implementations?
reef fulcrum
implementations for what?
frail girder
For messages to server without players.
reef fulcrum
Not built into the server, but pretty sure you'd be able to pull something together using sockets, or use redis/rabbitmq and use a library to send messages
frail girder
k, thx. Going make a lot of bad code for send messages. Thx again.
spring shale
What is this error please :
[02:52:15 ERROR]: SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder".
[02:52:15 ERROR]: SLF4J: Defaulting to no-operation (NOP) logger implementation
vagrant sequoia
It means you're trying to use SLF4J API without an implementation.
spring shale
How to resolve this ?
vagrant sequoia
I don't think you should be getting that error in Waterfall, because Waterfall should have an SLF4J implementation?
I'm also using SLF4J in my Waterfall plugin, and that doesn't happen for me.
Waterfall comes with the log4j-slf4j-impl implementation.
spring shale
What is that implementation ? What type of plugin ?
vagrant sequoia
It's an implementation of SLF4J that logs to Log4j. Waterfall comes included with it. The only way I see that error happening is if you're not using Waterfall (e.g. using BungeeCord instead).
Or maybe a really outdated version of Waterfall
spring shale
I updated WaterFall, and tried bungeecord too, but the error is already here, but I have another huge problem I got this error : https://pastebin.com/WmnZgdJu
It makes I can't connect to my server Faction, but the Lobby works well, it's very strange
spring shale
What can do that ?
reef fulcrum
Generally down to plugins on the server
spring shale
I tried to remove all my plugins, same in vanilla the problem occur
I just use BungeeGuard, that protect my server, but It won't works on waterfall
errant pivot
i know its diff thing but can anyone help me with bungee
players on my server getting kicked randomly
same thing on my other server
5 players left
bungee console log
mint spoke
Any way to remove this exception from waterfall?
java.lang.NullPointerException: info
at com.google.common.base.Preconditions.checkNotNull(Preconditions.java:787) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at net.md_5.bungee.UserConnection.connect(UserConnection.java:282) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at net.md_5.bungee.UserConnection.connect(UserConnection.java:273) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at net.md_5.bungee.connection.InitialHandler$6$1.run(InitialHandler.java:541) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at io.netty.util.concurrent.AbstractEventExecutor.safeExecute(AbstractEventExecutor.java:163) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:404) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at io.netty.channel.epoll.EpollEventLoop.run(EpollEventLoop.java:333) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at io.netty.util.concurrent.SingleThreadEventExecutor$5.run(SingleThreadEventExecutor.java:905) ~[BungeeCord.jar:git:Waterfall-Bootstrap:1.13-SNAPSHOT:0691b63:269]
at java.lang.Thread.run(Thread.java:844) [?:?]```dont pass nulls
mint spoke
I think its coming from bc I'm clearing the server set
marble kelpBOT
yeah, it probably doesn't work properly if there are no servers defined to connect a player to xD
mint spoke
I add them back though
I basically, have a dynamic setup so I can add instances as and when I want with random port generation and shit on startup so I can't add a server with an IP, and port hence me clearing them then I add my own set.
It allows me to join bc I define myself being teleported to a lobby.
marble kelpBOT
well the ping/connection probably happened in the time between you clearing and filling it again
or maybe if it occurred afterwards you didn't do some re-registering of the servers that is normally done on startup?
maybe look into how BungeeServerManager does it: https://github.com/tavonkelly/BungeeServerManager might be something special that you are missing
(although I had issues with that plugin and adding/removing servers before so I dunno, there might be deeper issues in Bungee directly when doing that dynamically)
silk gust
why does waterfall send legacy:fml|hs (which is an invalid identifier) in register for 1.13+ clients?
forest halo
bungeecord does, so we do
mint spoke
phoenix616: Just had a look the only difference is that they add it to a config, but that shouldn't make a difference.
errant pivot
players on my server getting kicked randomly
https://gyazo.com/977a04977d591349d976b746202bbf65
same thing on my other server
5 players left
https://hastebin.com/witumexoci.shell
bungee console log
spring shale
@tidal musk Yes I have skin restorer
tidal musk
has anyone been able to get Buycraft push commands working on server that uses 127.0.0.1 for the paper server ip, on a waterfall proxy?
errant pivot
so ok no one can help ?
