#waterfall-help
1 messages ยท Page 35 of 1
@severe pine CPU does matter, but depends on plugins and player count
even an low-end Sandy Bridge Xeon E3 will handle 100 players on BungeeCord, you'll be fine
Is there a reason why players might be randomly kicked from the proxy?
Hi, any idea what's this caused by?
>.... [15:36:50 WARN]: Event PlayerHandshakeEvent(connection=[/127.0.0.1:25565] <-> InitialHandler, handshake=Handshake(protocolVersion=578, host=play.mcreunion.eu, port=25565, requestedProtocol=1)) took 443ms to process!
>.... [15:39:17 WARN]: Event PlayerHandshakeEvent(connection=[/127.0.0.1:25565] <-> InitialHandler, handshake=Handshake(protocolVersion=578, host=play.mcreunion.eu, port=25565, requestedProtocol=1)) took 346ms to process!
I'm thinking about if it's not by TCP Shield? Because there is local IP and TCP shield make it +50 ping more for players
something took too long processing an event
you'd need to use a profiler to see what
yourkit, visualvm, spark, etc
Yes, there is a version for bungee
Just make sure that it'll look at the network threads
Will try and post here results, thanks for advice.
Hi, any idea what's this caused by?
>.... [15:36:50 WARN]: Event PlayerHandshakeEvent(connection=[/127.0.0.1:25565] <-> InitialHandler, handshake=Handshake(protocolVersion=578, host=play.mcreunion.eu, port=25565, requestedProtocol=1)) took 443ms to process! >.... [15:39:17 WARN]: Event PlayerHandshakeEvent(connection=[/127.0.0.1:25565] <-> InitialHandler, handshake=Handshake(protocolVersion=578, host=play.mcreunion.eu, port=25565, requestedProtocol=1)) took 346ms to process!I'm thinking about if it's not by TCP Shield? Because there is local IP and TCP shield make it +50 ping more for players
@meager citrus you can try to searchPlayerHandshakeEventin pluginsJAR, to find plugins those use this event.
For example you can use WinRar
F3will open search menu
Wait and I can make zip of all plugins and then use F3 or I have to use it in every plugin jar?
you can just open folder in winrar and then search
Does waterfall has fixes against crash methods and malicius packets?
you can just open folder in winrar and then search
@lusty badger open one plugin jar with winrar and then double click to this folder with..
Wait and should I search for PlayerHandshakeEvent or PlayerJoinEvent?
PlayerHandshakeEvent
try to remove antibot, if not then write to TCPShield
I prefer Jh_AntiBot
ok this was maybe solved, TCP shield authors made better version. Now I'm having problem with this.
... [16:56:10 WARN]: Event PreLoginEvent(cancelled=false, cancelReasonComponents=null, connection=[/xxxx:52951|EPGMooky] <-> InitialHandler) took 260ms to process!
So I could use same method to find it? But now there's more plugins using this event, so.... any idea how could I identify more which one is doing this?
could run a spark sample when logging in
ok will try
their plugin is dodgy at best
does what it says but last time i checked their github then they uploaded decompiler output
Event PreLoginEvent(cancelled=false, cancelReasonComponents=null, connection=[/xxxx:52951|EPGMooky] <-> InitialHandler) took 260ms to process!
why dont add a method name to log
same q
what does that even mean
.g serverlistplus placeholders
(DiscordBot) https://github.com/Minecrell/ServerListPlus/wiki/Placeholders -- Placeholders ยท Minecrell/ServerListPlus Wiki ยท GitHub: "A flexible Minecraft plugin to customize the appearance of your server in the server list - Minecrell/ServerListPlus."
well ok slp does not do such placeholders out of box
you can easily add your own placeholders though
Hello!
Does anyone knows a good plugin for waterfall, that let's you use the same chat in all servers? I need the following features:
-Global Chat
-Msg
-Socialspy
-Staffchat
-Luckperms Prefix support.
And I'm looking for the same but with tab. Just a tab for all servers, only need intall for bungee. And it shoud support supervanish.
@delicate phoenix the part about supervanish will only work if thatโs a plugin designed to be used on or with a bungeecord companion plugin
Supervanish I mean
Oh okay.
BungeeTabListPlus is pretty much the go-to global tablist plugin
Right not I'm trying to insert lp prefixes before names
I don't think that one works still
hm, I guess it was updated for 1.13 
BTLP is by the same author and way more supported
and you can create a simple global tablist with it too
Is there a way to send someone to a server without saying they were summoned in chat?
Use a proper plugin for it
What plugin
Ref the convo that people 10/10 shoved in #paper-help
@junior talon here you go, exact same as /send except doesn't send the summoned person a message https://github.com/Proximyst/silentsend/releases
how much ram should I allocate for 50 players?
