thanks, that's what I was looking for. But I don't quite understand how to include it in the code. Should I use some method of Tag.java and pass it as a string?
#waterfall-help
untold zenith
royal hawk
@knotty basin That’s done with full serverside rewriting. Essentially you need to make server B not intrude on what server A has already communicated with the client. It’s so messy and hard to do right that not even md_5 would do it with bungeecord
@untold zenith with the exception of the biomes see
And not the main NBT Tag but the other one that you need
Packet logic:
zinc sapphire
<url> lol
royal hawk
zinc sapphire
if you wrap a url with <> it doesnt do any fancy embeds
so my screen doesnt fill with velocity's, admittedly quite nice, logo
royal hawk
I usually just click remove all embeds after but I’m on mobile right now
zinc sapphire
ah fair enough
untold zenith
<@!710888431483486318> with the exception of the biomes see
https://github.co...
thanks, I hope it helps
untold zenith
<@!710888431483486318> with the exception of the biomes see
https://github.co...
I don't actually see any json references there. Is it unrelated to https://paste.gg/p/KennyTV/ad46ef552c6443cc808385f060564550 ?
reef fulcrum
it's an NBT structure which can be represented by json
untold zenith
It's quite a large structure to write it with code. Is there a way to just parse the json or convert it to nbt and parse nbt?
reef fulcrum
the data is sent from the server
you do not wanna just pump and dump you're own info
royal hawk
You’ll need it as constants either way. For the client to function if you’re emulating a server you need at least one biome with at least one matching dimension and one world name
untold zenith
the data is sent from the server
It's created without the server, but directly in waterfall
royal hawk
You’ll need it as constants either way. For the client to function if you’re emu...
Again, this applies.
untold zenith
You’ll need it as constants either way. For the client to function if you’re emu...
I need it as a constant, right
ok, I look up the links you sent and come again later
but thanks for helping
knotty basin
<@200139178938335232> That’s done with full serverside rewriting. Essentially yo...
Where can my devs start?
royal hawk
Where can my devs start?
I strongly advise against it. You won’t get a lot of help with this either. It’s frowned upon by the community for the ridiculous challenges it poses;
Anyway:
Essentially you’ll have to do one of two things:
1: Implement all the packets relevant to world and entity states and entity tracking on the bungeecord with all versions that you want to use
Then implement a packet tracking and storage logic to remap locations/dimensions to a tracker that you keep on the bungeecord for those things.
The logic has to ensure that nothing overlaps.
If you want truly seamless transitions you’ll also have to rewrite relative coordinates (strongly advise against that)
Bungeecord can do basic entity ID rewriting. You’ll want to hook into that system. You’ll also need to hook into the server switching logic to do it.
2: Implement a tracker for all of this on bungeecord and use plugin messaging to the downstream servers to communicate the tracked values. You’ll have to then use protocollib on the downstream servers to do what you’d do on the bungeecord in terms of remapping or rewriting.
Upside: you don’t have to implement the specific packets/versions from scratch
If that all sounds ridiculous- It is. It’s a philosophical question in the end. It’s very hard to maintain and hard to do right. And if you’re paying devs it’s gonna be very expensive if it’s done right
marble kelpBOT
Salute
Just built waterfall on freebsd 12.2-stable
But I got an error running original-Waterfall.jar
Oh
I don't remember that there
Spooky
Nevermind
It's supposed to be Waterfall.jar, I don't know how I didn't see it
hollow grove
Hey, is there any way to make a network with both a PaperMC 1.16.5 and a Forge 1.16.5 with some kind of server switching? Or would I have to stick with 1.12.2 since SpongeForge hasn't been updated yet?
reef fulcrum
velocity was working on something, no idea if it's even close to remotely stable but 🤷♂️
hollow grove
Hmm. Forgot to add the vanilla server is running Geyser. It's a nightmare of a setup, really
So until some mix of Forge and Spigot update to 1.16.5, that's a no go?
reef fulcrum
forge obliterated the means for server switching, velocity was working on something but it was still a WIP
reef fulcrum
via doesn't support mods
hollow grove
Oh, I didn't know that. Seems to me I'm SOL
Unless I host it out of the network, but then the proxy isn't particularly necessary now
brazen flume
I have a question about Waterfall. It seems that my network only sees how many players there are on the primary server (the lobby). How can I make it so it sees all players in total on the server lists etc, instead of just the lobby?
My current plugins are: https://file.coffee/u/Z4Wk4dOZIy.png could it be that one of them conflicts with it?
round ocean
So /glist doesn't work for you?
marble kelpBOT
disable ping passthrough
silver walrus
First time Waterfall user here.
I see that there is the Priority server, where you can set where people will first go, but how do you set up a "Fallback Server". Meaning, if the server someone is playing on goes down, it auto reconnects to the lobby? Right now, it just disconnects my players. They also can only connect through the proxy, no direct connections.
marble kelpBOT
you would need a plugin for that
e.g. I made one myself to handle that based on the priorities order: https://github.com/Minebench/KickInfo/
silver walrus
ooh, ok. I thought that was built in. Thank you!
trim pawn
im quite confused about what waterfall is, on the paper download page it says it has improved forge support, but the getting started page doesn't say anything like that
marble kelpBOT
it's a bungeecord fork
trim pawn
I'm running mohist right now and it doesn't really work and is kind of bad
what's bungeecord
trim pawn
yeah i googled it it's not like uh that obvious
marble kelpBOT
that wouldn't have provided a proper answer anyways, lol
trim pawn
does waterfall have all the normal paper features or is it separate
marble kelpBOT
it's a proxy software to connect multiple servers and allow switching between them
trim pawn
what's something for running a modded server no plugins
server modding is giving me a headache
trim pawn
just forge
marble kelpBOT
lmao
trim pawn
just forge mods
I wanted both but that seems like it's not a thing for 1.16.5
or like ~forge mods + backups that's it
im gonna try just a forge server and pray that it works
thanks for your help
gritty kestrel
after a high ping the player get "timed out" and is not kicked from Waterfall and player get "you already connected tot this proxy" when he want to connect again. Any ideea about this?
mystic sky
My server hit 90 online today players got 4K ping. What can I do to know the cause of the issue ?
Do I need a good cpu?
Is there any overload message like there is on spigot
delicate phoenix
Is there any overload message like there is on spigot
No
delicate phoenix
Do I need a good cpu?
Ideally yes.
delicate phoenix
My server hit 90 online today players got 4K ping. What can I do to know the cau...
You can use a plugin that listens for packets, and generally, if you have os access you can do stuff there as well.
terse cedar
how can i add a new server server to my waterfall like redirect to it with /server minigame for example
delicate phoenix
Please go to the root folder of your proxy (where you have the jar in most cases), and open the config.yml file
Under servers: you can add/remove/manage servers
terse cedar
how to add default server and none default server
also how can i dont let players join to ports?
just come from 25577
terse cedar
ty
delicate phoenix
Cloudflare wat are you doing
delicate phoenix
also how can i dont let players join to ports?
Use a firewall like ufw or iptables
Actually they are the same 🤔
terse cedar
ty but sorry i ask another question how can i use my dns connected to my website to redirect subdomain to waterfall?
for example mc.example.com will redirect to i.p.addr.ess:25577
delicate phoenix
Either use tcpshield or
https://www.namecheap.com/support/knowledgebase/article.aspx/9765/2208/how-can-i-link-my-domain-name-to-a-minecraft-server/
delicate phoenix
ty but sorry i ask another question how can i use my dns connected to my website...