reef fulcrum
Your logs only really show networking issues
errant pivot
is it server or players
reef fulcrum
players
errant pivot
so all 8 players exact same moment ?
reef fulcrum
It's a general network error
reef fulcrum
All that error means is that bungee didn't recieve a packet in so long from the client
errant pivot
same moment
reef fulcrum
Generally boils down to network issues somewhere
errant pivot
hmmm
reef fulcrum
Be that an intermittent issue or "your host is bad" issue
errant pivot
connection_throttle_limit: 3
irrelevant
reef fulcrum
The proxy is not the issue here
errant pivot
i never had issues with ovh as i remember
reef fulcrum
So, could easily have just been an intermittent issue
if you keep having issues, try using tools like mtr to monitor the connection to the server, see if there are any packet loss issues, etc
errant pivot
ok ty
tidal musk
@spring shale make sure that you're on latest version of sr
errant pivot
/entrypoint.sh: line 15: 33 Killed java -Xms12G -Xmx16384M -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 server.jar
[Pterodactyl Daemon] Server marked as OFF```anyone can tell me why it caused sudden kill for server
unreal stag
smells like a Docker thing, specifically speaking, your server has allocated too much memory and the kernel mauled it
allocate less memory - 50% of the Docker memory limit is what I'd recommend for OpenJDK
reef fulcrum
Disable the OOM killer, I have no idea what the thing was, but there was a way to do it in ptero
errant pivot
i have 64 GB
reef fulcrum
Basically, their default flags are stupid, and while without aikars flags it won't crash immediatly, their setup basically means that you're in a condition where the server could crash; The other alternative as Tux stated would work, but you'd be wasting memory
errant pivot
and gave 16gb to this server
reef fulcrum
Yes, but you're using docker containers
errant pivot
so what i can do to fix plz
vagrant sequoia
The issue is not about how much memory you have
To fix it, disable the OOM killer
errant pivot
how
vagrant sequoia
Ask Pterodactyl support how to do that.
errant pivot
ok
reef fulcrum
The idea is to not disable the systems OOM killer, but the one used on containers
Ptero had some config file which would let you specify that somewhere
unreal stag
Ah! The Docker flag --oom-kill-disable
reef fulcrum
But, issue with ptero is that if you tell a container it has 16G, it has 16G, use 16G on the heap = boom
errant pivot
but i have one server with 1 player and 10gb and that also crash
vagrant sequoia
Again it's not about how much memory.
It could be 1 GB or 64 GB, the outcome would be the same.
reef fulcrum
The issue is that AIkars flags configure the server in a way that your JVM crashes early
unreal stag
The difference is Pterodactyl, which uses Docker, which uses cgroups, and yes, Aikar's flags.
errant pivot
i was using Pterodactyl
vagrant sequoia
It's about the extra memory consumed by the JVM when using Aikar's flags.
reef fulcrum
This is because pterodactyl is stupid and configures servers in a state that you can crash them just by using the heap used by default
unreal stag
Okay, fair enough.
Yeah, Pterodactyl is at fault here. You can either configure Pterodactyl to disable the OOM killer on the containers or reduce your memory allocations.
vagrant sequoia
You have to account for heap size + JVM additional footprint when using those flags.
errant pivot
ok!!!
vagrant sequoia
The problem with reducing the memory allocation is that Pterodactyl passes the configured memory as the heap size.
As arguments
unreal stag
I've seen a lot of voodoo magic that tries to get Java to play nice with Docker
vagrant sequoia
Although you could of course hardcode the startup arguments for that server.
reef fulcrum
There are flags to make java more aware of it's cgroup limits
unreal stag
None of it worked, I eventually gave up and used OpenJ9, which worked perfectly at the cost of reducing my memory allocations. OOM disable isn't an option for us because we use managed Kubernetes.
vagrant sequoia
I'm super impressed with OpenJ9's memory usage. I give it 1GB and it uses like 600MB lol
It doesn't ever seem to actually use more or equal to the Xmx I give it
Xms*
Which is weird
But okay 😛
You'd think it'd allocate the Xms you tell it to, but surprisingly doesn't seem to.
unreal stag
English.
tidal musk
wat
abstract bolt
hey, i'm having some difficulty getting waterfall up and running. Always times out when trying to the send the player. If anyone would be open to hopping in a call with me (share screen), I'd appreciate it very much.
tidal musk
does this happen with bungeecord as well?
abstract bolt
yes, exact same thing
i had the base bungee working with the bungee server and a vanilla server, but after i changed the other server to sponge/forge and installed spongepls it would just time out
so, i found someone being recommended to use waterfall. I'm still encountering the same issue though
tidal musk
sounds like a generic shitty connection scenario?