512MB
I have 700MB and it still running out of memory
512MB is more then plenty for waterfall on its own
bungeecord completely become unresponsive and 100% RAM and 100% CPU Usage
My plugins
people don't like crack around here
- piracy is fun
- Literally one of the most common plugins for #1, which often creates its own persistent stupid stream of issues
don't know
online-mode=true solves a lot of issues
it also reduces my player base by 99%
too bad
You're preaching to the wrong people, lol
Quite simply, I don't care if you run online/offline/whatever
I know but in my county people don't like to buy games
not 100% saying that it is them in this case
but, I have 0 interest in tryna solve issues for people running plugins which cause known issues
๐ฅบ
give it a gb and take a heap dump and see what is using the excess of memory
I am sure it's something to do with how they store skins in the memory.
I can't take a dump in a docker container
then get rid of the docker container
yes you can take a dump there...
.ask
If you have a question, please just ask it. Don't look for staff or topic experts. Don't ask to ask or ask if people are awake or available. Just ask the question to the channel straight out, and wait patiently for a reply.
What was that link about asking for experts?
.dontask
thank you
that?
Hello, I need some help. I have a bungeecord proxy server, that I have converted using waterfall.jar and placed it in the folder, after restarting the whole node, now in my hub server I'm not able to do /server
Is there a list of permissions that I am able to look at to set up?
the permissions are the same
As the permission node is bungeecord.command.server still this?
It says this
are you sure that you're editing bungee perms properly?
OH i realize one mistake, friend was adding it to other server than the proxy..
[11:26:18 WARN]: Event PreLoginEvent(cancelled=false, cancelReasonComponents=null, connection=[/<IP>:20572|<Username>] <-> InitialHandler) took 1,126ms to process!
Logging in has been pretty slow, we even removed all plugins at one point but it was still being slowish
Anyone else experience this
either a plugin or general overload
Any recommended jvm profiler for CLI?
No, you really need a proper GUI
There is spark, but, it's not alltoo much better than timings v1
But, 'least it works without external stuff
Spark doesn't run on Bungee?
there is a bungee version
.ask
If you have a question, please just ask it. Don't look for staff or topic experts. Don't ask to ask or ask if people are awake or available. Just ask the question to the channel straight out, and wait patiently for a reply.
I opened a port at my router config, but I don't know why my friends don't detect the server
I use waterfall, so, I need to open only the bungee port, or all server ports
And should I need to puc Tcp port and not more? 25565
only the bungee port
you should open & forward the port of bungeecord
did you forward it to the machine running the bungee?
Yes
wat
Well, did you configure the proxy to use 25565?
In query is 25575 but on the another servers is 25565
You don't put the servers themselves on 25565
I send the config
Unless you're tryna shift from the default port, you want to configure the listener to use 25565
We're not going to configure it for you
backends servers do not require to be port-forwarded for the public irc
Look at the host in the listener section
Ok
Project Waterfall build #348: SUCCESS in 3 min 1 sec: https://papermc.io/ci/job/Waterfall/348/
One final question, should I need to put in port forward 25565 or 25577?
depends... on which one are you using
Whatever one you have the proxy on
In query I have 25577
I usually just set my proxy to 25565, forward it, and then add backends by adding +1 to the proxy port for each one
you should read bungeecord documentation tbh
Ok
i think everything is explained there
So, you forward the host of the proxy
If it's not working, you'll have to start debugging
often times it's down to firewalls, routers being janky, or ISPs blocking people from hosting servers, etc
idefk why I auto added // Travertine, but, bleh
Project Travertine build #144: SUCCESS in 1 min 14 sec: https://papermc.io/ci/job/Travertine/144/
oh ok, then i guess you are using Iptables\UFW
It doesn't work
in canyousemee.org says that the port is open
But when I intent to enter to the server with mobile data's it don't detect
How much RAM Waterfall needs? I have 5 servers with Mysql and about 30-70 players
Half a gig should be fine.
and another thing. Is it normal that Proxy (waterfall) have A LOT OF threats?
I see A lot of proxies in htop
Seems it duplicates for each player (?)
Half a gig should be fine.
I gave it 2gb before
Then that should be more than enough.
๐ฅ
Do you recommend ServerJars to keep server updated?
never update without testing
Why the shit would we recommend some 3rd party platform ran by some sketchy character when we already provide jars officially?
Amm I didn't mean the 3rd party platform. I meant keeping to the latest version is good or not, and I got my answer. thanks
Is it a good idea to have every gamemode in one vps/vds?
depends on if you can do that or not
It's not too hard
it may be, performance wise
Is there a reason for waterfall to cause high ping-lag across the entire network? Usually when >300 players are online
well, that would suggest that it's overloaded or something
Can't really say without digging into a profiler
CPU usage on proxy is like 10%
CPU usage is a useless metric
Can you recommend a profiler?
yourkit, jprofiler, visualvm, etc
zzzCat what OS do you recommend for mc?
something debian based
Not that I should really even be looking into this bc I have literally no need for it... but it would be pretty cool to setup. Has anyone tried running multiple bungee instances and then load balancing players across those instances while somehow keeping the player count accurate. I looked into Redisbungee but it seems to be abandoned and cost money. Just wondering if there are any suggested resources for running multiple instances
im running redisbungee just fine
its free, you have to build it tho
and you have to fix this with it, too https://github.com/minecrafter/RedisBungee/pull/68
Hmm, ok
I wouldn't recommend using RedisBungee anyway, just saying
I say that as the former developer of RedisBungee
it's not exactly code I'm proud of
Hi, I have this issue where players will disconnect from a server, but not the proxy itself. The proxy will report the proper player count, but still show player names which have disconnected with /glist. I have seen this most commonly happen when a player has their internet obstructed.