Use forced hosts for that
More info can be found in the bungee config guide
terse cedar
Use forced hosts for that
i used it
changed it to
but
it does not work
it need premission frrom domain provider?
delicate phoenix
No, it should work.
terse cedar
wait let me send the forced host
forced_hosts:
mc.domain.com: lobby
listeners:
- query_port: 25577
motd: '&1Another Bungee server'
tab_list: GLOBAL_PING
query_enabled: false
proxy_protocol: false
forced_hosts:
mc.domain.com: lobby
ping_passthrough: false
priorities:
- lobby
bind_local_address: true
host: 0.0.0.0:25577
max_players: 12
tab_size: 60
force_default_server: false
remote_ping_cache: -1
network_compression_threshold: 256
permissions:
default:
- bungeecord.command.server
- bungeecord.command.list
admin:
- bungeecord.command.alert
- bungeecord.command.end
- bungeecord.command.ip
- bungeecord.command.reload
log_pings: true
connection_throttle_limit: 3
server_connect_timeout: 5000
timeout: 30000
stats: 8fc38b5c-670d-469a-b3b0-fe6e82c30eef
player_limit: -1
ip_forward: false
groups:
md_5:
- admin
remote_ping_timeout: 5000
connection_throttle: 4000
log_commands: false
prevent_proxy_connections: false
online_mode: false
forge_support: true
disabled_commands:
- disabledcommandhere
servers:
lobby:
motd: '&1Just another Waterfall - Forced Host'
address: localhost:25565
restricted: false
Set ip foward, and online mode to true. As a server, it's better to specify your adress under host: properly.
terse cedar
i also connected my domain name servers to the server
delicate phoenix
0.0.0.0 means all interfaces
terse cedar
ok ty
delicate phoenix
And, at the backend servers, set bungee mode to on, and stuff in spigot.yml
terse cedar
And, at the backend servers, set bungee mode to on, and stuff in spigot.yml
set bungee mode to on? on spigot.yaml
delicate phoenix
Set it to true.
terse cedar
ok
delicate phoenix
And in the server propertes file, specify the server-ip: and port as the same as in the bungee config
delicate phoenix
https://www.spigotmc.org/wiki/bungeecord/ make sure to read these treads 👀
delicate phoenix
No.
If running on the same machine, the two ips will be the same so technicly yes. But host in the bungeecord config.yml file is the adress your bungee will bind itself to.
In the server propertes file, the ip you specify will be the one your backend server will bind itself to
terse cedar
but i dont find any ip address
delicate phoenix
Do you have os lvl access?
terse cedar
in bungee config
delicate phoenix
Run ip a
terse cedar
[19:21:16 INFO]: That user is not online.
delicate phoenix
in bungee config
Your ip that you set in the backend servers server properties ,must match the one you set under servers.
delicate phoenix
Yes. ip a is a linux command used to indentify the machines public adresses.
terse cedar
my server domain ip ok
delicate phoenix
Uh sorry, it's getting late here, so I'm heading to bed rn. Other people from here like @nocturne fog will help you I'm pretty sure! Good luck with the configs.
terse cedar
ty bye good night
delicate phoenix
See you 👋
faint elk
Is there a new URL to wget the latest version? I'm getting a 403 error on https://papermc.io/ci/job/Waterfall/lastSuccessfulBuild/artifact/Waterfall-Proxy/bootstrap/target/Waterfall.jar
young rampart
check the pins in #announcements
faint elk
So now I have to be a programmer to download from a headless server
Guess I'll just do it manually and upload
maiden fox
is there a way to get connected back to my lobby server when i restart the skyblock ?
when im restarting its disconnecting me
maiden fox
nvm i found a plugin for this
zealous pelican
whenever i run my waterfall .jar with sudo java -Xmx1024M -Xms1024M -jar server.jar nogui (my waterfall is named server.jar) it shows this and i have another server running on the same vps i just dont know how to fix this
however it created some files
delicate phoenix
Adress allready in use
zealous pelican
yes
zealous pelican
i don't know how to fix it already being in use
reef fulcrum
something is already running on that port
reef fulcrum
find out what's running on that ip/port
zealous pelican
no i know what is
reef fulcrum
No, you don't need a seperate system
reef fulcrum
99% chance you wanna run bungee on 25565 anyways
marble kelpBOT
run netstat tulpn | grep '25577' that will show you what's on that port
but yeah, why do you not run that on 25565 
delicate phoenix
Phoenix was faster than me 😔
marble kelpBOT
I had to use that command way too often 🙈
delicate phoenix
Networking issues 
zealous pelican
but it doesn't give me a server.properties, should i use the config.yml instead?
reef fulcrum
ctrl-f "host"
reef fulcrum
there is a config option called host, iirc
which literally has the ip:port assignment
reef fulcrum
I'd heavily suggest reading the install guide
zealous pelican
ok
reef fulcrum
bungee/waterfall is not something you should be messing with unless you have some understanding of networking, etc
marble kelpBOT
"host" is clearly in there
slow oriole
Hi Guys, this is weird started about 10 Days ago, but it is starting to become a major issue, I suspect that there is some network or internet instability happening, at which point water fall looses track of things - my best guess so far.
so far the only thing that resolves it is to restart the whole mc network
this is an issue with 200+ players and 12 servers
but we are accumulating ghosts at a rapid rate
reef fulcrum
99% of issues like that are caused by plugins
everytime I've wasted time tryna help diagnose some issue where it seems players are leaking, it's always been down to a plugin messing around with internals
slow oriole
weird because if anything we have removed plugins - but it started before then - as far as bungee/waterfall plugins are stable, we have tried updating
What I have seen though is along this period there have been regional net issues
reef fulcrum
yea, it's generally some network issue which causes some derpiness which some plugin hooking into generally doesn't deal with too well and causes a player ref to stick around
slow oriole
not saying nope - saying this is what i have seen
reef fulcrum
am guessing that you're also in offline mode, which is generally around where the plugins cause issues
at least with online mode the old connection would be booted if one was caught dangling
slow oriole
yeah - cracked
is there any way to cause a resync without restart (doubt but worth an ask)
reef fulcrum
with a plugin, sure, you could maybe force cleanup, as for how you detect that stuff when something is breaking that, god knows
slow oriole
I not sure even he does
reef fulcrum
the proxy has timeout handlers so even in worse case the connection would time out after like 30 seconds and be disposed of
slow oriole
😉
reef fulcrum
Basically, you're tryna work around issues something you've installed are likely causing
slow oriole
or hasnt been an issue and with instability has become an issue
which makes it even harder
other than players ghosting first sign of the issue was redis between 2 bungee servers not syncing
ok that gives another point to look at - plugins that sync on bungee
any suggestions - ie usual suspects amongst plugins?
wispy relic
[16:32:03] [Netty Worker IO Thread #4/WARN] [io.netty.util.concurrent.AbstractEventExecutor]: A task raised an exception. Task: net.md_5.bungee.connection.InitialHandler$6$1@7f1207f1
java.lang.IndexOutOfBoundsException: Index 0 out of bounds for length 0
at jdk.internal.util.Preconditions.outOfBounds(Preconditions.java:64) ~[?:?]
at jdk.internal.util.Preconditions.outOfBoundsCheckIndex(Preconditions.java:70) ~[?:?]
at jdk.internal.util.Preconditions.checkIndex(Preconditions.java:248) ~[?:?]
at java.util.Objects.checkIndex(Objects.java:372) ~[?:?]
at java.util.ArrayList.get(ArrayList.java:459) ~[?:?]
at net.md_5.bungee.api.config.ListenerInfo.getDefaultServer(ListenerInfo.java:91) ~[bungee.jar:git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:2b1ac19:398]
at net.md_5.bungee.connection.InitialHandler$6$1.run(InitialHandler.java:566) ~[bungee.jar:git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:2b1ac19:398]
at io.netty.util.concurrent.AbstractEventExecutor.safeExecute(AbstractEventExecutor.java:164) ~[bungee.jar:git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:2b1ac19:398]
at io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:472) ~[bungee.jar:git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:2b1ac19:398]
at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:497) ~[bungee.jar:git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:2b1ac19:398]
at io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:989) ~[bungee.jar:git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:2b1ac19:398]
at io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74) ~[bungee.jar:git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:2b1ac19:398]
at java.lang.Thread.run(Thread.java:834) [?:?]