1.12.2+ is more fragile with poor latency
abstract bolt
that's what i would have thought, but I've had other people try and they had the same thing
in sponge/forge server that it's trying to connect to doesn't even recognize something's happening in the console, and the only info the waterfall/bungee gives is timed out
abstract bolt
any suggestions would be appreciated
reef fulcrum
Depends on why it's happening, forge has some java flags to control the tiemout values, waterfall.yml also has an option to disable entity metadata rewriting, which can help fix compat issues with some things; Also, SpongePls should be removed when using waterfall, we have the relevant patches already in the proxy
abstract bolt
spongepls has been removed, the default timeout is 4k which should be way more than enough time for the server to process it
ill try the waterfall.yml
is was already disabled
i may have found it, and if this is the issue will feel incredibly stupid, is it necessary to have a game version in the waterfall.yml
currently it's blank
abstract bolt
Also, waterfall should be ran as the server jar right? Rename it to custom, etc. it’s not like a plugin like sponge pls
marble kelpBOT
Project Waterfall build #270: FAILURE in 1 min 24 sec: https://papermc.io/ci/job/Waterfall/270/
reef fulcrum
Grrrr
marble kelpBOT
Project Waterfall build #271: STILL FAILING in 2 min 32 sec: https://papermc.io/ci/job/Waterfall/271/
Yippee, build fixed!
Project Waterfall build #272: FIXED in 2 min 34 sec: https://papermc.io/ci/job/Waterfall/272/
marble kelpBOT
Project Waterfall build #273: SUCCESS in 2 min 34 sec: https://papermc.io/ci/job/Waterfall/273/
rough warren
That was the happiest webhook i've ever seen
marble kelpBOT
Project Travertine build #92: SUCCESS in 56 sec: https://papermc.io/ci/job/Travertine/92/
marble kelpBOT
does anyone know why when I join my second hub that doesn't have any plugin show the same tablist and scoreboard even tho it doesn't have plugins I use paper
modest pagoda
I need help, the channels are out of order D: https://i.imgur.com/wO2GLaJ.png
reef fulcrum
Discord needs a way to lock the channel order
It's so stupid how easy it is to move a channel without noticing
balmy crag
[20:48:37 INFO]: [FireIsH0t] disconnected with: Exception Connecting:NativeIoException : syscall:read(..) failed: Connection reset by peer @ io.netty.channel.unix.FileDescriptor:-1
<-> ServerConnector [pixel] - NativeIoException: syscall:read(..) failed: Connection reset by peer
anyone know how to fix??
unique linden
@balmy crag you dont, its an issue outside of your control
oh wait server connector may be bungee <-> MC
peer means "someone else killed this"
sick pulsar
what's the newline code for motd's?
reef fulcrum
\n
sick pulsar
it doesn't work
just shows up as "\n"
for some reason, waterfall uses a different symbol for color codes
hardy egret
its from the plugin that u use for motd
marble kelpBOT
it should support any standard yaml multiline string
reef fulcrum
Ah, yea, use " instead of ' otherwise the newlines will just be ignored
shy yoke
I need help. My server recently got hacked and the hacker managed to login with my account with a different UUID, I'm not sure if thats a waterfall hack or I'm missing a configuration. Please help.
I'm using a shared host with a dedicated IP for my waterfall and a normal IP for my other servers.
reef fulcrum
You need to use something like bungeeguard, especially on shared hosts, there are other plugins which have ip filters, but those don't work as expected on shared hosts
solemn flicker
Will 8MB heap suffice for waterfall to run?
vagrant sequoia
8MB? no, at least 512MB for a few hundred players
Should be able to get away with 128-256MB for anything less
vagrant sequoia
If you're using any plugins in Waterfall, then it might require a lot more, but those estimates are based on a pluginless set up.
marble kelpBOT
Does buycraft maybe using another port for that? Would seem odd that it would hijack the existing system
reef fulcrum
oh, it does, how nice
errrm
Can they not connect through bungee?