6 Players Online, Ram Usage: 813M
Startup Script: java -Xms128M -Xmx768M -jar server.jar
Is this normal?
My Plugins
does aegis antibot recommended?
why needing an anti-bot
<dependency>
<groupId>io.github.waterfallmc</groupId>
<artifactId>waterfall-chat</artifactId>
<version>${project.version}</version>
<scope>compile</scope>
</dependency>
``` why this don't work?
What are you trying to do ?
I want to make a project of it
I also did with PaperSpigot, but it's different here
Opera browser be like not safe.
๐ฆ ๐ฆ 
anyone knows a plugin that allows you to tablist player names across all servers, currently I have a plugin but it only works for <1.12 clients but >1.13.x sometimes it shows as ~BTLP(random numbers/letters). On my 1.12.2 client I can tab all player names, but with 1.15.2 it shows as ~btlp if the player isn't on the same server. Not BTLP plugin fault, but PlayerTabComplete is. Couldn't find an alternative on spigotMC
This has been happening for the last 3 days. I am losing my patience with this. I am trying to connect to the proxy.
proxy ip is the ip i am trying to connect to
This is what happens when I connect to the lobby
What kind of plugins?
what plugins do you have installed
on which server
proxy and the lobby server then
Anyone know why this is happening?
you have ipwhitelist which configuration is very likely wrong Sneakyassasin01
use BungeeGuard rather
bUt whY 
kek
more often than not its just wrong file permissions
he is on mUsTdie
your os, your file system
@kind karma Its a new install of windows 8.1
probably some program locks file
how else are we gonna torture you
Idk i just had it on a usb -__-
Donโt use Windows, and donโt run your server off a USB stick would be great ways to start
INSTALL GENTOO
use BungeeGuard rather
@tidal musk BungeeGuard keeps crashing my servers
report it then
yeah and ipwhitelist can be bypassed like a joke
basically equal to having no protection
you should really just do the sane thing and set up a firewall
bungeeGuard doesn't even do anything deep enough to cause a crash...
Without info of said crash, literally nothing anybody can say, but, like ๐คทโโ๏ธ
.paste
Please paste large logs to a pastebin: https://paste.gg
Hi please, why it kicks all my players when connecting? Survival server is normally online but we canยจt connect wtf.
>.... [17:24:33 INFO]: [Rainbow_Matt|/x:64090] <-> ServerConnector [Survival] has disconnected
>.... [17:24:48 WARN]: [babus33|/x:58774] <-> ServerConnector [Survival] - read timed out
>.... [17:24:48 INFO]: [babus33|/x58774] <-> ServerConnector [Survival] has disconnected
>.... [17:24:49 INFO]: [Bxrd3r|/x:53231] <-> ServerConnector [Survival] has connected
>.... [17:24:50 INFO]: [Rainbow_Matt|/x:64090] <-> ServerConnector [Survival] has connected
>.... [17:24:52 WARN]: [Tokfa|/1x:49793] <-> ServerConnector [Survival] - read timed out
>.... [17:24:52 INFO]: [Tokfa|/1x49793] <-> ServerConnector [Survival] has disconnected
>.... [17:25:19 WARN]: [Bxrd3r|/x:53231] <-> ServerConnector [Survival] - read timed out
>.... [17:25:19 INFO]: [Bxrd3r|/8x53231] <-> ServerConnector [Survival] has disconnected
>.... [17:25:20 WARN]: [Rainbow_Matt|/x64090] <-> ServerConnector [Survival] - read timed out
>.... [17:25:20 INFO]: [Rainbow_Matt|/1x:64090] <-> ServerConnector [Survival] has disconnected
>.... [17:26:19 INFO]: [Tokfa|/x:49793] <-> ServerConnector [Survival] has connected
>.... [17:26:21 INFO]: [minec123_|x:63409] <-> ServerConnector [Survival] has connected
>.... [17:26:21 INFO]: [Rainbow_Matt|/x:64090] <-> ServerConnector [Survival] has connected
>.... [17:26:21 INFO]: [babus33|/x:58774] <-> ServerConnector [Survival] has connected
>.... [17:26:21 INFO]: [Renzotom|/x54651] <-> ServerConnector [Survival] has connected
>.... [17:26:21 INFO]: [Bxrd3r|/x53231] <-> ServerConnector [Survival] has connected
>.... [17:26:24 INFO]: [/x:49793|Tokfa] -> UpstreamBridge has disconnected
>.... [17:26:24 INFO]: [/x:49793|Tokfa] <-> DownstreamBridge <-> [Lobby] has disconnected```
@shadow iris the error tells you what to do
read timed out means that it wasn't sent a packet in so long
You should install Paper
zzzCat but why it's happening?
ominous chanting
Isn't it caused by protocolize?