Im recently getting this error
Anyone know why?
reef fulcrum
empty priorities list?
wispy relic
Wait
wispy relic
empty priorities list?
in config.yml?
reef fulcrum
I mean, if it was empty on startup, it should fail to run iirc
that would more maybe point to a plugin modifying it
wispy relic
im using CloudNet but i dont know if its a Waterfall error or CloudNet error
reef fulcrum
probably cloudnet
yea, that list can't be empty normally, if it is empty bungee just auto-adds lobby into there
exception is because something emptied a list which shouldn't be empty
wispy relic
so its a cloudnet error?
reef fulcrum
maybe, probably 🤷♂️
jaunty crag
Is this pretty typical cpu usage for waterfall with ~100 players on, about 60 or so of them in a big pvp fight? top server is the survival server, bottom is proxy
I didn't think proxy usually ate that much cpu
is 512mb ram too little? Causing any issues rather?
zinc raven
https://paste.gg/p/Walrus2020/6327c508d86846ef8a2ad333d7eadfac
BUMP anyone know why?
reef fulcrum
anyone know why what?
you can't just dump the entire log on the channel and expect us to magically know what you're asking about
zinc raven
It crashed
vast crow
Hello, im running a normal vanilla but with paperspigot. The server continues "crashing" ||(Can't keep up! Is the server overloaded? Running 5496ms or 109 ticks behind)||
Or creates threads dump, how can i resolve all this problem?
I run a 1GB server and have only 2 plugins...
reef fulcrum
we've not been called paperspigot in years
basically, paper does more at once, stuff like chunk loading, etc, is more parallel, means that smaller heaps are muuuch more of an issue on paper vs vanilla, for example
reducing view distance, etc, if it's memory related can help, but, in general, you need more ram
you'd wanna use timings to see what's taking too long, but if it's underpowered hardware, vanilla being a bit slower around a few things can be beneficial for reducing the cpu load
vast crow
you'd wanna use timings to see what's taking too long, but if it's underpowered ...
https://pastebin.com/aJ2gXf62 so this error is caused for poor ram?
reef fulcrum
Please don't ping, your server was so overloaded that it took too long to even dump the "we're stalling" message on time
I'd suggest using timings
./timings report
(also, #paper-help for paper help)
vast crow
f*ck i thought i was on paper help, sorry
mystic sky
How can I know that my waterfall server is overloaded or needs more ram?
delicate phoenix
If you experience issues like oom with large playercounts than that's a sign
reef fulcrum
attaching a profiler is only real approach to monitoring that stuff
gaunt linden
I want to make a broadcast/alert command in my waterfall plugin
i used this to check if the player has the right permission
(I'm using luckperms to manage the permissions)
if (proxiedPlayer.hasPermission(("example.permission")) {
//Rest of the code
}
But when i want to perform the command it says: i dont have enough permissions
Can somebody help with that problem?
reef fulcrum
hasPermission basically just fires an event for plugins to control, you'd need to see if LP has those perms setup properly, lpb command for only the proxy unless networking is setup properly
slow ermine
is this issue related to bungeecord setting or a plugin issue?
[05:31:10 INFO]: [Cipetvzon7] disconnected with: The server you were previously on went down, you have been connected to a fallback server```
all of my player are having the issue connecting
tidal musk
hi there i got a bug to report but maybe it is not a bug but im not totally sure... but i wanted to change the server icon and removed it but the old icon is still there on the proxy while it isnt.
reef fulcrum
if you're using ping passthrough it will ask the server for all that stuff, otherwise it's in the waterfall folder
the client is also generally known for not being great at updating those icons
tidal musk
ah and how do i fix this? or should i check the waterfall folder first
reef fulcrum
check waterfall first
tidal musk
so the jar folder right
reef fulcrum
the folder where the proxy files are
tidal musk
so the main directory
reef fulcrum
yes
tidal musk
its not there
reef fulcrum
did you restart the server?
tidal musk
yes
reef fulcrum
best guess is try hitting refresh in the client a few times
tidal musk
ive done that already tho
but i asked the help support over at my server hosting and they told me that waterfall stores pngs somewhere or something with a plugin similiar to bstats
altho i got these 2 folders and dont know that they are
reef fulcrum
if you had it installed on the proxy the proxy might cache it, I forget if the proxy does that
tidal musk
if you had it installed on the proxy the proxy might cache it, I forget if the p...
that yes, correct
reef fulcrum
restarting or reloading (maybe) would clear the cache
tidal musk
alright ill try ok
tidal musk
restarting or reloading (maybe) would clear the cache
the icon is still there sadly...
reef fulcrum
do you have a server-icon.png?
if not, afaik it doesn't generally care to drop the old copy of it
tidal musk
yes i have a server-icon.png but its not in the main directory
also when i put one in the old icon will still be there
reef fulcrum
has to be in the main directory if you want it to read it
put it back and greload/restart
tidal musk
ok btw also inside the waterfall.zip folders is this:
reef fulcrum
I think that the server has a cooldown on how often the icon will be sent, so it caches it locally
tidal musk
reef fulcrum
and god knows, nothing to do with waterfall itself
tidal musk
and god knows, nothing to do with waterfall itself
do you know what that means in the picture?
reef fulcrum
No
it's a zip file, you'd need to extract it to see what's in it, but, that's not waterfall doing that
tidal musk
oh ok
so where can i enable restart cache or something
because restarting the proxy wont help
reef fulcrum
if the new file is in place and you've not configured ping passthrough or some other plugin, it should work, might need to refresh a few times
tidal musk
so if i drop in the new icon right now and do all that what u just said, will the new icon load even tho the old one shows up in the server list?
reef fulcrum
should do eventually
tidal musk
do you know which file has ''ping passtrough'' and what does this feature do?
tidal musk
ill try it
it seems its also put on false so
welp i enabled it but it didnt work either
reef fulcrum
you'd want it false if you don't want the server to have the image
tidal musk
it was on false before
tidal musk
you'd want it false if you don't want the server to have the image
the old icon is still loading while its on false
brave ether
Is there a lib similar to the Paper lib to make plugins work with Waterfall (Using its features) while being still compatible with BungeeCord?
Also, any changes to what Waterfall uses for Chat (Like does it still only use BaseComponent or does it use Kyori?)
young rampart
Waterfall does not use adventure (yet)
tidal musk
zzzCat i fixed the issue, there was a icon folder in the MOTD plugin.
brave ether
So is there a lib for Waterfall similar to where there is one for paper?