Exposing ports to the outside world seems a bit iffy, unless they have some documented IPs that connections will come from?
ah, that's fine then
rough warren
Custom afaik
unique linden
thats a very simple waterfall option we can provide, proxy ports - on connection to other ports (as defined), round robin to a defined server list or all active servers, so you can listen to votifier proxy and using SNI route for forced hosts even.
though votifier uses plaintext+encrypted payload doesnt it
not TLS
reef fulcrum
If we was to do that, it would basically need to be dumb forwarding
unique linden
can do port: server or port: <list of servers> or port: <ip>
reef fulcrum
No idea what buycraft is doing, but more than likely they're doing something that's not going to be compat with the protocol or will likely upset bungee
unique linden
right i meant to simply be a netty proxy, nothing to do with mc protocol
only on diff ports
for me votifier it round robins to random game servers for me
atm i have 3 mc servers all asking for the port, only 1 gets it
then votifier goes to 1 of the 3 machines
which then hits whatever process won the port
a bungee level proxy could simplify that
spiral crater
just setup a firewall 😃
unreal stag
I believe BuycraftX is using an HTTP server.
tidal musk
Aikar that sounds like a cool idea
reef fulcrum
You need libxrender installed on your system
lilac quarry
how do i install it? I have the vpn recently and I don't know some things
abstract bolt
hello, i found an error in my system and honestly i don't know how to fix whats wrong (even though it seems straightforward)
[Server] WARN Could not bind to host /158.69.55.120:25606
[Server] INFO io.netty.channel.unix.Errors$NativeIoException: bind(..) failed: Cannot assign requested address
What can I do to fix this? IP correct, online mode false ( true on bungee), so not sure what to double check from here
unique linden
@abstract bolt check listeneres: section, host: where you have 158.69.55.120, sounds like that not the IP the host gave you
either should be 0.0.0.0 or ask your host what IP to use
abstract bolt
that's probably it, thanks
Next problem I'm getting:
[Server] WARN -> UpstreamBridge - read timed out
[Server] INFO disconnected with: ReadTimeoutException : null
Seems like just a common time out, but it's not like I or the service has poor connection... any suggestions?
unique linden
@reef fulcrum ^
marble kelpBOT
That's pretty much your standard timeout, depends on when it's happening as to what you'd need to do
abstract bolt
The only information I’m getting are those lines, so I don’t have much to go on... unless I’m missing something
abstract bolt
How can I determine when it’s happening/the solution then?
If there’s no easy way to tell I’m down to hop in a call and share screen. I’d appreciate it
reef fulcrum
Best suggestion would be to use tools like mtr to keep an eye on the network, see if you experience any packet loss/latency spikes
abstract bolt
Alright, thanks. Give it a look soon and see what I get. Thanks for the help
stone aspen
Yo, I've been having an issue with waterfall recently where people are randomly timing out and getting kicked to the hub, but theres no error on either server or proxy logs, but the other servers are fine on the network
some of my players says it only happens around 10ish PM EST and I have a theory it might be someone with bot accounts trying to connect with names of online players but fails cuz online mode but it still kicks us, but that's just a theory
Anyone else had this issue and know the fix? D:
stone aspen
@full cypress nope, it just freezes the server then kicks us to hub, but it only freezes for the people who get kicked which isnt everyone on the server, and when it happens it happens many times for like 30m then it stops
full cypress
Its hard to determine if thats server sided or waterfall sided then.
hmm. how much memory you have allocated to waterfall?
reef fulcrum
being kicked to the hub suggests an issue between bungee and the server itself, what's the disconnect reason on the server?
stone aspen
@reef fulcrum https://hastebin.com/ucutebuvam.md
reef fulcrum
Something caused the connection to close
Anything in waterfalls logs? Looks like a network issue between waterfall and your servers
stone aspen
@reef fulcrum
[03:29:59] [Netty Worker IO Thread #5/WARN]: [/(ip)|Purpdan] <-> DownstreamBridge <-> [UHC1] - read timed out
[03:29:59] [Netty Worker IO Thread #5/INFO]: [Purpdan|/(ip)] <-> ServerConnector [PvP] has connected
the time is a bit earlier cuz it takes like 10s to switch server but thats my plugins fault so ignore that
reef fulcrum
Yea, something killed the connection
Are you running everything on the same box or shared hosting or something?
