❗-1-18-experimental

[neon shard]

wth are you doing

and yet still this

(rendering live map)

[peak pulsar]

[strange compass]

Unlimited CPU usage? Nice

[merry haven]

Wow 16 cores/threads

[neon shard]

16 threads yeah

[vivid iron]

message.txt by @next owl: https://paste.gg/afd0249563b54db58ca131b52ed3248f

[peak pulsar]

Offline mode isn’t supported here

[marble badger]

is world generation a bit broken in the latest snapshot?

specifically this being the generation

and yes in that order

oof rip

[marble badger]

3 client restarts and it finally appears

[calm crag]

how can it be not responding for -3 seconds?

oh alr

[sweet shadow]

They made it insist on never falling behind on ticks and crashing if it just can't keep up?

[gray yarrow]

Merry Christmas

[late topaz]

Joyeux Noël !! 🎄 🎁 🥳

[uneven pagoda]

So, because of upstream bugs with comments in yaml files, it will not be updated anymore? :D

[dapper lotus]

no?

[uneven pagoda]

There's no PaperDocs for fix-target-selector-tag-completion

[topaz hemlock]

Anyone facing any chunk unloading issue

Like chunk getting unloaded in front of eyes

[dapper lotus]

never heard of that one

I mean, theres the clients years of bugs of rendering shit which can create some fun oddities

but, never heard of seeing stuff unload inside of the vd range

[patent epoch]

We should have this link pinned

https://github.com/PaperMC/Paper/issues/6136#issuecomment-982297893

[dapper lotus]

there is a command for it

it's literally linked in the obvious channel where info is

Like, pull ones finger out of ones ass...

[patent epoch]

Hm let's see. Not in Paper-help, not in the 1.18 channel, not in announcements, not in general, not in this channel... can't find a channel named 'obvious'
Just saying, it would be nice to have pinned.

[dapper lotus]

It's in the 1.18 channel

[patent epoch]

Cool, I only checked pins in the channels

[eternal terrace]

will the paper be able to reduce the consumption of RAM in future? (1.18.1)

[dapper lotus]

Java uses as much ram as you tell it

most of the ram consumption is churned memory by the game itself which, we like to reduce that where possible but 🤷‍♂️

[eternal terrace]

okey

[urban epoch]

want to not use ram? simple fix - just don't run code on your machine

[eternal briar]

-Xmx16T

[vivid bluff]

127T.

[patent epoch]

OOM

[spark panther]

obligatory "mounting google drive over rsync as swap for infinite ram"

[cunning stump]

Amazon S3 as RAM

[sour token]

you do not mean that

[dark nymph]

How about get a bunch of em and; -Xmx1P

[eternal briar]

yes yes -Xmx1Y

[sour token]

i think at that point you should start using aikar's 12+ gb flags thank me later

[cunning stump]

[timid wolf]

is tps being over 20 a bad thing

[main mesa]

[drowsy zephyr]

It appends when the server was lagging in the first place. When the lag is finished, the server temporarily boosts the TPS to catch up.
In this case, this is not a bad thing.

[rough crow]

Hi, what should I do.. I have a server that has 9 GB RAM and when the server uses 9 GB RAM out of 9 GB RAM it freezes after some time and the CPU usage is at 0.05

[gusty trench]

How are you hosting the server?

[rough crow]

On the Pterodactyl panel , which runs on my dedicated server

[novel briar]

Hello, can anybody tell me what the current state in paper 1.18 is against seed cracking?
I've heard paper has some countermeasures against that and I'd like to learn more.

[gusty trench]

Try lowering Xmx

[feral spire]

feature seeds and manually adjust the structure seeds in spigot.yml

the only 2 structures that aint configable is the stronghold and mineshalf

[rough crow]

So I set Xmx to 8192 and leave 9216 MB of RAM allocated on the server?

[novel briar]

generate-random-seeds-for-all can I set that to true after already using a world?

[feral spire]

that would render knowing your seed useless but the only real way to prevent seedcracker is custom world gen.

it will only work on newly generated chunks @novel briar

[novel briar]

alright thank you

[gusty trench]

That should work

[patent epoch]

That depends on whether or not you see tps catchup as a bad thing. I do, so have it disabled. Another option that would be nice for Paper to have tbh

[sweet shadow]

At that point you can just cut out the garbage collector, the server won't run out of memory until you need to restart anyway

[jolly mortar]

Don't you have a problem with teleporting mobs into the portal?

[novel briar]

Alright question: I changed the settings but ship wrecks for example can still be tracked using https://www.chunkbase.com/apps/shipwreck-finder for example.
I noticed that on paper.yml world-settings.world.feature-seeds.* has a number of sub seeds saved. I assume I need to remove some to regenerate them from the random seed. is that correct?

[feral spire]

for shipwreck, it's in spigot.yml and default will follow your world seed yeah so you need to "manually" set it and delete the world again

(delete as in if you are making a new world)

one in feaeture seeds cover everything that's not in spigot.yml

[novel briar]

if I want to keep the already loaded chunks on my world I should be fine just deleting the 10 or so lines from the spigot config correct?

then just the newly generated chunks would be affected?

[feral spire]

I mean if you change it after, it will only change in the new chunk yes

[novel briar]

alright and (just making sure) settings generate-random-seeds-for-all in the paper.yml and deleting the seeds from the spigot.yml will generate new random seeds for the structure that are not derived from the world seed?

[violet wing]

not sure if it does it for the spigot ones, i just wrote mine in manually cause there aren't too many

and yeah old chunks dont gen new features

[feral spire]

I think spigot one follows whatever world seed is and you have to do it manually

but I could be wrong though

[novel briar]

Alright then I can do them by hand, whats the range on those seeds? so i can randomize them

[serene temple]

is updating with 75 or more players not a good idea?

[violet wing]

i had 50-60 on and it was working fine for me but as always the current builds are ✨ experimental ✨

[topaz nymph]

Just make sure to backup

[violet wing]

the main thing at scale was that missing the chunk loading patch sucks

there is quite a bit of stalling if a lot of players are moving about

[serene temple]

hmm ok, how long does the blending take usually?

i've got a 7.5k border

[violet wing]

blending just happens when people load the chunks

[serene temple]

oh lol, i would probably do it with chunky then

[violet wing]

yeah either way

the only thing to watch out for is dont use the forceupgrade flag cause iirc it doesn't do the blending

[serene temple]

oh ye, and ty for the help though

[feral spire]

idk default has 9 digits so I just randomly change 1 or 2 digits

[novel briar]

Did that, works perfectly Thank you!!

[feral spire]

just keep it in mind, if they want to burtforce your world seed they still can but almost all benefit of actually cracking the seed is really small now? aside from knowing terrain (?) @novel briar

[novel briar]

Yea I figure they need to crack the individual seeds for villages / shipwrecks now if they want that

[feral spire]

good luck to them

[slim pivot]

If you do force upgrade, will chunks blend afterward when players walk through them?

[violet wing]

i dont believe so, just dont force upgrade

[slim pivot]

O dear

[violet wing]

just normally load them is the safest bet

[dapper lotus]

nobody afaik has conclusively checked but it's what we've generally heard

if somebody has evidence otherwise 🤷‍♂️

[violet wing]

true im just going based off the rumors lol

[dapper lotus]

but, unless you're using 1.12, the upgrade process is cheap af, so literally irrelevant

[violet wing]

but also the currently widely held mentality is not to use forceupgrade so I guess that too

dataconverter is pretty fast anyway

[junior gorge]

c

[cunning thicket]

what bugs is there in 1.18 paper?

[dapper lotus]

see the issue tracker

[half walrus]

https://github.com/PaperMC/Paper/issues/6136#issuecomment-982297893

[novel schooner]

Hi, does anyone know a plugin that limit the RAM usage ? I know we can do that with the startup command, but our host doesn't allow us to modify it and don't understand why he shouldn't allow 100% of the server RAM to java

[timber ridge]

No, a plugin can't do that.