I mean
if that's sitting on the network stack and stopping packets from being sent...
Well, that lines up..
And why is this error from protocolize in proxy console? I tried report to author, but it's weird that only I have this issue, maybe something related with Java 14 there?
Here is the general gist:
We can only tell you stuff based on what you show us
Protolize is throwing a ClassCastException; Why, I can't tell you, it's literally not in there
@shadow iris the error tells you what to do
@unreal stag I'm an actual idiot. Thank you
That was the problem. Thanks for your help
Got a new problem. Was able to log in and got a message saying that the server was closing, then when I logged into the proxy, it gave me this message.
the connection between the proxy and the server failed
Fixed that and now my other servers wont show up
the motd in the config doesn't seem to support the same data normal servers do? tried using motd: \u00A73\u00A7l... - paper shows the unicode characters but waterfall just shows it as text
put it in ""'s iirc
None of my other servers show up
that would suggest your config is busted or you have perms misconfigured
what perms would i need
depends on if you set the servers to restricted or not
they arent
Please paste large logs to a pastebin: https://paste.gg
A sensible, modern pastebin. Share text and source code snippets with no hassle.
You literally only have 1 server configured in there
no
i put them in there and they go elsewhere on the config
one is goose and one is survival
then you probably misformated it or something and it removed it
ill try putting them back
that worked
I'm debating making a second survival server and moving our "resource" overworld/nether/end there. I feel like that would cut down on the performance needs for our primary server. has anyone done this and did it feel worthwhile
been testing the mysql data bridge plugin and it seems to work OK, but not all of our plugins work cross-server so some things would be main-server only.
trying to decide if the perf gains would be worth minor things like that
mysql data bridge is meh
@reef fulcrum hey I also tried removing skin restorer plugin but it still ran out of memory and freezes the server
I am doing a restart at every 24 hour
did you figure out how to take a thread dump
Then as I said, you'll need to look into heap dumps
maybe look into that
Is there any plugin?
I can't help you without them
heap dump* not thread dump
;-;
(DiscordBot) https://www.spigotmc.org/resources/spark.57242/ -- spark | SpigotMC - High Performance Minecraft: "spark is a performance profiling plugin based on sk89q's WarmRoast ... Heap Dump - take a full (HPROF) snapshot of the servers memory."
spark should work on bungee/waterfall as well right?
or they have a version for that i mean, iirc
with the bungee version, yes
When an application experiences java.lang.OutOfMemoryError, itโs ideal to capture heap dump right at that point to diagnose the problem because you want to know what objects were sitting in memory and what percentage of memory they were occupying when java.lang.OutOfMemoryError occurred
Ok i will try spark then
@acoustic adder why is it meh? just wondering. I'm not sure the benefits to a synced second server outweight the negatives
The problem is generally "the plugins work, but make a lot of bad assumptions to work instead of tryna manage these odd state things properly"
https://spark.lucko.me/#Mfyxdk9D2f
Heap summary after installing the plugin but I think this won't be very useful cause memory usage is not very high@reef fulcrum
heap summary is basically useless anyways
Yes...
Ok
Also, it needs to "leak" first
how fast we reckon 1.16 support will be added to mainline waterfall?
Yoo jake
PR is basically already sitting there, I could maybe pull it later today if it's updated
And for paper?
That would be pissing in the wind
LOL
๐คทโโ๏ธ
pull it later if what is updated
the PR
to the latest protocol?
Yea
ah
tbh, just set the flag to heap dump on OOM
... heap dump being 167mb is nothing my dude
๐ฅบ
i've tried examining one as large as 12gb
only 12G?
.hprof extention
cute
So how do I send it drive
compress it...?
Ok
with decent compressor
Ok
winrar is not decent
use 7zip
tbqh, easier to analyze it yourself
Idk how to do that
Well, you'd have to learn, but, can't really say looking into dumps is top priority
I can hardly read me screen right now
.g how to analyze heap dump java
I will try to read it
think some bastards are mowing the grass
(DiscordBot) https://dzone.com/articles/java-heap-dump-analyzer-1 -- Java Heap Dump Analyzer - DZone Performance: "Aug 3, 2018 ... Learn about heap, memory, and garbage collection in Java applications and how to take the heap dump of an application and analyze it..."
If I keep running the server for 10hours does heap dump size will also increase?
Or it depends upon the ram or something?
are you able to make a commit to a PR @reef fulcrum
literally just need to change one line lmao
F
i think I can, I just can hardly read right now
are you drunk cat lmao
๐
No, pollen I think
ah
https://github.com/lucko/spark/wiki/Commands#spark-heapdump
Should I use any arguments except compression once?