stone aspen
same box
stone aspen
is what I'm using
which is basically 127.0.0.1
The issue only seems to be happening on the NA server, the EU server is fine
and both have the exact same setup
I also got this incase that tells you why they get kicked to hub and not fully disconnected
stone aspen
um
after updating to the latest waterfall version the server isn't getting the right uuids anymore
ipforward is stil on, bungeecord is still set to true
but uuids are not correct for some people
reef fulcrum
example of a uuid? online mode? plugins? any errors in the logs?
stone aspen
https://namemc.com/profile/danielhaich.1 is one example
{
"uuid": "68d3cd94-de08-33ac-839d-4ae0aeeced06",
"name": "danielhaich"
},
is what got sent to me when whitelisting
online mode is on in the bungee config
reef fulcrum
That's an offline mode uuid
stone aspen
and some people are still getting the right uuid, some are not
@reef fulcrum no errors in the logs, just checked
plugin list
UHC is my custom plugin
full cypress
Thats a known issue
you have to whitelist the UUID, not the playername
or have them log in once and then whitelist
stone aspen
@full cypress whys it never happend for the year I've used bungee then until this weekend
full cypress
In the bungee config there is a section for server connect priority.
delicate glen
Can I get here help for Travertine? If yes, is there a public maven repository for it and where?
reef fulcrum
Yes, there is no public repo for it, just use the waterfall-api
novel saffron
Hi there.
There seems to be an issue where you can't place NBT signs (gotten from cmd+copy (middleclick)).
unique linden
wrong channel
twilit seal
Howdy feller
I must be stupid
but this error pops up when attempting to go to my hub server
however everything else is the same
shrewd zealot
does the server have that port open
twilit seal
yes
marble kelpBOT
Project Waterfall build #274: SUCCESS in 2 min 37 sec: https://papermc.io/ci/job/Waterfall/274/
tidal musk
I am trying to use this plugin https://www.spigotmc.org/resources/party-and-friends-for-bungeecord.9531/ and it sends output to players on the servers without any server-side plugin/addon.
I am able to see the formatting of the messages (colors, underline, bold, etc) but another player is not. We have the same permissions through LuckPerms, neither are opped, both using 1.13.2 vanilla clients.
He is able to see formatting in the output of every other plugin and messages I send.
What could cause formatting codes to not work on his end?
marble kelpBOT
@tidal musk are you sure that the other player is not using any mods? 🤔
the only thing that I can notice regarding messages is that the plugin isn't using the chat component api but the server should automatically convert the legacy format. it's doing some interesting "fixing" of colors though: http://git.io/fjtd7 I think that's just for getting rid of line break issues that exist when not having non-chat-component colors
but even if it sends raw §'s and there is an encoding issue with them it should display some error characters in the chat instead of simply no colors
tidal musk
So there is a newer, better API (the "chat component api") that would be better to use? I'll forward that along.
marble kelpBOT
minecraft is using a json based chat components now
there are severl different implementations, Bungee (and spigot) include the bungee-chat compoenent API
but internally it just converts legacy codes to that so I really doubt that's what causing the issues
twilit seal
How do I setup a fallback server when another server goes down
or restarts
reef fulcrum
You need a plugin like MoveMeNow for that
twilit seal
rip
paper fox
I'm using pterodactyl and am having issues with getting Luckperms to work with the mariadb server
It's really, really weird
strange brook
does that actually have anything to do with waterfall? luck has his own discord fyi
shrewd zealot
is port 3306 open and listening on that machine?
paper fox
my bad wrong channel
anyway, figured it out, it was me putting unnecessary quotes in
apparently it took them as part of the password
paper fox
actually, that might not have been the issue
vagrant sequoia
Pterodactyl runs servers inside of Docker containers. A container's network is isolated from its host's network.
You will probably want to ask the Pterodactyl support how to normally deal with that.
You'd probably need to supply a different IP, instead of 127.0.0.1.
rough warren
Its 172.18.0.1 by default if you did not change anything
But its defined in the daemon's config file iirc
mint spoke
Hey,
So I'm using waterfall as a proxy server and it's not giving the IP of the players, it's giving the IP in the range of either our rack, or an ip of our ddos protection.