If your server only has 8GB of RAM and you give all 8GB of it to Java/Minecraft then it'll crash.

You need a bit to actually run the system.

[novel schooner]

I know that, I've already explained it to the host, but he didn't care about it because for him "server crash because 1.18 use more ram"

[timber ridge]

Oh you're trying to limit it. Yikes.

Unfortunately the only thing that can change the RAM usage are the JVM startup flags.

You may want to consider getting a new host if you can't get what you want out of that one.

[novel schooner]

I think that what we'll do when the server will expire and if he's still doesn't want to fix this simple issue.

Thanks anyway for your help.

[timber ridge]

Sorry I couldn't help more.

[spark merlin]

did they setup the servers launch expression correctly?

[timber ridge]

No, they said the host doesn't want to adjust it.

[feral spire]

Sounds like you need to switch host

[novel schooner]

They're using Pterodactyl, we have 10G of server ram and 100% of it is allocated to Java. They didn't want to give us 1G more of unallocated neither reduce the java tag to 9G. Also they didn't wand to add the Aikar flags.

[timber ridge]

Yeah you definitely need a new host.

[strange compass]

What host u are using?

[novel schooner]

I won't name and shame them.

[dapper lotus]

So, basically

You're using a host who has half a clue of what they're doing and won't even work with you to make the thing actually work

Like, imho, if you stick with them, you're dum

[novel schooner]

Well, I'm not the server owner, I'm just taking care of the development & updates. And sadly, the actual server owner want to stick with this host.

[clear steeple]

shame them regardless

We may also know people who work there if they're a decently sized host

[feral spire]

If I am being given a captain of a sinking ship, I would glazing abandon it. Choice is yours.

[half walrus]

https://eternity.community/index.php/paper-optimization/

[ocean copper]

Wouldnt be a name and shame, would be helping the community to not choose bad hosts imho.

[winged vine]

Still got a weird issue with mob spawning, i've checked out the guide. It is as if stuff isn't despawning as it should, if i play the game for abit then go to my AFK platform almost nothing spawns. If the server reboots and i load back in while on my AFK platform the mob farm works extremely well

[gilded ice]

anyone knows this bug?

https://ptero.co/yfifuqumil.md

[inner crypt]

That plugin has a bug

Talk to the author

Or you messed up the config

Basically when creating a gradient it divides by 0 which is highly illegal

[gilded ice]

mh ok, thank you

[novel schooner]

I've sent a mail to the CEO with all the necessary explanation, maybe the ship can be replenished

We'll see if he care about it. Otherwise...

1.17.1 required about 2 reboot/day to not crash, but since we've switched to 1.18.1 we need about 5 per day and that's annoying

[amber mica]

My 1.18.1 server runs fine for several days, so that definetly seems like an issue on your end

[regal meteor]

i remember seeing some very very big iron veins is that normal?

[gusty trench]

For 1.18, yes
https://minecraft.fandom.com/wiki/Ore_vein

[haughty kiln]

Hey, people on my server crashes when they break spawners. Any known bug or anything I can do about it?

[amber mica]

Is there an error message in the log? Or a crash-report?

[haughty kiln]

no only
[15:15:07 INFO]: player lost connection: Disconnected
[15:15:07 INFO]: player left the game

[white robin]

probably a plugin issue

[amber mica]

Do you have any plugins that make spawners breakable or something similar? Try removing those plugins first

[haughty kiln]

nope

[violet mist]

uh

[haughty kiln]

here is my plugin list:

[violet mist]

https://mcpaste.io/9ee3ea627080841c

[haughty kiln]

all plugins updated to latest

[amber mica]

Not paper

[violet mist]

Wait this is not

Fuck

I should switch to paper

[amber mica]

Still, try removing plugins, you are the first one with that issue so it's probably a plugin issue.

[violet mist]

Alright

[amber mica]

Also make sure it's not your players using some weird client mods that make them crash.
And sorry, that was meant for Jonakoli

[violet mist]

Uh

I try doing /pl with op

It says i cant use it

[solemn wyvern]

when will 1.18 come out thay said it took like 1 - 2 weeks

[half walrus]

__There Is No ETA__

Updates to Paper do not have any sort of estimate for when they release, ever. Any and all updates will arrive when they are ready, and the only thing to do is wait for them patiently along with everyone else.

ETA means 'Estimated Time of Arrival'!

[feral spire]

Who said 2seek? Go slap the person

[solemn wyvern]

a lot of peoples said 1 - 2 weeks after realese

[feral spire]

It’s out already you just need to do regular backup

[half walrus]

https://papermc.io/downloads

[solemn wyvern]

yea

[feral spire]

I guess we are in agreement

[winged vine]

Any ideas on this please? 🙂

[feral spire]

Any player online during?

Despawn range is lower than your simulation distance?

Read this part and see if this is your issue https://eternity.community/index.php/paper-optimization/#per-player-mob-spawns @winged vine a very common misunderstanding

[winged vine]

Servers empty except for me, despawn range soft is 32 hard is 128, simulation range is 10

[feral spire]

View distance?

[winged vine]

12

[feral spire]

Take it down to 10 to see if it spawns more. As a test

Your lowest point of farm (pic looks like it’s on the sky(?)

Is at least 128 blocks right?

I mean from your afk spot

[winged vine]

I'll give that a go, and it is yeah

[feral spire]

You can remove it with fabric client + sodium extra

[vivid iron]

message.txt by @gilded ice: https://paste.gg/058cec21c86245e18346b01c9349baa5

[topaz nymph]

That's a plugin issue.

[rough crow]

After some time the server freezes, do you know where the problem is?

[topaz nymph]

logs?

[rough crow]

Without errors

[topaz nymph]

Ur ram is full

[gilded ice]

I'm thinking it only happens to me

[topaz nymph]

It's this command, it is causing an ArithmeticException somewhere in its logic.

[feral spire]

Container is misconfig’d if you are using a host. Ask them

[rough crow]

And how do I explain it to them?

[peak pulsar]

Tell them to reduce Xmx

[half walrus]

-Xmx controls the amount of heap memory assigned to the JVM, this does not include other memory used by java, or native memory used by other libraries such as netty (for networking) or SQLite. Please do not allocate all of your memory!

[rough crow]

They already did that to me, they reduced my Xmx by 1 GB

[gilded ice]

mhh it doesn't happens without <>

[feral spire]

Y’all fuck up my insurance. Do you know JAVA also need ram bro.
I mean just reduce more 1.5G

[topaz nymph]

I don't know what you mean.

[topaz nymph]

This is a plugin logic issue, not a paper issue.

[gilded ice]

thank you

[rough crow]

And there's no other solution?

[sweet shadow]

Any way to fix this?

Is it a plugin causing it maybe?

[rocky nest]

Send the full thread dump

[half walrus]

Please send large files/logs to a pastebin

[peak pulsar]

you're 20 builds out of date

[sweet shadow]

Oh...

[rough crow]

Sorry to bother you, but my server froze again and this error appeared in the console

[1149.792s][warning][os,thread] Failed to start thread - pthread_create failed (EAGAIN) for attributes: stacksize: 1024k, guardsize: 0k, detached.

[covert jolt]

contact your host

[rough crow]

Sure

[feral spire]

The solution is lower the ram even more so yesh

[arctic oracle]

Really wish hosts showed actual ram usage and not just Java heap

And y'know, made it so users could change flags themselves

[final mango]

Because they have no way to know without invasively instrumenting the java process

As far as the system is concerned, the java process is using that much ram, so that's what the system reports and that's what the panel reads, because that's all it can actually see

[arctic oracle]

Would they not just be able to keep track of the entire container?

[feral spire]

You are giving too much credit for your average MC host

[arctic oracle]

So yes, but no then?

[feral spire]

Correct

[covert jolt]

no because then you'd be showing the same thing that they're already showing

[final mango]

The system will report the size of the allocated java heap (if you're using aikar's that would be the max heap size) plus any overhead.