@proven crypt apologies, I pushed it to 735 now. Have been running tests with ViaVersion for consistencies sake till now. Seems all good
@reef fulcrum we ready to roll tomorrow? ;)
Please don't mention project developers and please don't ask for ETAs, there are none.
@serene tide dude- I'm the one who made https://github.com/PaperMC/Waterfall/pull/494 and we were just talking about that
question is
is it worth porting that patch to travertine
because this patch breaks travertine
I might end up doing that early tomorrow. Who knows
lmao
Tomorrow is going to be fun
thanks dude

@proven crypt i fucked that PR regardless apparently
i forgot about the whole patch system shit w/ file signatures
so it's not a proper PR or something idk
you what
the waterfall 1.16 pr lmao
oh the code change works
https://github.com/Xernium/Waterfall/pull/3#issuecomment-647739550 the patch pr isn't
lmao
clrxbl is you?
yes
nice
lmao
can dl here: https://jenkins.krusic22.com/job/Waterfall-Xernium/
oh yeah he has his own CI
epic
eh doesn't matter it'll be in waterfall waterfall tomorrow or something
maybe md_5 will surprise everyone and dump spigot like 2 hours after
if only
Well itโs still going to take a while to rebase the other waterfall patches

Also not my own CI, sponsored by yours truly @covert portal
๐
-gang
Gang gang.
ok
I mean without we have no information to go on so yes
?paste
Please paste all your code/errors/startup logs at http://hasteb.in/
is that address correct?
(and the port open/free I guess, not sure if it sends a different message in that case)
the query port is only important if you enable you need to enable the query for some reason
default should be 25577 iirc
did you enable it
that IP address is invalid
from the log, it's telling you that the IP address is invalid
How are you hosting the server?
Best guess would be speak to your hosting provider/check their FAQ
if they're on ptero, just set it to 0.0.0.0, but, might need to figure out what port they expect
hi guys
I'm looking to turn my server into a bungee one, and I take it that waterfall is an optimized bungeecord?
correct
Alright
now can someone pls explain how bungee servers work/what I would need? I'm new to this kinda stuff
.g spigot bungeecord wiki
(phoenix616) https://www.spigotmc.org/wiki/bungeecord/ -- BungeeCord | SpigotMC - High Performance Minecraft: "Dec 13, 2014 ... The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy."
The home of Spigot a high performance, no lag customized CraftBukkit Minecraft server API, and BungeeCord, the cloud server proxy.
doesn't look like an issue ๐คทโโ๏ธ
hmm
should I send you the file I have no idea about this ;-;
there are the memory leak suspects
leak suspects are generally useless unless you're familiar with it
General gist is that you want it to OOM and grab a dump, that way you can see what's consuming the heap when it blows up
ok
currently it's not consuming a lot
java -Xms768M -Xmx768M -jar waterfall.jar
This server is running for like 4 hours only and the usage is 701 / 1024 MB with only 4-5 players online
Hey cat just replay me on this:
can I use this on my startup script?
-XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/home/container/heapdump.bin
yes
ok Thanks
currently I configured my containers to use more memory in need, I will revert that
Can Waterfall run on 1.16?
Not yet 1.16 isnt out yet
there's forks with 1.16 support
hi guys, quick question. Running the waterfall jar for the first time but it's not generating any config files etc? Any ideas? Thanks
@dusty rampart Does the folder permissions allow it to make the configs?
it does, yes
Did you triple check? lol
lmao
it's been too long
problem solved!
thanks
is there official support for fabric servers on waterfall?
there's FabricProxy
and fabric's registry sync protocol should work fine over proxy
if anyone is impatient and wants to update their waterfall to 1.16
xernium's waterfall 1.16 patch build can be retrieved from https://jenkins.krusic22.com/job/Waterfall-Xernium/lastSuccessfulBuild/artifact/Waterfall-Proxy/bootstrap/target/Waterfall.jar
do note that this build is unofficial, use with caution
works fine in combination w/ viaversion
figured i'd send it since there's a lot more people going to ask for it
Hey cat please reply, I just took a heapdump of my proxy when the ram usage was 900MB/1024MB
and my startup script is
java -Xms768M -Xmx768M -jar -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/home/container/out-of-memory/heapdump.bin waterfall.jar
heapdump size is 16MB after compression (original size 200+ MB), I already looked into it but can't find anything, could you please help me ;-;
compression done by the spark plugin itself
Project Waterfall build #349: SUCCESS in 3 min 8 sec: https://papermc.io/ci/job/Waterfall/349/
๐

hELLO!