- I've got ip_forward: true in config.yml (Bungee)
- I've got bungeecord: true in spigot.yml
Very confused xD
reef fulcrum
stop all servers, ensure that you have everything configured properly and then start it
mint spoke
Tried that, its showing the same with /ip aswell
reef fulcrum
no idea what /ip is
mint spoke
reef fulcrum
If the proxy itself is seeing the wrong IP, sounds like you have an odd network setup
mint spoke
Every single one of our dedis are setup the same,
Except they don't use Bungeecord / Waterfall. (GotPVP, Arkham, PlayMC etc)
reef fulcrum
For an IP of one of your other servers to show up in there, sounds like you're using a tunnel or something
mint spoke
It ain't showing an IP of one of our other servers tho,
It's showing one in the range which is weird asf
reef fulcrum
I'd suggest speaking to your hosting provider
rough warren
/ip is a bungee command :)
simple silo
@mint spoke what do u use for ddos protection
strange brook
turns a ddos into a dos just like that :^)
shrewd zealot
quick maffs
unique linden
@mint spoke if your DDOS protection is using simple tunnels, thats expected. it would have to tunnel using proxy protocol to pass the real clients IP down. Unless the DDOS is filtered at the routing level to your host such as OVH where real IP is maintained.
If your routing packets at iptables level from 1 rack to another, that would be a network level issue not masquerading the connection correctly.
someone mentioned bungee has built in support for proxy protocol? is it same as what this does? https://github.com/MinelinkNetwork/BungeeProxy
rough warren
ye
cold jay
How can i use game_version?
reef fulcrum
It's for the version shown in the server list when it doesn't detect the correct version
cold jay
Oh, ok
distant kite
waterfall used to have PacketAPI?
tidal musk
no?
reef fulcrum
There was plugin libraries for it
Anybody else just reflected into the proxies packet mappings
full cypress
^ I reflect onto the map to allow the ability to filter chat messages at a bungee level
rough warren
Why not just chatevent
full cypress
Cause from Server -> Client there is no chat event. And setting the text in the chat event doesn't actually change the message, so I need to rebuild the packet after I recieve it
marble kelpBOT
okay so how does one get tab completion to work on Bungee? xD
neither the event nor the command TabExecutor seem to be resulting in suggestions? 🤔
sour raptor
There is a config option in the bungeecord config
marble kelpBOT
what setting would that be?
sour raptor
There is a setting "inject_commands" which should inject the bungeecord commands.
Are you missing spigot tab completion too?
marble kelpBOT
no, only bungee plugin commands
sour raptor
inject_commands=true should solve the problem.
marble kelpBOT
thanks, gonna try that
rough warren
That is default behavior i believe
Inject_commands disappeared in upstream and its always true now
marble kelpBOT
uh, I guess my version is a bit older... xD
unreal herald
Did travertine still get support ?
Well some one else having issue when some times its kick the plyaers
from the servber
reef fulcrum
Some else should provide information about their setup
reef fulcrum
Closed due to inactivity and likely to not be our issue
Errors like that literally just mean that something threw an exception somewhere down nettys pipeline
tidal musk
that issue is absolute cancer
reef fulcrum
tidal musk
obligatory if you're running anything other than latest then you should be able to read and understand errors that are thrown
thorn magnet
Does Waterfall improve plugin performance or proxying in general?
tidal musk
that's really vague question
thorn magnet
Nevermind, I'll probably use Velocity.
reef fulcrum
Waterfall isn't a plugin
tidal musk
About to switch from bungee to waterfall on spigot. And partly because we have 14 servers and one is a forge that we would like to connect to the network using the spongeforge setup. Anything I should be aware of before we pull the trigger
marble kelpBOT
I guess, make sure that forge support is enabled in config.yml as bungee has it disabled by default, there is also an entity remapping setting in waterfall.yml, which might prove handy if you start having issues with some mods
marble kelpBOT
Project Waterfall build #275: SUCCESS in 2 min 40 sec: https://papermc.io/ci/job/Waterfall/275/
celest maple
i need help
tps server lost
https://timings.aikar.co/?id=9a02460bdcc74681a2bdd093cf786b69
abstract spindle
@celest maple Hi, this channel is only for waterfall, still...
Lag reasons:
PermissionsEx regex system
QuantumRPG
See the count value, maybe too many entities?