Always.

[acoustic yacht]

Depends on how deep you opened up your wallet when signing your agreement

:P usually i wouldnt expect them to do anything at all without pressuring them

[vapid prawn]

Doesn't the JVM have a standard instrumentation API?

You just need to wire something up to understand it and force a certain JVM flag to tell it to talk to you

[peak pulsar]

^

[vapid prawn]

I'm saying that's not that invasive

It's not like you need to ptrace it or run a modified JVM

It just has a standard way to ask it "how much RAM you using?"

[peak pulsar]

It's a lot more invasive than just reading the memory usage of the container

[vapid prawn]

Eh, I suppose, but it's just a sidecar

[peak pulsar]

True, but most panels hosts use aren't designed for just minecraft/java servers

[vapid prawn]

So you're saying they'd have to actually do some work?

I guess that rules out most hosts

[peak pulsar]

exactly

[feral spire]

[winged vine]

Sadly no change

[feral spire]

Something built wrong then but I can’t really tell

[shadow oasis]

I have people that keep getting kicked due to keepalive timeout, is this 1.18 beta related?

[half walrus]

https://github.com/PaperMC/Paper/issues/895#issuecomment-372175216

[shadow oasis]

Thanks

[surreal raven]

I'm pretty sure it's worse than 1.17 because some optimizations are not implemented yet (the chunk loader rewrite maybe?)

I "fixed" it by reducing the view/simulation distance

less chunks sent to the player = less chance it's going to timeout

[merry scaffold]

I just noticed a typo in that GitHub comment. Oop.

Meh, not hurting anyone.

[shadow oasis]

That makes so much sense!! It only started after I adjusted view distance....thanks for jogging my memory lol

[brisk canopy]

if you have antixray enabled, you can enable permission antixray bypass and give that perm if you trust struggling player well enough

that's what we did on our friends smp, seem to have mitigated the issues, along with increasing keepalive timeout

[shadow oasis]

Thanks

I don't actually have anti-xray enabled though, so thats not the issue

[brisk canopy]

welp, render distance is still a good suggestion that applies

it's actually a very strange issue in 1.18, seems that client is perfectly able to send the packets and then just doesn't receive anything back..

like players that timeout can send messages to chat and they're immediately sent, however they themselves do not see them, because they still didn't receive it back

[shadow oasis]

My guess is its just a way to optimize the networking. Less data sent means you don't need as high of a connection

[feral spire]

That’s correct.

Until all patch is back in, Thsts a good mitigation

[timid wolf]

what would be the best workload to lag a server

testing purposes of course

[feral spire]

1. Craft a post with everything 4chan hates
2. Post on 4chan
3. Switch IP and reply under your original post saying you found OP’s Minecraft server.
4. Get test result.

[dire tartan]

I usually lag a server by disabling entity cramming and summoning 100+ chickens

That is, slowing tps

Not internet connection

[timid wolf]

yeah might try something like that

[peak pulsar]

brb, gonna test this

[feral spire]

That’s not really organic and doenst really tell much

[timid wolf]

I made it just set thousands of blocks per tick since I couldn’t think of something else

just to test monitoring

[vague falcon]

1 billion TNT

[dire tartan]

From my experience setting blocks is one if the easiest actions

[timid wolf]

someone should make a plugin with a bunch of different workloads for testing

[feral spire]

Repeating commandblock that spawn end dragon in overworld

[timid wolf]

dropped tps to 16 for about 2 mins

[timid wolf]

Only ran for 5s

well really it ran for around 10 or so

but it’s programmed to 5

[dire tartan]

Yeah, I think paper distributes workloads over multiple ticks, same with tnt

The heavy ones, of course

[timid wolf]

when is ram used?

obviously always but I mean like what specifically causes more ram to be used

[feral spire]

Chunk loaded is a big one

But you can’t really quantify things like that

[peak pulsar]

I believe that’s a thing already. Look into Stress by pop (chunky’s dev)

[timid wolf]

ye but that's only a few workloads

[peak pulsar]

I don’t know if it works on 1.18 or not though

Ah didn’t know what you were looking for

[timid wolf]

yeah I meant like a large amount of workloads

even some real world ones potentially

[violet wing]

yeah i mean i basically made that to check chunk loading and entities for the most part (which are the two main concerns I usually have)

players on my server always like to make mob farms and it's nice to just spawn different densities of those mobs to see how much they actually cause lag when spawned in the world

[raven musk]

Hey does anyone have issues with nether causing server crashes?

its a weird issue thats been happening

[topaz nymph]

Logs?

[raven musk]

https://paste.helpch.at/hodaqipuxi.md

thats from yesterday

[dapper lotus]

Line 2188

HushTowns deadlocked the server, gg

[raven musk]

tf that mean

[dapper lotus]

don't mention

[raven musk]

wasnt a mention?

I replied my bad

[dapper lotus]

it was a reply with the stupid mention thing enabled

[raven musk]

my bad

[dapper lotus]

HushTowns tried to load a chunk off the main thread in the perm calculation logic for LP

[raven musk]

hmm weird

[dapper lotus]

when something on the main thread asked for a perm, it was stuck waiting for LP to finish the perm calculations

But, that will never arrive, because LP is now waiting to check a perm on the main thread, but, that requires the perms to be calculated, and it can't calculate the perms because hushtowns blocked the calculations loading a chunk which requires the main thread to be ticking, but it's not and will never happen now

[raven musk]

hmm lemme contact the dev

[dry robin]

is it possible there's a bug where going through a nether portal generates chunks at the overworld coords (instead of dividing by 8) even though it teleports you to the proper coords

or am I going insane

it seems there's now a bunch of regions in the nether at 8x the distance that would make sense

and every bit of this newly generated land is the same coordinate as an overworld portal

worth attempting to replicate this issue?

[dapper lotus]

I mean, it's a potential

that code is only through several years of butchering from everybody

[honest path]

my god theres like 2000 lines in Hellshowers' log dump, and only three (3) obscure mentions of Husktowns plugin. and ElectronicCat (no mention) finds the culprit

WHAT IS THIS WIZARDRY

[timber ridge]

You can usually ignore a good portion of that stuff.

It also cat is a wizard.

[honest path]

yeahbut

three.

way down in teh middle of nowhere

[dapper lotus]

It's simple

The main thread is blocked behind a LuckPerms call

I just scrolled down the thing until I saw what was holding the lock there, was LP, calling HuskTowns calculator, which was loading a chunk

familiarity with the app tells me that that is just not gonna work

ez

or maybe am just autistic

[timber ridge]

Also wizardry.

[dapper lotus]

Wrong discord

[cloud haven]

Ah sorry

[timber ridge]

Yeah try Purpur

[cloud haven]

using purpur atm just to see if it'll help, didn't really notice anything unfortunately

[solemn ether]

My CPU is really good
but my Ram is really bad right now

thats like 10 players XD

[feral spire]

What’s your startup flag?

[solemn ether]

its the normal 1

[feral spire]

I have no idea what that mean

[solemn ether]

java -Xms128M -Xmx8192M -Dterminal.jline=false -Dterminal.ansi=true -jar server.jar

[peak pulsar]

Ram usage means almost nothing if you set it up properly

[feral spire]

So you tell your jvm it can use anywhere from 128M to 8192M

[solemn ether]

yea i did

[feral spire]

So it’s pretty normal and it has nothing to do with your performance

Please understand how java use memory and what you said makes no sense

The allocated ram has zero relationship on your server performance

[violet wing]

most people use aikars flags so their server is permanently at 100% memory usage (not really, but it will display like that in your panel)

[half walrus]

https://startmc.sh/

[violet wing]

it will just all be allocated

[feral spire]

Get yours from here

[solemn ether]

what i mean is the Version is not stable yet lol
and some of my plugins are having memory leaks

im just saying that my ram is on a bad spot lol

[feral spire]

How do you determine you have memory leak?

I’ve seen no sign of it

[uneven pagoda]

How much time Chunky will generate 16k*16k map? Anyone tried this/know?