Do you guys recomment using https://mcflags.emc.gs/ for waterall too?
waterfall works fine without gc flags tbh
Check pins.
no, pins
heya please don't mistake this as me asking "When it 1.16 available" I'm not asking that. I want to ask if travertine fully supports 1.16 as of right now?
thank you
no, it does not
Alright thank you.
waterfall does though ๐
Oh is waterfall updated for 1.16 already? >.>
I use travertine since it's easier on my modpacks
But thank you
It's okay i stuck viaversion on the bungee for the time being and that works
if you have it I can take it, won't touch it till I get to my desk though
I'll push once I get home
also
how can I help to earn that nice purple contributor role? 
once md_5 updates bungee-
eh you'll get it sooner or later
cat you want me to do this
Once again
For the last time
You want the proxy to OOM and generate the dump, then look into it
It's a complete waste of time taking dumps and tryna look into them when there are 0 issues
Yup got it
should I send that .bin file here?
ok leave it I will not ask questions until I have that file
is there any way to make waterfall ignore unknown/invalid packets and just forward them?
i know it is a security risk that i'm willing to take
just install viaversion
we already do where viable
afaik
(Unless there is some oddball specific case am missing)
Oh, during login sending invalid data is not allowed
The protocol is pretty strict there, essentially
is there a way to reduce the strictness?
yes. Bungeecord as well as any other sane proxy requires that data
99% of the time, that's down to janky plugins, etc; ignoring that stuff doesn't really fix anything
i need to forward fabric's registry protocol for my modded server
wich happens during login
Are they just using plugin messaging channels?
nope, they use custom packets
fucking...

Mojang LITERALLY added shit for this
imagine using the system mojang dedicated for this
Basically, no, it's not possible to just ignore that stuff
There is a whole lot more state involved than just a singular packet, e.g. jumping between servers will cause issues, etc
fabric registry sync uses the plugin data message, don't know what you're talking about martinz
like at the moment i don't think fabric itself even uses login query -- everything's done during PLAY
@reef fulcrum as promised https://github.com/PaperMC/Travertine/pull/128 seems to work alright from my testing but it wouldnt be beyond me to do a rebase screwup. Please have a look yourself, thanks :D
Multiple people have reported that npcs/entities from one server are being seen from another server
I think the 1.16 build broke it
the 1.16 build won't break anything with existing versions
it literally only changes stuff for clients with the new protocol version
Weird
unless you have something like VV on the proxy
well, suggest not putting it on the proxy if it's there
it's all a pipeline, unless there is some oddball race condition, ain't really a reason how entities would leak across servers
One sec, I am checking if they have it on the backend or proxy
They have it on the backend
@royal hawk
Iโm investigating already
And it only happens on 1.16
one thing that would hit me would be if there was any client changes here, especially around how dim changes affect entities
Itโs oughta be keep player meta
Is it correct that at the moment Travertine does not support 1.16?
yes
Gotcha gotcha, but waterfall does right? Is it usually a quick process for travertine to follow or should I just move to waterfall for a bit?
basically, waterfall supporting 1.16 makes doing travertine a pita, not to mention there is apparently an issue with the 1.16 stuff
Iโm still looking
What's the issue? Should I not support 1.16 yet?
It doesnโt clear players during server switching
not only players, all entities
Also I am not really sure what you meant by "a pita" (nvm)
yes i see armor stands
Gotcha
i have the issue only when i keep entity rewrite
dang so if you swap servers anything that was around you on your previous server will appear to still exist when you swap servers?
Yep
do you have entity metadata rewriting disabled?
Lol

i think there is a missing respawn packet
Last question is there any reason I can't just swap out the travertine jar with the waterfall jar? Like would anything break?
1.16 requires only one
You'd just lose 1.7 support
Big oof, I canโt use that emoji here since I no have nitro
great ty
[13:13:00 ERROR]: [Plutoren|/XXXXXXXX] <-> ServerConnector [snap] - encountered exception: net.md_5.bungee.util.QuietException: Unexpected packet received during server login process!
1b81047b227472616e736c617465223a```
I get this while trying to join on 1.16. This is a message from the proxy
Kicked whilst connecting to snap: Outdated server! I'm still on 1.12.2
The server is for sure on 1.15.2
That server clearly thinks otherwise
Do I need to adjust a config somewhere to tell my server otherwise?
All other versions besides 1.16 work
lol
don't you know that almost 50% of trans-version servers commit suicide because of this kind of bigotry

See
This is why I just disable metadata rewriting
entity destroy packet? Is the client not killing that data anyore?
fine. I'll disable it and throw you the patch again
I dont have time to debug this either

Should I be using ViaVersion and ProtocolSupport to get 1.8-1.16 support, or is there a better solution?
I'm just trying to figure out what might be causing that 1.12 error ๐ฆ
oh no you posted the transparent one
๐
dammit glare
People here have told me it's possible to get your backward compatibility from ProtocolSupport and your forward compatibility with ViaVersion at the same time, and that this is actually suggested
and hypnotic
@ zzzCat to disable meta rewriting for the version I just need to replace the entitymap with the dummy instance or am I missing something?