This seems to be lagging too, and see the count is pretty high
Gangsplus is lagging too
celest maple
Thank you for your help
And I'm sorry for the chat channel waterfal
😭
abstract spindle
No problem :)
abstract spindle
I have a question, why is waterfall separated from travertine instead of having a option on waterfall to be 1.7 compatible? Isnt travertine literally waterfall with 1.7 support?
stone aspen
ok yeah I just downgraded waterfall version to 258 again and I'm now getting the right uuids of offline players when whitelisting etc
reef fulcrum
Bungee doesn't have an impact on that unless you're trying to ban players who haven't joined yet, which is problem with bungees online mode setting not handling that properly
stone aspen
¯_(ツ)_/¯
tidal musk
@abstract spindle that's right, travertine is literally waterfall with 1.7.10
reef fulcrum
Travertine sucks, well, 1.7.x support in general sucks
We're talking about 5 year old versions of the server in which a lot has changed, the underlying protocol is the same, but things like what packets require or other limitations have changed, you're much more likely that you'll have a clean migration from bungee > waterfall vs waterfall > travertine
tidal musk
nou, 1.7.10 is the best! >:(
abstract spindle
@reef fulcrum I agree that 1.7 sucks, but most of the pvp community still uses it plus more mods are releasing that forces players to use 1.7.
People uses that mods/clients even if in the terms it says you will give them access to all your files. (lol)
sand steppe
any examples? :>
abstract spindle
You know the one im talking about
sand steppe
no I don't actually
reef fulcrum
Doesn't matter that people use it still, it just sucks in terms of supporting it, and each time mojang starts going ham with the protocol, it becomes more to maintain
We still have to limit stuff like prefixes because of 1.7.x
abstract spindle
Well just making it a option for travertine saying "WE WILL GIVE NO SUPPORT FOR 1.7" could be good maybe
reef fulcrum
It just's not viable for it to be in the same repo as paper due to the complications it has
I mean, same repo as waterfall
reef fulcrum
and adding configurations for stuff like that gets a bit ott in terms of getting config info around properly
shrewd zealot
@abstract spindle I thought pvpers preferred 1.8 was the diff
abstract spindle
Mostly 1.7 because of the delay bug of 1.8, 1.8 is prefered by more "casual" players
reef fulcrum
There was a bug with the hitposition of the cursor being behind a tick or something
THere was a mod to fix it iirc
tidal musk
yeah, prplz made a tiny mod to fix this
ivory basalt
is there a page which explains the options in waterfall.yml?
reef fulcrum
documentation psh
I was going to work on the docs for waterfall, but kinda got about as far as reorganizing a few things before I started questioning my own judgement
ivory basalt
ok, use_netty_dns_resolver. What's this used for? Converting IP addresses to hostnames?
reef fulcrum
Yup, I forget what it was used for exactly, think it was more dns handling for mojangs lovely servers
ivory basalt
game_version is it " '1.13.2' " if I want to limit to only 1.13.2 clients to connect?
allow_empty_packets Whats the disadvantage of allowing empty packets?
reef fulcrum
game version is what shows in the server list
ivory basalt
if I set game_version, can otehr versions still connect though?
reef fulcrum
allow_empty_packets is one of those "shooting yourself in the foot type things"
The client is more tollerent to bad packets them bungee, that setting basically edges it a lil' closer to vanilla in what it will handle
And yes, others can still connect if there is a server that will accept them
ivory basalt
ok
marble kelpBOT
does anyone have an idea what could cause this error when loading a yamlconfiguration? https://i.moep.tv/ksz5H1fj.txt the code in question is just loading a config? http://git.io/fj3lm
What's the file it loading look like?
https://i.moep.tv/fmnwLJSt Imo only standard stuff
(and yes, we currently have a merged Bukkit and Bungee config for that :S)
strange brook
ok but wtf is that copyright LUL
marble kelpBOT
:P
hm, I have the suspicion that the configuration in bungee doesn't like map-lists or something like that? gonna remove all of the non-bungee stuff from the config and see if that changes anything xD
Yea, having a look at the code and it basically looks like that's what it's upset about, looks like it expects everything to either be a Map or a value or some type
Project Waterfall build #276: SUCCESS in 2 min 37 sec: https://papermc.io/ci/job/Waterfall/276/
Project Travertine build #93: SUCCESS in 1 min 11 sec: https://papermc.io/ci/job/Travertine/93/
white lantern
[22:18:42 INFO]: Listening on /0.0.0.0:25577 trying to run a server but it never goespast this
reef fulcrum
That means that it started fine.