In terms of being compatible with Bungee
reef fulcrum
Nope
am not even sure how it would bridge some of the API stuff, am not even sure exactly what API stuff would concern people and if that was even viable to throw some compat workaround lib together
brave ether
ok
brave ether
Does Waterfall have any noticable differences compared to BungeeCord in terms of stuff like PreLoginEvent and ProxyPingEvent?
delicate phoenix
Waterfall has a few improvments yes, but if you are looking for something feature rich, check Velocity
slow ermine
[10:17:50 INFO]: [JustzZ] disconnected with: ReadTimeoutException : null```
what is this error?some of my player cant conenct to serverserver took too long to respond
slow ermine
any where i could increase the time before it kick player?
next birch
```[10:17:50 WARN]: [/*.*.*.*|JustzZ] -> UpstreamBridge - read timed out
[10:17:...
config problem
are you using restrict?
next birch
any where i could increase the time before it kick player?
not sure there are such function on waterfall but i know there are such function on server.properties
slow ermine
are you using restrict?
restrict is set to false
next birch
all the server?
slow ermine
yes
next birch
and make sure your ip type correctly and the server are online
slow ermine
all is correct
next birch
better to put the higher priority that the server online
slow ermine
but my i set my prior to hub
next birch
any plugin installed?
slow ermine
i have tried removing all the plugin,but still it doesnt work
marble kelpBOT
is it only happening to some players? if so then they have network/connection issues
slow ermine
that weird...
because when he join other network
it has no problem
like i have multiple server setup
survival,vanilla,skyblock and faction
he cant join survival
but the other server has no problem conencting
marble kelpBOT
are they all on the same machine?
slow ermine
yea
next birch
the hub server.properties ip is 127.0.0.1?
marble kelpBOT
that has nothing to do with his issue
next birch
or what you type in the config on waterfall?i mean the hub
marble kelpBOT
it's possible that they are logging into a place where the chunks have tons of data, had a similar issue before. iirc you could try disabling the network compression on the paper server if you haven't done that already
next birch
127.0.0.1 or localhost or what?
next birch
it's possible that they are logging into a place where the chunks have tons of d...
yes that maybe
or the console of the hub said the player have login?
marble kelpBOT
knowing the exception that the server throws on disconnect would also be helpful
slow ermine
it's possible that they are logging into a place where the chunks have tons of d...
in where?
bungee config?
marble kelpBOT
no, in the configs of your paper server
bungee's config only deals with the proxy -> client connection, not server -> proxy
marble kelpBOT
yes, seeing as that's already quite high I would just leave it
tidal musk
Pls GIVE me an tutorial how to use waterfall
royal hawk
Read or watch any tutorial on how to use bungeecord
reef fulcrum
should be safe to ignore iirc
austere viper
but waterfall doesnt do anything 😢
reef fulcrum
ouch, erm, last I recall it was generally down to misconfigured VPSes
austere viper
sad
reef fulcrum
looks like log4j added a workaround for such broken systems a while ago
reef fulcrum
wait for the next build of waterfall
reef fulcrum
somethings wrong with the build server stuff
use travertine in the meantime, that built fine for some reason
austere viper
okay thanks 🙂
slow ermine
Do I have to use travertine for 1.7 support??
Because I have viarewind install,people are still able to use 1.8 to connect
austere viper
travertine didnt work 😢
reef fulcrum
define 'didn't work'
austere viper
the same error
reef fulcrum
if that was build 175, am not sure
somebody would need to find a way to reproduce it and report it to l0g4j
what's the mac address? 🤷♂️
austere viper
00-15-5D-00-64-10
reef fulcrum
on the surface looks fine
bungee command output?
Oh, wait, no, full startup log?
.paste
marble kelpBOT
Please paste large logs to a pastebin: https://paste.gg
reef fulcrum
in the logs folder for waterfall
austere viper
didnt get a log folder :/
reef fulcrum
try just copying the full output from waterfall/travertine
reef fulcrum
I said 175, not 174
reef fulcrum
I then told you to try travertine, I even said, build 175
and you're here showing me build 174
reef fulcrum
Please, read and stop wasting my time.
austere viper
175
reef fulcrum
refresh the page
austere viper
i did
reef fulcrum
force refresh
austere viper
its still 174
reef fulcrum
175 is 100% on the downloads page, you're likely hitting some oddball caching or something
austere viper
its there now
reef fulcrum
if you refreshed bypassing the cache, it should be there unless cloudflare is taking a moment
crimson epoch
If I want people to be able to join my local network do I portforward all of the servers or just the bungee?
reef fulcrum
just the proxy
crimson epoch
Alright, I opened the port to my proxy (25577) and I went to these two websites that check if a certain port is open. They both said 25577 was open on my IP. What's weird is that when I try to connect to the proxy MYIP:25577 it doesnt let me. Bungeecord is enabled in my spigot.yml settings, ip-forward is enabled, servers are offline, and proxy is online. Do you know what the issue may be?
reef fulcrum
use an status testing website
not all routers allow reverse NAT (connecting to public IP inside of the network)
crisp folio
Are you able to have the servers attached to bungee/waterfall on a different network to the bungee server its self? If the bungee is on a hosting site
somber geode
Does waterfall have a command for manually creating heap data report? Kinda like /paper heap
reef fulcrum
nope
somber geode
Darn, thanks anyway!
warm bramble
Can I have servers on a different hosting site than each other but still connect them with bungee/waterfall?
Thank you!
reef fulcrum
yes, it just connects over the network
warm bramble
Oh thanks!
reef fulcrum
only caveat is that you add more latency
warm bramble
That’s gonna be rlly useful
reef fulcrum
so, ideally aim to host in the same place
warm bramble
Yea
I do for my proxy and 3 other servers (All I have right now) they have good service and servers etc just one downside they cost a bit more than the others.
So I was looking to host my main ones on them and others on a cheap one.
Tysm!
celest solar
did anyone ping me
hollow prawn
How would I get an automatic download of Waterfall when there's a new build?
royal hawk
Make a script to query the API
hollow prawn
Ok, not sure how to do that, but thanks for the tip.
royal hawk
vagrant lily
I have some problems connecting to the server
royal hawk
Network or firewall issue on your side
slow ermine
network_compression_threshold: 256
if i change this to 512 in bungeecord config will it be benefit for player that are having high ping?
royal hawk
No, quite the opposite as you’re sending more data that way
vagrant lily
can I have multiple lobbys at the same time? for example my lobby a is down and users would connect with lobby b
reef fulcrum
yes, priorities list
vagrant lily
oh ok thanks
btw what is the restricted option on the servers catergory? what does it do?
reef fulcrum
needs a perm for /server, please read the docs
bungee's config guide goes over all of what you've asked
reef fulcrum
?warn @tidal musk not the place to advertise
rotund valleyBOT
Sambor99#0045 has been warned. || not the place to advertise
still warren
hollow grove
I think someone's being kicked from servers over a timeout, they're on Java, any setting I can increase to keep it from happening? It's fine on my end so I can't replicate
I'm getting "disconnected with: The server you were previously on went down, you have been connected to a fallback server" but I'm on the server, it's fine
hollow grove
I've already set connection-throttle: -1 on the backend server's config, I've increased the timeout on the proxy, no matter what, they're getting booted. Managed to play for hours yesterday, today can't stay on longer than 3 seconds
errant pivot
how does redis bungee works? anyone knows?
warm violet
is it easy to make one .bat file to start a whole set of servers?
bright crater
Yeah
The way I would do it is have one bat script that starts x amount of bat scripts
x bat scripts for x amount of servers
If theres a way to run a .jar file in a new terminal window, not the existing one, then I guess you could run all the java comands to start the server in one file
bright crater
how does redis bungee works? anyone knows?
are you looking for a simple or comprehensive answer
reef fulcrum
grab the messages file from inside the jar and throw it in the waterfall folder
humble thistle
throw it in the waterfall folder?
reef fulcrum
yes?
humble thistle
that is, do you say to change the messages of the .jar file every time I update it? Is there no way that even if I update it, it will keep what has been translated?
crisp folio
Is it possible to have a bungeecord attached to another bungeecord on two different networks (without any weird stuff setup, one needs to work on mc host)?
crisp folio
bungee IN bungee, no
could you have a set of spigot servers on one network and a bungee on the other?
reef fulcrum
yes
it connects over the network
so long as there is a network connection between the proxy and your server it will work
not that you should be introducing latency into such a mess, but, you can
heady wave
guys i'm having this issue
io.netty.channel.AbstractChannel$AnnotatedConnectException
how to fix?
heady wave
I have some problems connecting to the server
just like him
reef fulcrum
generic network issue
heady wave
so is hosting related?