[vapid prawn]

I don't want to do the math on how many chunks that is but a top of the line system can be expected to do about 90 chunks per second on 1.18

I think that's around 21 million chunks?

So about 64 hours

Oh wait that's 21 million blocks, lol

So about 2.5 hours

That seems way too fast though

[vapid prawn]

So yeah, assuming my math is at all accurate I'd say somewhere between 2.5 and 4 hours on a powerful machine

If you're doing it on some random shared host I'd multiply those by 3-10x

That might be a bit pessimistic but better to overestimate these things

If you just start the generation task chunky will tell you how long it thinks it'll be and then you'll know

Let it run for a few minutes for the estimate to settle

[uneven pagoda]

It's physical server with 8x Xeon on 4 GHz, SSD and 8 gb allocated RAM

[vapid prawn]

If that Xeon is at all recent it'll be anywhere from twice the time I estimated to slightly less than I estimated

iirc Intel CPUs were slightly better at terrain gen, a 5950X was getting 90 cps while a 11900k was 100 cps iirc

I only have single points of reference for either of these though and am going on memory but I haven't seen anyone claim to get better than 100

[uneven pagoda]

Thanks for reply

[solemn ether]

Well it's my plugins
Called Advanced Crates
The Animations were lit that crashed my server 3 times now
Thanks for the aikar-flags
It helps a little bit right now

[slim tree]

Any plugin to show memory usage of the server in tab list

Pls

[dapper lotus]

use a placeholder plugin thing

install an extention to show that

[ocean copper]

took 35hrs to generate a 40kx40k border (so upto 20k from 0), was 6.25m chunks, done at around 40cps

[slim tree]

Thanku

[ocean copper]

pretty much any tab plugin will do this.

Most of them offer tons of extra stuff you may or may not need.

getting a "small" one instead of one that changes the whole tablist might be an idea if you dont want all the extra magic

[swift dust]

using 1.18.1 build 101 and facing issue with bukkit scheduler using more CPU with 0 players , 104 entities and 675 chunks loaded
https://timings.aikar.co/?id=31b3de7acfc94551be233ad100243ba7#plugins

never faced this issue with 1.17.1 😐

[covert jolt]

try a spark sample

[swift dust]

5 minutes. running it

this is 35k x 25x overworld, 10k x 10k nether and end pre-generated world
https://spark.lucko.me/2alY3TtvfO

[dapper lotus]

dynmap is going brr

that's all I can see

[swift dust]

🤦 i just now saw it

but in timings it shows this

Minecraft::Bukkit Schedulercount(18)  total(66.66% 0.763s, 84.80% of tick)avg(42.40ms per - 42.40ms/1.00 per tick)```
so was bit confused

[dapper lotus]

there was some ticks where it took stupid long for stuff

The server also spin waits between ticks so it screws up the CPU usage thing

[swift dust]

oh.. so i wait for dynmap thing to complete and all should be fine ?

maybe i should pause and check again

ya, that seems to be the culprit

[uneven pagoda]

love biome blending

mythical floating village

[topaz nymph]

Ain’t perfect but makes for some interesting interactions with structures

[feral spire]

that's a cool base if you ask me!

[uneven pagoda]

wait it's like a man holding a sword in his hand

and his pet/dog at the right lol

[upper drift]

Hahaha

Cool

[potent pilot]

is there a workaround to barely get working perplayerviewdistance in 1.18.1? Just for myself cause my internet crash with more than 2 viewdistance
(ofc i cant set the serverviewdistance to 2, player would blow up at me)

[feral spire]

Just wait

[half walrus]

https://github.com/PaperMC/Paper/issues/6136#issuecomment-982297893

[feral spire]

Or really upgrade, I cannot imagine playing on a hardware that can barely run 3 vd

Have you try all the fabric optimization mods? @potent pilot

[potent pilot]

yes but thats another problem... in this case its my network, when i join normal server , server is sending ((10+1+10)*(10+1+10)) chunks to me and then my client is downloading them and i get timeout

but might it be possible to handle that with a network packet listener when cancelling chunkpackets?

[dapper lotus]

can't cancel

well, you could basically try to queue them in some manner

[dark wigeon]

there's this monstrosity of a plugin https://www.spigotmc.org/resources/fartherviewdistance.84950/
i've had it on my server and it caused some weird chunk loading, but maybe that's been fixed since then

looks like he updates pretty often

[odd stirrup]

i tried that but it didnt really work that amazingly lol

[dark wigeon]

then why'd you rate it 5 stars lol

[odd stirrup]

fuck it

[dark wigeon]

Fix the problem of block loss caused by prevent-xray don't know if i recommend this lol

[potent pilot]

mh

If the player's view distance is less than or equal to 8, the view distance will be fully provided by the server
looks like it wont solve my problem

[dark wigeon]

so why not have server view distance of 4 or 5, and give everybody else who wants it a view distance of 10 via permission

[feral spire]

Or just wait for leaf to finish the thing

[gusty trench]

It's just that the author isn't a native speaker, the bug was with blocks not being displayed in the client

I've been using it since 1.14, in the beginning of 1.18 the plugin was a bit iffy but it's been fixed in my experience

[grave egret]

works

[dark wigeon]

i had issues where certain chunks would simply not load

relatively sure removing it fixed it

[gusty trench]

Oh, yeah, that's what the anti xray stuff refers to (the plugin's internal anti xray was causing the missing chunks)

[potent pilot]

4 or 5 will need at least 1mbit and my max stable speed is 0,3mbit

[grand raven]

Hey I cant find the option for anti xray for other worlds (Like the nether and end), do I have to type it out manually or something?

[polar quarry]

have you considered giving up?

[polar quarry]

yea, manually add them

heres a good guide https://gist.github.com/stonar96/ba18568bd91e5afd590e8038d14e245e

[potent pilot]

why wdym?

[rose falcon]

I had this issue with the plugin as well, but one of the recent updates seems to have fixed it

[dark wigeon]

yeah looks like the author is pretty consistent with the updates

[polar quarry]

i meant that you shouldn't apply all of these workarounds such as view-distance of 2 if you are on a satellite-tier connection with <0.3 mbps bandwidth, because your player's experience will be so terrible that why even bother going through all of that

[potent pilot]

the server have a pretty good connectiom, just my computer have 0,3mbit

its not satellite, its lte

[scarlet pawn]

measured in gifs of swallowed rats leaf posts, how many?

[peak pulsar]

At least three

[uneven pagoda]

Could we add location to this message? Server floods with it every time after restart and we just can't find this spawner :/

Invalid SpawnData: No key entity in MapLike[{id:"minecraft:zombie"}]

[dark wigeon]

i think it's a vanilla message

[feral spire]

Soon™️

[scarlet pawn]

I seriously expected rat.gif as the reply, but that is good as well :p

[fickle fossil]

So when I upgrade my server to 1.18 the worlds will auto convert correct?

[covert jolt]

yes

[fickle fossil]

sweet ok

[heavy sandal]

Well, I got a lot of errors when joining my vanilla 1.16 world using the latest Build

[covert jolt]

provide the errors

some errors can be ignored

[heavy sandal]

(After which my client even crashed)

(A moment, just starting the server again...)

[noble juniper]

Is anyone else having the prob where %bungee_total%
and %bungee_<servername>% wont load?

[long edge]

no, but reach out to https://discord.gg/helpchat for papi help

[heavy sandal]

@covert jolt Hm, not happening again (but I can restore the old world state if wished), but it was fairly similar to https://github.com/PaperMC/Paper/issues/7131

[noble juniper]

I joined and got automatically banned

wtf

[long edge]

@merry scaffold ^

[dark wigeon]

kek

[long edge]

you've got a new account, they've been dealing with bot raids

[long edge]

would have to get the exact error, check in your logs folder, all previous logs are stored.

you'll need to use a program like 7zip/winrar to extract them though

[heavy sandal]

Already checked that; for some reason they are not

[long edge]

odd

[heavy sandal]

The latest one is from February, the last time the server was online previously (on 1.16)

Wait, I'll just rollback to the snapshot just before the upgrade

[merry scaffold]

I lifted their ban. They gotta join quick before I turn it back on.