Yes
userscript that injects that image as the background of every webpage when
Idk if this could possibly help but plotsquared claims to have solved the issue, even tho they are just a plugin
https://gyazo.com/08254e6b4d0287fc4f952ddc0dc8da53
should be something else
@signal trout No, replacing ViaBackwards/ViaRewind with ProtocolSupport wonโt work
Well atm I am using ViaVersion and ProtocolSupport and I seem to have 1.8-1.15.2 support on 1.15.2, but I have that issue where my proxy says my server is on 1.12.2 when connecting with 1.16. Do you think removing protoclsupport would fix this issue?
@signal trout You need waterfall
I have it, just got latest
And as I said, replacing ViaBackwards/ViaRewind with ProtocolSupport wonโt fix it
This server is running Waterfall version git:Waterfall-Bootstrap:1.15-SNAPSHOT:265e719:349 by md_5
And lets go to the ViaVersion discord
I'm not sure this is a ViaVersion issue is it? Since my error is on the proxy
if youre still interested

sorry for the trouble
Project Waterfall build #350: SUCCESS in 3 min 3 sec: https://papermc.io/ci/job/Waterfall/350/
@signal trout Update ViaVersion with the one from #links in the ViaVersion discord
Ok Buzz
Well if you do close the PR
@calm ether should be fixed in 350
@covert portal that goes for you too 
Well, yes but no.
I updated everything but still have the same issue ๐ฆ I guess I will try on a test server to see if ViaVersion is really the issue
is VV on the proxy?
@signal trout do /viaversion dump and post the link here
No zz, should it be?
no
This is the error, which is only seen in the proxy console https://gyazo.com/75db8c04824f9c5e4c8271a2f572da6f
with 1.16*
That's only thrown when the server detects a proto mismatch
Did you change the outdated message in spigot.yml or whatever it was? o.O
Unlikely
let me try to bare bones this on a test server, it could be related to protocolsupport+viaversion. Also I might just need to update PS
defo update PS
@royal hawk what made me say there is a respawn problem is the fact that the world is not unloading
my client lag as possible when switching and the chunks are progressively overlapping
that does happen on vanilla clients <> servers too
Is the dev build for PS for 1.15.2, or 1.16? The website doesn't really tell you anything about the versions
I'll just ask PS discord nvm
i dun think it's PS
But, that fact that you're getting the proto mismatch is a "somethings interferring somewhere"
Yea, it wasn't PS, oof. Time to setup a test server
i don't understand why only one packet would be needed. yes the client is not blocked anymore in the dirt screen, but the purpose is to force the client to unload world by telling him to switch to another dimension
my problem is that this doesn't happen anymore and i see ghost chunks depending on the server view distance
Does waterfall work with 1.16, I saw that there are 2 commits but I'm not sure what they mean?
@calm ether yes but you cannot reliably double dimension switch anymore. Switching dimension requires rekeying the registry
And that means relogin
For now Waterfall supports 1.16
Ok, any eta for travertine? I can't image it would be very hard to pull the changes?
Well I did the patch for it so itโs on zzzCat to either pull it or donโt
ok
I have confirmed on a blank server that running PS and VV together yields the 1.12.2 mismatch error when connecting with 1.16 through waterfall
@signal trout if you donโt need to rely on <1.7 then you could use ViaBackwards and rewind instead of PS?
No?
I have a bad history with ViaBackwards stuff
I forget exactly what the issues were but I suppose I will try it again
When i try to connect with 1.16, console says "No client connected for pending server!"
for some reason, I'm getting these errors with my resource pack on 1.16
and even when I host it on my website in a nice clean url
proxy logs?
k one sec
I think there is a proto messup
can't reproduce personally
same
I use bungee resource packs
why?
Because protocol IDs changed
use forceresourcepacks
ah
it works fine with that plugin
Well, packet IDs
in the original irc bot, people in irc could mention discord users just by saying their name
just like how it works in IRC
turns out, thats really annoying for people not used to it and everyone complained
lol
tried myself extensively, cannot reproduce
with bungee resource packs? @royal hawk
resource-packs through bungee
hmm okay
are you using the vanilla servers resouce-pack= in server properties or a plugin?
teleportation and death seems to kick you from the server
Did you remove that resource pack thing?
The packet literally sits around the respawn one
Pluto
Check your logs/remove PS, just saw from somebody that PS apparently prevents via hooking a class
I am currently running via only which seems to be alright atm
You are correct, here's some logs for reference https://hastebin.com/newapodomi.md
Plugins like that usually need update with new releases
Somebody got a server they can throw trav on and test on a few versions for me?
GOD FUCKING... 8mb limit...
8mb limit what?