reef fulcrum
I can't say
I generally refuse to provide support for such issues
literally all that tells us is that the client failed to connect to the server
that's not even the remotely useful part of the exception, so, you've given us nothing to go off at all
heady wave
i've tried to do a rollback to travertine 172 (that worked fine since last night) but same error
reef fulcrum
nothing to do with travertine
(and this is why I don't provide support for this stuff)
heady wave
well as usual i was an idiot, the default fallback server was offline so....
tidal musk
what java version should I use 8 or 11?
reef fulcrum
11
tidal musk
k thanks
stray stag
hello all. can anyone assist me with default fallback server issue
i think i have everything here correct however when the server restarts it says cant connect to fallback server
i have the force host for the lobby, and the priorities of lobby
tidal musk
I think yes
severe badger
set online-mode: false in the server.properties and enable bungeecord in the spigot.yml
slender terrace
we need more people like you in the world lol
but no for the proxy you need the backend servers to be in offline mode
the proxy handles the authentication to make sure people have legit accounts
warm violet
ok
warm violet
we need more people like you in the world lol
#no-more-cracked-accounts!!!!!!!!!!
slender terrace
yeah just make sure the proxy is set up correctly though
slender terrace
what NotMyFault said
reef fulcrum
invalid configuration
warm violet
witch config file?
reef fulcrum
config.yml
warm violet
so... do i delete the file and start again?
reef fulcrum
remove 63, move survival into the servers section properly
warm violet
ok 🙂
brittle siren
Anyone want to share tcpshield?
marble kelpBOT
wat
vagrant vortex
it's literally free
delicate phoenix
If you are offering a part share of the company sure
fossil bone
hello i have a problem with my server I can no longer take a head, a banner etc it gives me a head of Steve
if you look in my inventory we see the head appear and become steve's head just after
Version 1.15.2 + waterfall
royal hawk
Something something disable tab list rewriting and ensure you have forwarding set up correctly
royal hawk
.paste your bungeecord config
marble kelpBOT
Please paste large logs to a pastebin: https://paste.gg
topaz mountain
Hi! ^-^
I'll go straight to the point: https://pastebin.com/PeiVnriH
No clue on what caused it, help me please c-c
Oh, yes, even if calling "ModernSurvival" the server could look egocentric due to my nickname, my nickname and my network's name are not completly linked dx </ot>
tidal musk
Good evening,
Will there be an update of waterfall in the next few weeks to make it compatible with forge 1.16 because it's really a pain since the proxy doesn't work with it? 😭
reef fulcrum
Velocity was working on it, but, experimental WIP at best last I knew
tidal musk
Thanks, but the big problem with velocity is that most of the plugins are not compatible.
reef fulcrum
They're doing work with forge to make it possible to do it in the first place
Once velocity had the technique, we can adopt it
Otherwise, work on it yourself and submit a pr
tidal musk
Thank you for the information, and I think I'll wait for it to be developed. xD
woven stone
Hello, here I am again. I have a question about the following screenshot. Can I edit this to say something else? Talking about the "Waterfall + version" text.
reef fulcrum
you can change the version numbers in waterfall.yml, can't modify attribution however (without a plugin)
woven stone
Oh, ok. Does this have a specific name so I can find a plugin? 😛
tidal musk
Because of what there may be an error in which the passenger of the boat is thrown to the original place where he sit in it? At the same time, there is no driver.
delicate ermine
does anyone know any bungeecord (waterfall) manager plugins on spigot? (for booting up hub servers when needed, or booting another minigame server when asked for)
restive wraith
Does anyone know why it says: Kicked while connecting to AUTH: Accelerated connection! Please wait before reconnecting.
strong jay
Only thing I could suggest is: Is the connection throttle set to -1 in your servers and on bungee? Other than that idk how to fix that.
tidal musk
is waterfall exactly like bungee but optimized? just like paper for spigot?
mild rover
yes
mental pawn
what versions does waterfall support?
nocturne fog
1.8-1.16
nocturne fog
correct
mental pawn
thanks
tidal musk
hello i have some big issues
from some reason unknown to me, i can't send the players from lobby to another server using any other plugin except the bungee commands written manually
for example I do have a lobby plugin (hub plugin) with a compass which executes some specific commands to send players form lobby to server A, B,C
reef fulcrum
bungee commands are bungee commands
those commands are not defined on the server and cannot be ran on the server
tidal musk
however it just stopped working completely, but it's not the plugins fault because 3 plugins stopped working with bungee commands
I mean, you know the command /server survival?
reef fulcrum
you'd need some plugin which can use the messaging channel to send a plugin message to the proxy to do the siwtch
tidal musk
or any /server command at all
reef fulcrum
yea, that's a proxy command, it works solely because the proxy intercepts it
tidal musk
yep, so let's say that
if the player does /server survival
it works
if he clicks an NPC which does /server survival
it sends a message in console
reef fulcrum
I already said
/server is not a server command
the server does not know that such a command even exists
you need a plugin which is able to send a plugin message to the proxy instead of tryna run a command which doesn't exist
reef fulcrum
if you got some plugin which simulates the server command or something, you'd need to take it up with them
make sure that the server name matches what's in your proxies config.yml
tidal musk
make sure that the server name matches what's in your proxies config.yml
it does indeed, re-checked everything 100 of times
but i used also 3 different plugins
Citizens + CommandNPC
DeluxeHub
Skript Hub
and a couple more
all have the same issue
reef fulcrum
I mean, we know that the plugin messaging channel works just fine and we're pretty sure we'd have a huge line of people screaming if it didn't
that would imply that you've got something misconfigured somewhere or some plugin on the proxy is saying nope or something
check your proxy logs I'd guess
tidal musk
it has 0 errors as well in the console, I verified that as well.
Perhaps if you tell me a good plugin which sends a plugin message to the proxy I can figure something out
tidal musk
Can I show you my config at least so you can see if something is misconfigured?
reef fulcrum
i mean, only useful thing in there would be the server names, which if /server works, implies that the proxy is working fine
tidal musk
yep, but for some reason it has issues to work together with all the other plugins
and i guess it's only something related to my server but it's a very big issue that i can't figure out on how to fix
reef fulcrum
What am saying is is that if /server works, the proxies config.yml is fine
tidal musk
I also tried actually uninstalling the plugins which use those commands (Citizens and CommandNPC)
and then i configured them from 0 again
it worked for some time, then stopped working again
talking about some minutes here
reef fulcrum
there's nothing in the proxy itself that would cause some time limit to stuff working, especially not without some form of messages, etc
do you have plugins on the proxy? if so, try without them
maybe try without any other plugins on the server itself too
tidal musk
i only have 4 plugins in the proxy which are nuvotifier, epicguard, skinsrestorer and bungeeguard
all of those worked perfectly for 2 months
perhaps it's an issue related to the lobby itself only
reef fulcrum
I mean, you may only have 4 plugins, but, one of those is known to create massive joyous issues on occasion
if you can't or are unwilling to replicate without other plugins, I can't help you
tidal musk
which one specifically?
anyways, seems like the issue is with my lobby specifically not being able to send players in other servers
i'll try to uninstall all plugins
and make it from scratch
magic shard
@elfin moth "java -Xms512M -Xmx512M -XX:+UseG1GC -XX:G1HeapRegionSize=4M -XX:+UnlockExperimentalVMOptions -XX:+ParallelRefProcEnabled -XX:+AlwaysPreTouch -jar waterfall.jar"
elfin moth
<@!684635192987877385> "java -Xms512M -Xmx512M -XX:+UseG1GC -XX:G1HeapRegionSize...
ty
is there a more recent tutorial on how to configure waterfall?
ionic wigeon
is there a more recent tutorial on how to configure waterfall?