Or join on an alt account, that works too lmao.

[heavy sandal]

@long edge Is there a (new) config option to dis/enable the logs?

[long edge]

no

[heavy sandal]

Wait, nvm I am dump

[long edge]

other than a custom logging config or something

ah ok

[heavy sandal]

I was in the mounted backup folder

@long edge https://pastebin.com/V7zh6vrX

@covert jolt Cc

[long edge]

ok yeah that's unfortunately pretty much expected

why your client crashed is likely unrelated, but check client logs

[covert jolt]

those jigsaw errors are the ones you can ignore

[heavy sandal]

Oh, how's that?

[fickle fossil]

Getting a ton of errors with jigsa-

[long edge]

if you weren't the one who stopped it/don't know why it stopped, enable debug in server.properties

[fickle fossil]

oh lol

[heavy sandal]

Nah, I stopped it because I thought something was severely going wrong

[long edge]

ah ok, then yeah you can pretty much ignore that specifically. there aren't any adverse affects

you'll get a similar thing on vanilla as well, pretty much every major upgrade

[fickle fossil]

How long does world conversion take?

My server seems sort of frozen

[long edge]

it's done on chunk load

oh, well it shouldn't be

[fickle fossil]

waht

i'm getting the jigsaw errors and such

[long edge]

as in the whole world isn't converted at the same time

yeah that's expected

[fickle fossil]

it's been doing it for a good minute now

[covert jolt]

world conversion takes more than a minute if you're doing a force upgrade or loading a ton of chunks at once

[fickle fossil]

i just put the jar in, upgraded my plugins, then restarted the server

[dark wigeon]

you're not doing a force upgrade are you

[fickle fossil]

keeps saying this

>....>....[372.241s][warning][gc,alloc] Paper Async Chunk Task Thread #0: Retried waiting for GCLocker too often allocating 256 words [395.560s][warning][gc,alloc] Thread-11: Retried waiting for GCLocker too often allocating 256 words [414.048s][warning][gc,alloc] Paper Async Chunk Task Thread #0: Retried waiting for GCLocker too often allocating 256 words [414.048s][warning][gc,alloc] Worker-Main-1: Retried waiting for GCLocker too often allocating 524288 words [428.859s][warning][gc,alloc] JDA MainWS-ReadThread: Retried waiting for GCLocker too often allocating 256 words [428.905s][warning][gc,alloc] JDA MainWS-ReadThread: Retried waiting for GCLocker too often allocating 256 words

[covert jolt]

you're running out of memory

[fickle fossil]

what do I do?

[covert jolt]

how much memory are you allocating to the server

[fickle fossil]

as much as my host allows, i'm not 100 on it

10gb

but i think it goes over that

[covert jolt]

what host

[fickle fossil]

dedicatedmc

[covert jolt]

would send in a ticket and ask about it

they'll help you out

[fickle fossil]

sure

stopping my server during this won't break it right?

[half walrus]

https://github.com/PaperMC/Paper/issues/6136#issuecomment-982297893

[fickle fossil]

well, i got the dedicatedmc guys to up my memory

and it did start

but if i try to join the server it just turns off

[polar quarry]

what does it say

[sick stone]

Can you create a timings report and heap dump

[fickle fossil]

I can get you guys a latest.log if you want

This is what I get when I try to join, let me get the log

[vivid iron]

latest.log by @fickle fossil: https://paste.gg/86d058d4e83b47688b5e1723506ef0cc

[fickle fossil]

It just sorta shits itself when I try to join

There's also something about a world in a newer version of the log

[vivid iron]

latest.log by @fickle fossil: https://paste.gg/b75e7deb8dc9408b9aaf7a3ca413f2bb

[covert jolt]

update glowapi

or remove it

[polar quarry]

that's a lot of nasty plugins

[fickle fossil]

unfortunately yeah, we've got some problem ones but they're integral features

[sick stone]

You need to remove every not needed plugin and update glowAPI

I would also recommend pregenerating all your chunks

[fickle fossil]

that is the newest version of that plugin

and also all unneeded plugins are not implemented

all chunks pregenerated as well hah

[crude orchid]

any idea why tps is fine at 20, then randomly goes to 10 for 5 minutes, then back to 20

[sick stone]

Create a timings report and heap dump

[fickle fossil]

garbage collection?

[crude orchid]

should i be rebooting twice a day

[sick stone]

The heap dump will tell you the source

[crude orchid]

or would 1 a day be fine

[sick stone]

Yes

[crude orchid]

i can send spark + timings

[sick stone]

With that many players

[crude orchid]

https://timings.aikar.co/?id=54115817794b4a1c8eb363e51738077a

https://spark.lucko.me/42IXT2dFRh

[sick stone]

Ok

[sick stone]

Looks like chunk loading

Have you pregenerated all your chunks?

[crude orchid]

yeah i have

basically i need to wait for leaf to drop that shit

den?

[sick stone]

Okay

@crude orchid

[crude orchid]

ooo me should go change to all those?

lemme see lemme see

[sick stone]

Ignore hopper-disable-move-event

[polar quarry]

those are quite low, players will definitely notice

[sick stone]

This works as well

[fickle fossil]

Well now I can join my server

but the problem is that the TPS is like, 10

[covert jolt]

avoid this

[half walrus]

https://eternity.community/index.php/paper-optimization/

[polar quarry]

no

[covert jolt]

go through the more up to date guide

the bot you're using is completely useless

you do not want to blindly change settings

[fickle fossil]

new log

[vivid iron]

latest.log by @fickle fossil: https://paste.gg/ac0720419f0d4301bde4212660f75d99

[polar quarry]

agreed, understand what they do and think about how your server and players will behave

[covert jolt]

the jigsaw errors can be ignored

[violet wing]

everyones server is always different, like on mine i intentionally ignore a lot of the more aggressive settings that don't actually change performance all that much

like players love making mob farms and stuff, i just tweak the spawning rules so that it works without having a huge impact

[polar quarry]

[16:52:55] [Server thread/INFO]: Done (314.962s)! For help, type "help"

i cry

[polar quarry]

if you have 96 heavy af plugins and it takes your computer literally over five minutes to start the server maybe, just maybe, you're doing something wrong

[violet wing]

world record 😎

[fickle fossil]

bruh mormon

i've got a host and this server runs 20tps just fine

a good chunk of em aren't heavy either

[harsh lance]

PinataParty

[crude orchid]

wbout it

is it bad

lol

[harsh lance]

[crude orchid]

ah

hi fren

good plugin yes

[violet wing]

omg hexed

[polar quarry]

well at least it doesn't crash now. can you post timings now that the server actually stays up?

[crude orchid]

yall want my timings every player milestone, to see how 1.18 jar runs

[fickle fossil]

will in a mo

im about to go out for a bit but when im back

[violet wing]

oh hexed made limitpillagers

[crude orchid]

he make pinataparty

!!

famous

[violet wing]

lots of plugins

[crude orchid]

we looking rough

[violet wing]

is it just generally lagging or is it just spikes

[crude orchid]

https://spark.lucko.me/xibw4xkNcp

mm generally lagging

theres new spark

[polar quarry]

what does this mean?

[violet wing]

plugins look generally pretty light though which is pretty good

[polar quarry]

also does this page have a not-dark mode?

[violet wing]

wait spark has a flame graph now

[covert jolt]

i'm pretty sure it used to have one

[violet wing]

this is a time to be alive

[covert jolt]

probably didn't survive the UI rewrite though

[polar quarry]

whoaaaaa incredible

[covert jolt]

has had one for a few months

but now there's an actual button top right to toggle it

[violet wing]

oh i see

[violet wing]

i swear i updated lately but maybe i havent done any profiles

[crude orchid]

if yall want any specific timings to see general 1.18 performance with decent amoun of players lmk

[violet wing]

oh wow right clicking on it does stuff

yeah ok never knew that i guess im a spark noob

[crude orchid]

12 tps 50 players, mmm

i disable mob spawning until more patches come?