I can test
is that a /dev/urandom file
nitro upload limit is 50mb iirc
100mb for nitro users
that has been bumped up recently then
i remember the nitro upload limit was 50mb at one point
it says 100 when I try to drag it over here
That travertine version you shared looks like it works fine ๐
Project Waterfall build #351: SUCCESS in 2 min 59 sec: https://papermc.io/ci/job/Waterfall/351/
Yea, the site needs updating, bleh
where can i download
the latest version of waterfall?
ah there ok
do you think its safer to use a previous version
or this version should be ok?
with the latest version of waterfall when i try to connect to a vanilla 1.16 server it disconnects me with
[16:08:02 WARN]: [/xx.xxx.xx.x:11797|blucobalt] <-> DownstreamBridge <-> [survival] - bad packet ID, are mods in use!? Did not read all bytes from packet class net.md_5.bungee.protocol.packet.EntityEffect 90 Protocol GAME Direction TO_CLIENT
looks like it only goes up to 1.12
this is the latest i see
Should be fixed in the upcoming build
Project Waterfall build #352: SUCCESS in 2 min 58 sec: https://papermc.io/ci/job/Waterfall/352/
Yea, I know, the site apparently doesn't auto-update
https://www.blucobalt.dev/i/PIxnK.png found a fork that might work
@tidal musk can u send that fork ?
let me comb through my history rq
even with vanillacord i still get DownstreamBridge <-> [survival] - bad packet ID, are mods in use!? Did not read all bytes from packet class net.md_5.bungee.protocol.packet.EntityEffect 90 Protocol GAME Direction TO_CLIENT
update waterfall
someone can send here waterfall-352.jar?
just refresh the web page
Can't download it
https://i.imgur.com/Q50kOZT.png
works fine here
probably that fun browser cache bug
hold shift and refresh again
Perfect, thanks!
351 was having issues with packet ID 90 (seen above), on 352 now and having issues with packet 88 bad packet ID, are mods in use!? Did not read all bytes from packet class net.md_5.bungee.protocol.packet.EntityRemoveEffect 88 Protocol GAME Direction TO_CLIENT
I erm... keyboard...
This is with ViaVersion 3.0.0 on top of Paper 1.15.2 #370 if that helps anything ๐
(or maybe it's just a typo as your message implies)
Project Waterfall build #353: SUCCESS in 2 min 55 sec: https://papermc.io/ci/job/Waterfall/353/
that was fast
trying ^
wat
it appears to work
Yea, I just mistyped a protocol number
download has been sitting at 100% for a while now...
yeah it looks like its working
whitelist works too
i think
yeah whitelist works fine
thank you
is there a waterfall /version command
/bungee
travertine pwease? ๐ฅบ
i was working on it but ran out of IQ points
understandable, no worries ๐
Project Waterfall build #354: SUCCESS in 3 min 2 sec: https://papermc.io/ci/job/Waterfall/354/
Hey cat see this, My server is out of memory but it's didn't created the heapdump and also I can't interact with it.
Project Travertine build #145: SUCCESS in 1 min 26 sec: https://papermc.io/ci/job/Travertine/145/
you should probably remove that screenshot >_>
@tidal musk delete that screenie, you're opening yourself up to DDoS
You might want to talk to the hosts about it
Looks like WitherHosting to me, open a support ticket on their discord and they'll help you out
It's self hosted
I am just trying somethings
Enabled OOM killer
Let's see if that works
And that screenshot don't contains my ip
kill the server, there is literally a button for it
oom killer only helps, if the setting is applied upon the container start, therefor needs a restart
also the server must be over its ram limit
btw look at pinned messages
https://pastebin.com/raw/mpm7425i this is completely force resource packs plugins fault right? not anything with travertine?
hey all, trying to get waterfall to connect to a vanilla 1.16 server and it's throwing a "Unexpected packet received during server login process!"
I'm on the latest build (#354)
Vanilla by default is not Bungee compatible. I guess waterfall isn't changing that state. Right cat?
full error log from trying to do /server to the 1.16 one -- no other messages https://gist.github.com/cyberkitsune/8e199bc41b05b22c5215b9081dc5d829
oh reading the scrollback I have to patch my vanilla jar
Waterfall fixes Bungeecord, 2020 (colourised)?
is that why the bungee one doesn't even look like you are joining a new server opposed to the velocity one?
oh god, custom world name blows it up?
This was fixed by last waterfall
LOOONG
I hope I got my point across
He responded
@ aikar I don't think investing the time in trying to make this even faster is worth it
but I want your opinion on a fast-switch patch anyway;;
see vid above
having the option to fast switch would be sweet
what do you mean by fast switch
first video
this one ? #waterfall-help message
yeah
if there was entities on the vid, they would have stayed between server switch
ah ok
@calm ether on velocity in fact not.
the fast switch actually looks pretty nice, i hope that it also does work as good as the normal one
anyone getting this on latest waterfall? happens when swapping between servers in game
Do any of the proxies work for 1.16.1? Waterfall, Bungeecord, and Taverntine dont work yet.
Guess which PR ain't getting accepted ๐
jup, know that message from my pr when 1.15.1/2 came out. ๐
Luckily ViaVersion works on Proxies as well, and they do very fast updates
Wrong download in Travertine 1.16 #145
Accessing this link the download goes normal
Ja this for example of the error:
or
@reef fulcrum
Wrong download link
dont ping