Follow any bungee guide, just using the waterfall jar instead
elfin moth
is it right to set the port host of the waterfall to 25565?
nocturne fog
if you want it to run on 25565, sure.
there isn't a right or wrong port. Depends what you want. 25565 is however the default
elfin moth
if you want it to run on 25565, sure.
if i dont run it on 25565 tho, people would have to join <ip>:<port> right? instead of just <ip>
nocturne fog
Correct
stray stag
hello all wondering if someone can take a look at my config and see if everything looks good. it seems when server restarts it just says couldn't coonect to fallback server
dire plank
How can I make
so when i restart one of my servers
they get forced move to the hub
is that via a plugin or the config?
reef fulcrum
plugin
dire plank
Do you know which one?
reef fulcrum
don't remember, try using the search
elfin moth
what should i put here?vb forced_hosts: pvp.md-5.net: pvp
if im portforwarding off my of my machine
and all servers are on the same machine
how can i leave this area blank?
reef fulcrum
leave it alone, read the docs
stuck nacelle
Hey, I have issue with waterfall performance on newest version. It freezes on 100% of one core every couple of minutes, then usually goes to 100% all cores, and back to normal. It causes massive lag on server
What might be the issue
On previous version I didn't have this 100% one core issue
stuck nacelle
IS anyone here?
From what I see there is issue with GC. However I don't know what to do with it. Can anyone please help?
stuck nacelle
Anyone familiar with GC? I'm really clueless now
What might be causing this issue
Players are mad
It was working fine, and now it has this issues
wtf
It was working for a long time without issues
And now this
why
I tried different startup flags. It slightly helped, but not completly
tidal musk
How do I make when people connect to my server they don't go in survival server but they got to lobby server?
marble kelpBOT
enable force_default_server and define the lobby as the first in the priorities list
tidal musk
enable force_default_server and define the lobby as the first in the priorities ...
Oh OK thanks
shadow dawn
Waterfall keeps shutting down on its own and im not sure why, this is the last log before it was just killed, any help?
`[03:28:59 INFO]: ["player"] disconnected with: The server is currently under maintenance!
Try again later!
[03:28:59 INFO]: [/ip address|player] -> UpstreamBridge has disconnected
Killed`
its running in an amazon linux 2 instance in lightsail, do their servers kill it?
next birch
Waterfall keeps shutting down on its own and im not sure why, this is the last l...
Is that there are no ram?
shadow dawn
uhh i think your asking about ram. its running with 215mb or ram
the server it self has 512mb
delicate phoenix
Aws is not good for mc, nor for proxyes, but all that error message tells us, it that your server got Killed by something.
Do you have any other errors/error files in the root directory of your server?
If any of them mentions OOM, than yes, allocate more ram.
next birch
Is that there are no ram?
You can try for server.pro(recommended) , ovh(recommended) , apexhosting , or china stuff(aliyun)
meager citrus
Hi, can't you please add way to have more worlds?
Preconditions.checkArgument( worldCount < 128, "Too many worlds %s", worldCount ); https://github.com/SpigotMC/BungeeCord/blob/master/protocol/src/main/java/net/md_5/bungee/protocol/packet/Login.java#L56
I'm using playerworlds plugin and I can't join server then... https://www.spigotmc.org/resources/⭐-playerworldspro-⭐-the-nether-and-the-end-support-expiration-1-8-1-16-4.72113/updates
reef fulcrum
try removing it and see what happens
I can only imagine that that was added for some reason, but 🤷♂️
hollow grove
So more people joining my server are getting " disconnected with: The server you were previously on went down, you have been connected to a fallback server"
The first person to get it fixed it by... Reinstalling the OS and everything from scratch
I'm on the server, other people are on the server, some keep getting kicked
royal hawk
@meager citrus this limit is there because 1.16-1.16.2 will misbehave if there is 128 or more
meager citrus
what do you mean by misbehave? And even if all worlds aren't working at one time?
royal hawk
Yes. Misbehave: If you send a world name that was in the index above 127 it will disconnect you
reef fulcrum
if it works fine on newer versions, can probably just be removed or shoved behind some version checks 🤷♂️
stuck nacelle
Hey, can anyone please help me with GC, and flags?
stuck nacelle
My waterfall usually goes close to 0 free memory, then does intensive GC which causes all cores of Ryzen 7 3700x to go to 100% (causing lag on server) and then it continues to function normally until next GC
Why is that?
Please help me
I've tried multiple things, and it didn't fix it completly
Right now I'm using absurd ammount of ram for 800 players with this flags (but it seems to be the least lagging startup flags)
java -Xms25G -Xmx25G -XX:+UseG1GC -XX:+ParallelRefProcEnabled -XX:MaxGCPauseMillis=200 -XX:+UnlockExperimentalVMOptions -XX:+DisableExplicitGC -XX:+AlwaysPreTouch -XX:G1NewSizePercent=40 -XX:G1MaxNewSizePercent=50 -XX:G1HeapRegionSize=16M -XX:G1ReservePercent=15 -XX:G1HeapWastePercent=5 -XX:G1MixedGCCountTarget=4 -XX:InitiatingHeapOccupancyPercent=20 -XX:G1MixedGCLiveThresholdPercent=90 -XX:G1RSetUpdatingPauseTimePercent=5 -XX:SurvivorRatio=32 -XX:+PerfDisableSharedMem -XX:MaxTenuringThreshold=1 -jar bungee.jar
Before for a long time I was using this startup without any issues, but out of nowwhere (without any of my changes) it become super laggy, with 100% on all cores every couple of minutes
java -Xms12G -Xmx12G -XX:+UseG1GC -XX:G1HeapRegionSize=4M -XX:MaxGCPauseMillis=200 -XX:+UnlockExperimentalVMOptions -XX:+ParallelRefProcEnabled -XX:+AlwaysPreTouch -jar bungee.jar
And this simple startup is just unplayable, with GC lasting for even 11 seconds! Why? 😮
I have to mention that I've used this simple startup for many years without issues before going to slightly more advanced flags. And now it just doesn't want to work
java -Xms10G -Xmx10G -jar bungee.jar
I even rented new dedicated to see if it's not an hardware error, since I havn't changed anything when the issues popped up. I don't know what else I can do. Can anyone help me with finding what might be the issue? Or how can I tune my GC, so it performs bettwe (I'm noob in GC tuning). When not doing GC bungee runs fine, with no lags, and ~200-320% cpu usage on Ryzen 7 3700x. Lags occur only when memory does GC.
Of course right now I'm on newest version of waterfall.
stuck nacelle
@unique linden I know that you’re GC master. Sorry for pinging. But maybe you could help me?
unique linden
see the pinned message
stop using my mc flags for bungee
and ffs dont give bungee 25g
stuck nacelle
I’ve used pinned flags for a long time
unique linden
just give it 4 gb
stuck nacelle
However it stopped working for me
unique linden
define "Stopped working"
how do you even know its GC
but sounds like you got a memory leak
stuck nacelle
Well, it was running fine with no laggs. And few days ago it started to go up to 100% on all cpu cores when freeing memory
I checked it was freeing memory wirh multilobby command, claiming that bungee’s free memory went up
In the moment of freeing memory it laggs
That’s why I think it’s Gc
But im no expert. So what do you think about it?
round ocean
just read.
tidal musk
nvm i got my answer
stuck nacelle
@unique linden ?
Or how can I find memory leak in bungee? Because it might be an issue since ram is eaten up quite fast
Free ram after GC usually goes up to 12GB (out of 25), and it gets usen in about 6-10 minutes (small lag, and then ram gets free again)
scenic pebble
Why are you giving 12GB of RAM to a proxy?