[violet wing]

i had 50 players and i wasnt hurting that bad i feel

is your view distance super high

or simulation

[crude orchid]

view-distance=16
simulation-distance=5

[violet wing]

holy shit XD

[polar quarry]

man to be honest all of this new stuff sucks for people who run servers. i remember running like 120 players in a 1.12 player just fine with like 4 gb of ram

[violet wing]

thats a lot of chunks

[polar quarry]

bad

[violet wing]

i have mine tuned to 4/8

[crude orchid]

oh uhm

[violet wing]

default view distance is 10 im pretty sure

[polar quarry]

sd=6, vd=8 here

[crude orchid]

ohhhhh

should i change that rn

[violet wing]

the VD for sure, 16 is crazy high

[polar quarry]

simulation distance can't go lower than 5 (edit: im wrong, it can go lower in servers)

[violet wing]

10 or 12 maybe you can do if you really like high VD

[polar quarry]

some limitation by mojang i guess

[violet wing]

i thought paper made it so you can do lower

[fervent lynx]

it can go as low as 0 even on the vanilla server software

[polar quarry]

oh cool

[fervent lynx]

it's only enforced to 5 on the slider in the client

[polar quarry]

nice, good info ty

[crude orchid]

ok i leave my simulation at 5

crank vd down to 8

that better?

[polar quarry]

yes! a lot

[violet wing]

i wouldnt go lower than 4 normally though i think, because 4=64 blocks, some achievements are at like 50 blocks

[crude orchid]

just changing my vd, do i need to change anything with mob spawning / despawning

or is that only svd

[polar quarry]

nah, default mob spawning configs work ok with 8 chunks

[fickle fossil]

the server times out every time I try to join

changed simulation distance as well

[native hornet]

differ to logs

servers shouldnt crash

or time out

[fickle fossil]

huh?

[native hornet]

look at your logs

[vague falcon]

I think defer is what they meant

[crude orchid]

changing my svd

and changed some mob spawning values

now we gucci

will report back at 80 playa

[vague falcon]

I think NMS got changed again, just the risks you take lol
Should be #paper-dev tho

[errant ivy]

i have a weird problem with zombie piglins, if the server is online for more then 12 hrs only the 1 piglin i hit gets agro, all other piglins stay idle. when i do a server restart i works again and all piglins get the agro. dont know what to do with that other then restarting the server every 12hrs. is there a solution for this problem???

[covert jolt]

do you have any plugins

[errant ivy]

i have 11 but those are the general plugins that everyone has so we have a pure vanilla play but with some anti grief

in 1.17 i used the same but no problems

[covert jolt]

provide timings

[errant ivy]

as soon as i have the problem again i will do a timing report. i just restarted it again so the gold farm works again

[fickle fossil]

Do superflat worlds have any problems converting?

My host is telling me my server is crashing because it’s trying to convert the worlds and overloading

[sick stone]

Hmm, maybe, try on a singleplayer world and see what the Y is

[spark bobcat]

Can anyone explain me why the villagers will only decrease trade prices on the first conversion only?

[grave charm]

Because it’s a bug that they kept decreasing. Confirmed as a bug on the mojang bug tracker.

Should only get the discount once, not over and over

[spark bobcat]

Yeah, I found how to disable it, thanks anyways

[fickle fossil]

What do you mean by this?

[sick stone]

See what Y level the ground is at, not 100% sure how to help you

[feral spire]

Please report back

Coz super flat world has an increased Y level in 1.18

I remember reading it somewhere

[fickle fossil]

maybe that’s what’s screwing up?

[feral spire]

I don’t know but it may as well be possible

[wise zenith]

Is anyone having issues with his sorting system ?

[grave charm]

With who’s what?

[fickle fossil]

seems my server has stopped crashing instantly but does after a while

still takes fuckin forever to start

[fickle fossil]

being told by my host that it's a memory leak from a plugin

lovely

[feral spire]

how did they know? did you provide heap dump?

server start time has no relationship with the issue you are currently have

[dark wigeon]

is there a crash log?

[fickle fossil]

I've got a latest.log I can give you but I see no crash reports

[timber ridge]

Post the log? Might be stack traces/etc in there.

[fickle fossil]

No, I don't believe so. I can't even get the server to start. It does the Gclocker error stuff. One of em told me that I had pushed my server too far and maybe this was the end or some shit lmao I'm just getting grimness

[fickle fossil]

I'll send the log in just a second. I've got a bout a hundred and I'm starting the server again I'm sure it'll crash in about 5 min

[fickle fossil]

#❗-1-18-experimental message
This is one from earlier. It's pretty much if not the same exact error at the end there

I may be mistaken though. Scoping through I don't see the one I'm seeing now. Good lord this has been confusing and stressful

[vivid iron]

latest.log by @fickle fossil: https://paste.gg/104c9c62dc844625ba97a653cc41d5be

[fickle fossil]

the latest log from about 30 sec ago with the error

again, I'm being told it's a memory leak from a plugin but I'm just not sure. Who knows at this point. I've been running servers for like 7 years now and I still learn something new every day

[dapper lotus]

You'd need heap dumps

[fickle fossil]

How can I get a heap dump

[dapper lotus]

well

/paper heap

or, there is the Jvm option to dump on an OOM

[dark wigeon]

well if there's no error it's probably the container being killed, no?

[fickle fossil]

Not sure how to do the latter. since my server won't even start i can't imagine /paper heap will work

[fickle fossil]

What would that mean?

[timber ridge]

Well I'd disable DeluxeMenus to start, as it's throwing errors. You have a bunch of worlds in multiverse that apparently don't exist.

[dark wigeon]

you will have to talk to the host

[dapper lotus]

i.e. the process is being killed by the host

I mean

at least

[timber ridge]

Seems like a few things are broken/misconfigured. Try fixing those and see if the server will start properly.

[dapper lotus]

it seems that your host has half a clue and is pissing in the wind to come up with a problem that isn't theirs

[fickle fossil]

I hate to be like this but all of those things have been like that besides say some of the new errors from these new plugin versions. Surely it's not the configuration?

[dark wigeon]

you could also try downloading a copy of your server and starting it on your own computer to prove that it's not an issue with the server itself

[fickle fossil]

true, I may try that. I do have a localserver

could take days to download all this though, agh

[timber ridge]

You don't really need the world file.

Your issue is likely not the map.

[dapper lotus]

I mean, Deluxe failed to load, it's not gonna yeet a full server out nder its own rug

[dark wigeon]

kinda do if you're converting it though

[fickle fossil]

I appreciate this kind of honesty. I'm paying a bill and was told that "1.18 was the straw that broke the camels back" like this thing hasn't been running for years with no issue lol

[fickle fossil]

That's what I had thought

[timber ridge]

Which host is it?

[fickle fossil]

DedicatedMC

[dapper lotus]

There is a vanilla issue in 1.18 that causes a memory leak but that's already patched in paper

and plz no mention

[timber ridge]

Yeah for sure. But you're gonna have to fix that stuff eventually anyway. If you're trying to figure out what is causing crashes then you may as well fix some of the other stuff also to at the very least reduce the logspam and eliminate variables.

[fickle fossil]

that's a fair assessment

i'll remove it and see what happens, not like it'll make it worse

[dark wigeon]

could try removing plugins, maybe regenerating server configs

backup first though

[fickle fossil]

i'm doing the plugin removal thing and getting nowhere

i've got a backup from 1.17

[timber ridge]

But yeah if it's just dying with no real error then most likely the host is nuking the thing. Maybe they're allocating too much RAM to the instance?

Are you able to modify your startup flags?

[fickle fossil]

hrmmm, we upped it temporarily to test it and it actually ran pretty fine but then another staff member for dedicated shut that down lol

I believe so, maybe

I can check

[timber ridge]

Which host is it again?

Oh you said

[fickle fossil]

dedicatedmc

[timber ridge]

Yeah they're less likely to have MC running over on the container

[dapper lotus]

I mean, 1.18 does increase the size of chunks "a bit"

[timber ridge]

So that's probably not likely

But hard to say.

[dapper lotus]

but, you said gclocker

[fickle fossil]

i'm not sure what all this stuff means so bear with me

[dapper lotus]

which, gc locker implies some form of memory pressure

[dark wigeon]

actually, are you close to using all your storage?

[dapper lotus]

so, it could be that you don't have enough ram

[fickle fossil]

i can modify my startup settings

[dapper lotus]

But, for it to blow up so early in those logs, er...

[fickle fossil]

they're unfortunately not gonna give me more ram, i think i'm on the highest plan available. i think i may be able to pay for a bit more but i'm not sure

[dark wigeon]

you can also try turning off keeping spawn loaded for all worlds in multiverse config so it doesn't load the spawn chunks

[dapper lotus]

how much ram do you have?

[fickle fossil]

i don't think so, i never really had any issue with it before 1.18

[dark wigeon]

it's 10G for highest plan

[fickle fossil]

10gb

[timber ridge]

10GB seems like it should be enough.

Unless multiverse is loading five hundred worlds.

And/or you're running a bunch of other stuff on the box.

[fickle fossil]

we've got a pretty big number of worlds but it's not five hundred and most are pretty damn small

[dark wigeon]

#❗-1-18-experimental message
try this

[timber ridge]

Ballpark how many?

Yeah that's a good place to start if there's a bunch of them.

[fickle fossil]

good question, let me see

[timber ridge]

If that's really what's nuking it.

[fickle fossil]

maybe 20 tops? likely a bit less but i can't see it being much more

[timber ridge]

I'm not 100% sure how much RAM each world adds in multiverse. But I'd try gecko's suggestion and see if it at least starts and runs cleanly.

[fickle fossil]

i was comically wrong it's like 27 but again most are like void worlds with a utility for a plugin or like 1k radius. that being said i've literally never had a problem before besides lagspikes here and there from plugin stuff

[fickle fossil]

lemme see what that'll do

gonna download some worlds that are not used often and take em out for now

[amber creek]

What if you rename the plug-in folder and start the server then will it start?

[fickle fossil]

I'll try that next. seems a little drastic lol

[fickle fossil]

Where is this located in the config

[dark wigeon]

worlds.yml

find and replace keepSpawnLoaded: 'true'

[amber creek]

If nothing works, go back to start and begin from there.
Then try adding your plug-in back one by one and see when it fails

[fickle fossil]

that's what i've been doing and have made little to no progress

[feral spire]

[fickle fossil]

Not seeing this as a setting

is it KeepSpawnInMemory?

going to assume it is since it's the closest thing to that

[fickle fossil]

With that on it still throws the error. No change
`[02:06:27 INFO]: [HeadDatabase] Successfully loaded 35597 heads!
[02:06:27 INFO]: [HeadDatabase] Successfully loaded 18 featured tags!

....[158.776s][warning][gc,alloc] Paper Async Chunk Task Thread #0: Retried waiting for GCLocker too often allocating 524288 words
[158.789s][warning][gc,alloc] Server thread: Retried waiting for GCLocker too often allocating 524288 words
[158.832s][warning][gc,alloc] Paper Async Chunk Task Thread #0: Retried waiting for GCLocker too often allocating 256 words`

I'm going to have to take a break with this but if anyone can please help me figure this out I will literally paypal you I am so stressed about this. We're hoping to release 1.18 by new years and my server simply will Not start

host has closed my ticket so i'm on my own

[dapper lotus]

gclocker screaming generally implies too much memory pressure

[fickle fossil]

so what does that mean i should do?

[dapper lotus]

I mean, reduce pressure

[timber ridge]

I wonder if multiverse loads those worlds at first regardless.

[dapper lotus]

increase Xmx

try to reduce memory usage

I mean, on the surface it's hard to say

[timber ridge]

Try removing some worlds and unused plugins.

[dapper lotus]

all we can see is that memory pressure is high

[fickle fossil]

i think the world conversion is maybe taking up all the memory?

[dapper lotus]

Or, at least, signs alluding to that

[fickle fossil]

host said it was a memory leak from a plugin causing that

[dapper lotus]

host just seemed they where pissing in the wind

[fickle fossil]

fairly true, i got told that this was like the end lmao

[dapper lotus]

You defo have memory issues but I'd take what they say with a grain of salt

[fickle fossil]

fair enough

[timber ridge]

Yeah it could be a memory leak but just act on what you can see for now.

[dapper lotus]

Like, at this point, you basically need to start hooking up a profiler and seeing what's going on

[timber ridge]

Try to reduce the memory pressure/footprint and see if you can get it started.

[dapper lotus]

but, nature of a shared host, good luck on that one

[timber ridge]

Yeah.

I'd remove some worlds from multiverse - as many as you can.

If you can get it started with 1 world then that's a start.

Then add some worlds back in slowly while checking it/etc.

If remove most/all the worlds doesn't fix it then it's probably another plugin. Back to binary search.

You're kind of hamstrung a bit on a shared host for what you can debug using proper tools.

[fickle fossil]

What good would the profiler be? hell would it even be worthwhile if i can't even get the server started

hmmm. may download all the plugins to my local server and run it to see if its a plugin or a world

[dapper lotus]

I mean, am a yourkit fan, but, visualvm, jprofiler, etc

[timber ridge]

Seems like it'd be easier to disable worlds and try restarting than downloading everything, but sure. Try both.

[fickle fossil]

How would I disable the world without deleting the file?

[dark wigeon]

you changed it to false right

[fickle fossil]

yes

[dark wigeon]

you can also try changing the keep-spawn-loaded setting in paper.yml

[timber ridge]

It looks like mv loads the initial world regardless. So there's likely some memory overhead just having it in the config.

I'd think you can remove it from a config to have it not load without "deleting" anything.

But I'm admittedly not very good with multiverse

[dapper lotus]

I mean, rename the mv jar and they won't load

[dark wigeon]

Correct, you can import it later, removing it from config doesn't touch the folder

Or do that

[dapper lotus]

each of the worlds has a config option which lets you disable worlds from loading by default

[timber ridge]

Yeah that's probably a good start.

Yank mv entirely and see if it starts.

[fickle fossil]

will it break my worlds though?

[timber ridge]

It shouldn't, but make a backup first if you're worried about it.

[fickle fossil]

right okay

well

i've got a backup from 1.17 so

i suppose it'll be fine

[timber ridge]

There you go.

[fickle fossil]

running it now

a couple plugins will likely disable due to MV not being installed

well

it loaded

and then it turned itself off

https://youtu.be/TnlPtaPxXfc

[timber ridge]

Log this time?

[fickle fossil]

sure thing. Just a sec. It just.. stops itself

[timber ridge]

I don't love seeing it as a .gz

Why isn't it latest.log?

[fickle fossil]

I had restarted it again with MV-Core

I can do it again if you want it to be latest.log

[timber ridge]

No, it's okay.

I have to step away.

Only thing I can really suggest is disable plugins until the server starts up and try a profiler if you can.

Good luck!

[fickle fossil]

fair enough

hopefully we can get it sorted

[fickle fossil]

It's not any plugins. Ran it through localhost and they all enabled and server runs fine

at least it's something

[timber ridge]

I would say that running it locally with all plugins and it not running out of memory isn't a comparative test for the host environment setup.

If it were me I'd still yank a bunch of plugins until the thing started. I have a sneaking suspicion that if you disable most/all plugins and multiverse the server will start successfully. But I'm always open to being wrong.

Running out of memory is an environment issue and you've changed the environment. You could use your local setup to more easily profile things and find out the memory usage / see if there's any leaking and so forth, though.