Nevertheless, you should be able to do a heap dump.
stuck nacelle
Well, it needs it
When I was giving it lower values it was unp;layable with more than 400 p;layesrt
calm ether
I use 1GB with 1000+ players, so it has to be with the plugins you are using
scenic pebble
Yeah definitely a problem issue since you don't need that much ram for it.
stuck nacelle
@calm ether You run on newest waterfall? What plugins?
delicate phoenix
Pjo, have you tried using Velocity?
calm ether
I run latest waterfall with only custom plugins
stuck nacelle
Well, currently I have some plugins that are not avaliable on velocity
I'll try switchinmg in future, but right now I need to make waterfall work
nocturne fog
waterfall/velocity is not the issue here
really there isn't much more than can be said other than go get a heap dump
scenic pebble
Yeah, again, make a heap dump.
stuck nacelle
Is there any good tool to do heap dump on waterfall?
scenic pebble
Spark
stuck nacelle
It works on waterfall? 😮
stuck nacelle
Oh, I see it
I wasn't aware
Thansk
I'll try that at night
But is there anyone who can help me analize it? I don't know what is normal, and what isn't. So I might be guessing
Or maybe anyone has correct heapdump that i can compare to
calm ether
There is no "correct" headdump. If you have a problem in your's, it should be easily visible
nocturne fog
yeah it will be super obvious if its doing what you claim
calm ether
I never used spark but with yourkit i search for biggest object dominator
nocturne fog
eclipse memory analyzer is free, and works well enough so you could check it out
but yeah else yourkit or something
spark is just a plugin that can capture heapdumps. Alternative to using jamp or something
calm ether
oh ok
scenic pebble
If you use spark for a heapdump just upload it to heapdump.io or whatever it is and throw a link here and someone can probably look at it with you.
nocturne fog
heaphero I assume
scenic pebble
Yeah
stuck nacelle
Thanks. I'll do it in couple of hours when there will be less players
So I don't annoy them even more XD
nocturne fog
if you don't want to use a plugin you can use jmap (assuming you have access to the machine)
syntax is just jmap <pid>
tidal musk
.optimize
marble kelpBOT
additional tips for improving server performance - https://www.spigotmc.org/posts/3684550
unique crow
Will a 1.16.5 forge server work with waterfall
scenic pebble
Holy shit
Yeah that looks like it alright.
Looks like it's doing something with scoreboard teams but it doesn't even release the objects after it's being done.
At least I can't image it would hold them for any real reason for as long as it seems to be.
It looks like it's storing some data in a map for scoreboards and players and then just not clearing data from it.
stuck nacelle
Okay. So I hope that's the issue. I've messaged dev about it
But it's weird, since it's a popular plugin
Thanks for help
stuck nacelle
Multilobby was patched. It seems that it wasn't directly it's fault. Now waterfall scoreboard handling eats up all my ram. Here is new heapdump.
https://heaphero.io/my-heap-report.jsp?p=YXJjaGl2ZWQvMjAyMS8wMy81Ly0taGVhcC0yMDIxLTAzLTA2XzAwLjQzLjUwLnppcC0yMy01MC02Lmpzb24=
Can anyone experienced look at it?
Is it really internal waterfall issue?
scenic pebble
It's hard to go off just that. All I see here is Bungeecord is handling scoreboards and players.
stuck nacelle
It's eating up ram fast
With 100 players it eats up so much ram
With 800 it eats up whatever I give it
And then there is issue with oldGC from what I see in spark
scenic pebble
Alrighty, well, have you ensured you are on the latest version of Waterfall and using the proper startup flags?
brittle siren
Hi my server is using about 300GB of bandwidth per day? Anyone knows how to reduce this?
Current network-compression threshold config: -1 on backend servers, 256 on bungee.
stuck nacelle
I hope that anyone can help me why waterfall uses so much ram for scoreboard handling, and it's not freeing it
scenic pebble
Well, you never responded to the questions I asked.
stuck nacelle
What questions?
sorry
I've missed them
Yes, im using correct startup flags
And latest waterfall
But I had to add loads of ram to startup flags, since it's unplayable on 1GB
scenic pebble
I think there's something else causing this then. If it was pure Waterfall causing it, I'd assume we'd have more reports and I think you're the only one having something this severe to my knoweledge.
stuck nacelle
Well, it's probably something related what I use
Waterfall is handloing something wrong
But it's still an waterfall issue
It's holding to my scoreboard data without purging it
stuck nacelle
Out of curiosity I started waterfall with no plugins at all
I'll check heapdump soon
There is heapdump with newest waterfall + no plugins. It seems to be issue with waterfall scoreboard handling.
Do you have any opinions?
scenic pebble
What are your startup flags?
stuck nacelle
java -Xms12G -Xmx12G -XX:+UseG1GC -XX:G1HeapRegionSize=4M -XX:MaxGCPauseMillis=200 -XX:+UnlockExperimentalVMOptions -XX:+ParallelRefProcEnabled -XX:+AlwaysPreTouch -jar bungee.jar
scenic pebble
Yeah might have to have someone else give their opinion. I just started up a test env and everything looks normal.
My only thing would be suggesting trying regular bungeecord to see if it still happens cause it might be stemming from upstream but that's up to you.
stuck nacelle
Regular bungee - issue is still there
scenic pebble
Yeah so it's not a Waterfall issue then.
scenic pebble
I'm just a mod, I don't work on the project itself. Now, I think scoreboards with the proxy also communicate to the backend servers so it's possible some kind of scoreboard plugin might be causing it.
And also you're the only person to report having an issue this severe. If it were really happening for everyone, wouldn't you think there'd be more reports?
stuck nacelle
I've disabled scoreboard on every server, and I'm testing now
Well, it's minecraft. And I have big luck for bugs
You think it's working perfectly? Give it to me XD
My dev really loves me for this
Could you present this issue to waterfall devs? I don't want to ping them
It seems to be really an issue inside of waterfall
And not every issue is the same across every server
scenic pebble
Wouldn't it make more sense to go report this to Bungee if it still occurs on Bungee?
stuck nacelle
I guess I can, but isn't waterfall supposed to be patched fork of bungee?
Here is heapdump with every scoreboard related plugin I could find disabled on servers. Still the same issue https://heaphero.io/my-heap-report.jsp?p=YXJjaGl2ZWQvMjAyMS8wMy82Ly0taGVhcC0yMDIxLTAzLTA2XzAyLjAwLjAwLnppcC0xLTMtMTYuanNvbg==
scenic pebble
Again, if it was a more wide-spread issue, I'm sure they would've fixed it by now. You're welcome to open a issue on the GitHub and they will respond to it.
stuck nacelle
Can't they respond here?
stuck nacelle
So I'll wait
naive sky
Hello, I have a server let's call it, HUB, I want players to ALWAYS go to Hub when they join, but when they join for the first time, they go to Hub but then they always go to the server they were in, how to make it so they always go to Hub server first?
scenic pebble
I think there's a setting called force default server
scenic pebble
I want to believe so, correct.
naive sky
It worked, but why does it say "Could not connect to target server....."?
reef fulcrum
Proxy needs to retain scoreboard info in memory for API reasons and for the server switching mechanism
if you wanna patch out scoreboard support, go for it, just, we can't do that without breaking an entire part of the API
royal hawk
@ zzzCat isn’t that hot-wired to be off if entity meta rewriting is too?
royal hawk
How does that affect it?
reef fulcrum
resetting the scoreboard manually might have been removed
plugins on the proxy use the scoreboard stuff in the proxy
it's a horror as I guess you could make it so ones recieved from the server aren't added in a way that it's exposed to the API and can just be passed through
would maybe need to be behind a config option as I got no idea why plugins would shit themselves over that, but, errrr...
flashback to that time he made scoreboards not just return json strings
royal hawk
I’m currently implementing scoreboards for velocity and I’m very much against storing what’s current
I’ll force the plugins to declare that they require this, otherwise it’ll be disabled
reef fulcrum
issue is that plugins rely on the fact that the API stores that stuff for others to get
or even for their own state management essentially
royal hawk
Well I’m setting the precedent for velocity
So I might as well restrict it as much as possible
brazen ruin
How to set up bungeecord fallback server
it never work for me help plz
reef fulcrum
"never work" gives nobody anything to go off
unique crow
Will a 1.16.5 forge server work with waterfall
reef fulcrum
no
unique crow
is it being worked on (modern forge support)
reef fulcrum
velocity was working on it, maybe with intents of getting forge to support it, but, not a priority for me with my health and no intents to conflict with others
royal hawk
@ zzzCat I’m offering to take over maintaining waterfall and travertine if you need a break-
tough island
@ zzzCat I’m offering to take over maintaining waterfall and travertine if you n...
IIRC Don't you work on velocity too?
next birch
.flags
marble kelpBOT
Optimized & Recommended JVM Startup Flags: https://mcflags.emc.gs
royal hawk
IIRC Don't you work on velocity too?
Yup. I’m currently Tux’s right hand man on that
tough island
Working on two competing products 😛
royal hawk
Working on two competing products 😛
In a way yes. I actively develop for velocity however and only do maintenance work for Waterfall
stuck nacelle
Well, so what can I do to get my proxy working fine, and not eating up ram like this? I can’t start proxy with less then 10 GB of ram, because it just crashes
And with big ram ammount it sometimes pause itself for Gc
And causes lag
tough island
what plugins are you using on your proxy... 10gb for the proxy is nuts... Though we haven't got any idea on what's behind it...
reef fulcrum
work out what's creating that many teams
stuck nacelle
How can I do that? Is there any tool that can help me wirh it?
reef fulcrum
look at a heap dump, maybe the team name is useful
stuck nacelle
I disabled scoreboard plugins on all my servers for testing
And still the same issue
reef fulcrum
end of the day is that if the scoreboard stuff is creating a stupid amount of data on your proxy, it's creating a stupid amount of stuff on the clients too
try renaming/deleting scoreboard.dat
some scoreboard plugins are basically trash and will flood that file with junk
reef fulcrum
I mean, it's scoreboard data
something somewhere is creating scoreboard data
work out what
you can enable string interning during GCs in the JVM which will save some ram, but, iirc does make GC's more expensive
pretty sure that's disabled by default for good reason
stuck nacelle
Well, good GC might help
And less cpu i tendive
Any ideas on how can I achive that?
reef fulcrum
GC ain't gonna save your ass if you got stuff throwing a ton of crap in the heap
Work out where that crap is coming from
there is a reason WHY I keep telling you to look into what is creating that
if something is throwing out a ton of crap, you JVM has to deal with that ton of crap in the heap, that memory in teams is also gonna be sent to the client
saying you have memory issues and not working out where that memory is coming from is just a waste of time
and if you can't see that, am wasting my time
royal hawk
If you have a memdump to share I’ll take a crack at that
calm ether
@royal hawk #waterfall-help message
calm ether
@stuck nacelle i think you should share the file
nocturne fog
Only useful when there is a single super obvious thing
reef fulcrum
the heap dump
stuck nacelle
Oh, okay. I'll put it onto some file host
But now, as for scoreboard.dat:
- On most of servers it's less then 1kb
- On my hubs it's about 6kb
- On one game server it's ~600kb
Is this 600kb scoreboard where might be an issue? It doesn't seem that super big.
Also, on my hubs villages.dat is 600kb, which is quite a bit. Can't it cause issues (since I've got minor performance issues on hub)
reef fulcrum
size is irrelevant per-sei until it's gotta, 1) decompress that 2) send all the data in there to every player on that server
stuck nacelle
So what I can check in that?
reef fulcrum
heap dump, as I've said multiple times
stuck nacelle
@calm ether @royal hawk Here is heapdump avaliable to download https://www.easypaste.org/file/KIqwLZ6p/heap-2021-03-06.02.00.00.zip?lang=pl
reef fulcrum
or, as I said, try renaming it
stuck nacelle
Rename it, when server is off, or just rename it while it's running?
reef fulcrum
when it's off
reef fulcrum
or, try deleting the teams manually
stuck nacelle
How can I do that? I'm not really familiar with scoreboards. Sorry
stuck nacelle
I'll try to read and learn about it
There are as many teams as unique players
Thats bad?
reef fulcrum
I mean...
that's all gotta be sent to the client
am surprised that players don't struggle to connect with that many
stuck nacelle
This is correct, right?
No issue here
And to this 125463 teams, how can I find what's making them?
reef fulcrum
🤷♂️
you'd need to start digging into your plugins or something
pretty hard to say
stuck nacelle
This colide rule is not an issue, right?
reef fulcrum
i question why it has so many players but no
I mean, it's not gonna help, but, it's not the pressing issue here
glacial jacinth
does waterfall work with Forge servers?
stuck nacelle
Okay. It's probably how many players has joined today
But idk why it';s not clearing it, since there are about 150 players onloine now
reef fulcrum
1.12.2 and below, yes
glacial jacinth
does it work with paper 1.16?
reef fulcrum
supports 1.8-1.16.5
glacial jacinth
so if I want to have a forge it has to be 1.12 but I can have the newest paper servers right?
reef fulcrum
yes
glacial jacinth
When I set Waterfall up, can I just follow a bungee-cord guide for setting it up, because It is backwards compatible
reef fulcrum
yes
glacial jacinth
Thank you
stuck nacelle
Okay. So that's scoreboards:
- On 1.15 and 1.16 survival server there are as many teams as unique players
- On 1.12 servers with quicoboard colliderule team has the same ammount of players as has joined before server restart
- On 1.12 servers with no quickboard there as many players in coliderule, as there are online
So, the main issue is 1.15 and 1.16 server? Other than that, should I look for anything else?
glacial jacinth
I just had one last question, what is Travertine
reef fulcrum
- 1.7 proto support
glacial jacinth
whats that?
reef fulcrum
1.7 protocol support,
glacial jacinth
oh thanks
reef fulcrum
paste config.yml into a yaml linter or try again
glacial jacinth
ok thanks
stuck nacelle
Hey, so I'd like to thank you zzzCat, and everyone who helped me. I've managed to find issue with scoreboard data not clearing up. One plugin was in fact doing this. That allowed my bungee to work with much less ram. I'm still in testing, but it looks good.
nocturne fog
glad you seem to have gotten it sorted
scenic pebble
Nice work! Which plugin was it?
calm ether
Great to hear that :)
stuck nacelle
It was one of my own plugins that was used to hide player nicknames for more survival experience
scenic pebble
Ahh
placid grail
is waterfall version git:Waterfall-Bootstrap:1.16-R0.5-SNAPSHOT:c031df1:395 able to run forge and spigot servers?
nocturne fog
forge as long as it's not above 1.12, spigot/paper yes
I think velocity was working on a way to do 1.12+ forge but don't know if it works or not yet
placid grail
ok good then i have nothing to worry about. ik this next question is not related to anything here but anyone know of a bungeeguard alternative for forge?
i would assume 1.12.2 is a part of 1.12...
nocturne fog
yes it is
nocturne fog
as far as bungeeguard, not aware of anything for forge
although I also haven't looked far
placid grail
ok ty
slate bobcat
hi everyone, just trying to add waterfall to my current server and I'm getting the error that I'm not currently whitelisted! It's directing only to my main server atm in which I certainly am whitelisted. Is there something I'm missing? Thanks!
ok, nevermind. It has randomly fixed itself. Ignore me.
royal hawk
hi everyone, just trying to add waterfall to my current server and I'm getting t...
whitelists will only work correctly if you have ip forwarding set correctly- just fyi
celest solar
Did somone ping me
tidal musk
Can anybody help? I'm trying to set the host: to my waterfall server's port rather than the 0.0.0.0 default one, but it keeps reverting back
mighty cloud
what are the differences between waterfall and bungeecord? and why is better?¿ there isn't a list of features or something similar?
reef fulcrum
bug fixes, performance improvements, more api