[fickle fossil]

That was sort of my idea was to see if anything was causing a memory leak because surely if it caused one on my actual server it would cause one in my local one, no?

This test I ran wouldn’t necessarily find the issue but at least rule out memory leak unless I misunderstood

[round widget]

You can’t really rule out "there is a memory leak" at all, because that would mean you covered all possibilities which could cause one

[timber ridge]

What SirYwell said. You also don't know if it's a memory leak. It may just be taking up too much memory, which you can't replicate easily on a different environment. Again: what I'd do is remove plugins to see if it will start. It's pretty easy to disable plugins and test.

[merry drift]

How far can I say it safe to use 1.18.1 now?
I’m running on a live server for few days with plugins,so far there no issue

[inner crypt]

Pretty safe I wanna say, else this channel would be on fire

There's 70k+ servers on paper 1.18 last I checked

[merry drift]

I see

[vapid prawn]

https://minidigger.github.io/paperstats/#1.18.1 still says 56k, guess it hasn't updated?

[inner crypt]

I said 1.18!

[vapid prawn]

Wow there really are 20k servers still on old enough builds to be 1.18

[feral spire]

There are people running build 1

For months

[inner crypt]

Yes

That's expected at this point, nothing we can do

[feral spire]

Actually you can spam print console message. They sure will notice

[clear steeple]

glad this guy is still running 100 leegacy plugins

[vapid prawn]

Crazy, for 1.18 there was only 1 day where Spigot had more servers, the next day Paper had almost 10x as many

Well, 2 days but the first day was 0 Paper and 3 Spigot, probably just md5

[tidal vigil]

when is paper stable for 1.18.1

[half walrus]

__There Is No ETA__

Updates to Paper do not have any sort of estimate for when they release, ever. Any and all updates will arrive when they are ready, and the only thing to do is wait for them patiently along with everyone else.

ETA means 'Estimated Time of Arrival'!

[feral spire]

Use with backup

[uncut oyster]

What does it mean, when plugin is legacy?

[feral spire]

It’s calling old shit. Replace it with updated version or find replacement

[uncut oyster]

Than how i know if its legacy? 😁

[feral spire]

/pl

One with *

[uncut oyster]

And its written with something old code or how it could be detected?

We have EpicGuard with * but whats different if it works correctly?

[dapper lotus]

depends on the plugin

anything touching the world with a * should be replaced

well, the world, items, etc

[feral spire]

Going by that name, I bet it does

[visual talon]

i'm not quite sure if this is the right place to ask, but are spectators supposed to make any noise?
if I set myself on fire while in survival, then (whilst still burning) change into spectator mode and enter a body of water, my friends report hearing the sound of a player extinguishing (which isn't ideal for snooping around)

[feral spire]

Probably

[visual talon]

woo, wonder how much plugin fuckery i'm going to have to go through to fix that one

[feral spire]

The best sneaking tool @visual talon https://github.com/mbax/VanishNoPacket

You won’t even be shown on server to your friends

[dapper lotus]

I mean, much stuff like that is generally gonna be "test upstream"

Sounds odd but it's not like mojang has a sane mechanism for dealing with that stuff

[feral spire]

I don’t even know how it is suppose to behave in vanilla

[dapper lotus]

it's mostly just hardcoded checks all around the place and 0 context of stuff down the chain

[visual talon]

I'll have a look - does this show me leaving the server when vanishing?

[uncut oyster]

Ahh, its doesnt defined a api-version in plugin.yml

Than it have *?

[feral spire]

Configable

Have leaving message is better. Looks more real

[visual talon]

I don't want to randomly leave the server whilst still talking to them lol

[feral spire]

https://dev.bukkit.org/projects/vanish/pages/configuration

You can set it either showing fake announcement or not

[visual talon]

do I just erase the messages from the config to silence it?

[feral spire]

The page explains it

[visual talon]

don't think it answers my question?

oh its controlled through a command

[feral spire]

It “can” be or also be permission based as explained on the page

[visual talon]

oh i really am blind

[feral spire]

[uneven pagoda]

What caused this? https://paste.gg/p/anonymous/a0847ac4edd44367a501e2d1c3dcfae6

I mean this is ProtocolLib-related but we don't use anything that modifies ClientboundLevelChunkPacket's

Or, stop, maybe this is Orebfuscator

[dapper lotus]

ConcurrentModificationException

something is modifying data in the packet as it's being wrote to the buffer

[uneven pagoda]

Yea, it remains to me only to understand what caused it... Player 1 was in spectator mode watching player 2, when suddenly player 1 was kicked with this error, and player 2 continued to dig the stone

[thick spruce]

Just curious, what happens with generated chunks when switching to 1.18? Will the area below 0 filled with bedrock?
I hope it's not much of offtopic

[gusty trench]

The area below 0 will be filled with the new caves

[thick spruce]

Oh? Good to hear.. thanks!

[cinder goblet]

I just dont get why dont you guys just want to make paper as a hardfork

[gusty trench]

Freedom is a big reason

[plain solstice]

It's a lot of work and it's currently very easy for switch to paper as it's a drop in replacement

[upbeat star]

if I can ask a quick dumb question: When updating a 1.17 world, is it normal for mobs to spawn mid-air and drop to their deaths? Meat-rain

I deleted all the chunks in the world that had been occupied for less than an hour, and updated to 1.18. This completely changed the biomes on the surface too. I'm guessing the mobs' positions weren't updated along with it?

[elder jetty]

You can always change the spawn point: https://shockbyte.com/billing/knowledgebase/133/How-to-Set-Your-Minecraft-Server-Spawn.html

This is something I plan to do when we’ve finally updated to 1.18

[upbeat star]

I don't think that has anything to do with it though?

[elder jetty]

Oh. Oops for mobs. Yeah that sounds like a bug

[upbeat star]

Thanks though

[elder jetty]

Maybe a Z value wrong somewhere?

I wonder whether it was adjusted for the new world height?

[upbeat star]

since cows are spawning above oceans too, My suspicion is that the terrain is completely different, but mob spawns are going off old terrain somehow

[elder jetty]

I now have that image of whales and petunias spawning and asking existential questions

Makes me want to re-read the hitch hiker’s guide

[upbeat star]

LOL yes

not again! ha

[elder jetty]

Frig, they need to do that for real

Maybe on the 42nd day of the year?

[real bolt]

After I updated from paper-103 to paper-105, beacons seem to have stopped giving buffs to players

[long edge]

something else must have caused that

no chance updating 103 to 105 was it, nothing changed that could have remotely done that

[real bolt]

Oh alright

[timber ridge]

Try reproducing without plugins

As a general rule.

[random nacelle]

Just here to drop a random thank you to the whole paper team for their hard work! ❤😁

And of course everyone else who contributes in any way ^^

(Oops, wrong channel for this though. I was certain I was in general channel.. Mobile discord, what are you doing?)

[violet wing]

interestingly some player complained to me about beacons yesterday as well

i didnt investigate it though

lol

i dont think im on paper 103 yet though

[long edge]

I'll look, but definitely isn't 103-105

[violet wing]

im on paper 81

dont shame me

[topaz nymph]

Well there was a beacon issue on 1.18 - Prolly not relevant here, seems to be a much earlier build.

[real bolt]

Huh, I downgraded back to paper-103 as a test and beacons still aren't working. For now this isn't much of an issue though

[warped thistle]

Beacons work fine for me

[violet wing]

i'm thinking it could be related to simulation distance tbh

[fickle fossil]

I've gone through and picked out every plugin that I can't confirm works properly

so it can't be a plugin for my issue

[feral spire]

Is your beacon range larger than your simulation distance?

[fickle fossil]

I'm going to try downloading each world individually and upgrading manually through vanilla

i imagine this will be a several day process but if it lets my server function lmao

[peak pulsar]

Is there a way to disable the -y world generation when importing an old world?

[clever topaz]

below-zero-generation-in-existing-chunks per-world setting in spigot.yml