#Self-ran Server Issues

22 messages · Page 1 of 1 (latest)

grave coral
#

Have you checked the forum and wiki to ensure your issue hasn't been resolved? (Y/N): Y

What launcher are you using? (Curseforge, ATLauncher, Prism Launcher...): Prism

Which modpack are you using? (BCG+, BCG+NC, BCGHG):
BCG+

Have you modified the pack in any way? (i.e. Adding or removing mods):
No

What version of the modpack are you using? (1.1.16HF, 1.1.15, 1.1.14...):
BCG+ 2.0.0

What version of the Java runtime are you using? (Java 17, 18, 21...):
21

Have you allocated 6GB of RAM or higher? (Y/N):
Y (8GB client side, 8-16GB for the server while testing for changes)

If playing with others, what method is being used to host/join the server? (Official server support can be found here: https://discord.gg/playcdu):
Host machine, DNS setup, etc

If playing with others, have you confirmed that both the client and server are running the exact same version of the pack (including any HF releases)? (Y/N):
Y

Describe the issue:
Started a fresh world after updating from 1.1.14(ish), so nothing carried over. Loaded the world up and started playing, just myself and a friend. The server is able to and has ran the older versions of the pack just fine, virtually zero lag or any other issues. Suddenly being met with massive lag spikes and interacting with the shop keeps in the Pokemart also leads to 30s+ in a lagspike. (No crashes yet)

Steps you've taken in effort to resolve this issue:
Next to none outside of adjusting the RAM. Not sure if/what should be disabled or changed to try to better the situation. Any and all help is greatly appreciated! <3

stuck herald
#

Try downgrading your Java to 17, while its always nice having the latest and greatest BCG+ is most likely tested on a JDK of 17 also try running /spark profiler for 10+ minutes

#

when u run the spark send the results here

#

/spark profiler start and then after ten minutes /spark profiler stop

#

try to reproduce whatever is causing you to lag that way the spark can do its best to pick it up

grave coral
#

Sorry I should have updated the post. Removing ModernFix (which was eating up 70% of the javamain thread) has alleviated most of it. I'll need to look at the shops and stuff still.

stuck herald
#

ok good

grave coral
#

After an RTP, generating new stuff in the overworld - just a massive spike that kicked me lol

stuck herald
#

send the spark link

grave coral
#

that's the last decent one i could get

stuck herald
#

So it looks like /lib/ld-musl-x86_64.so.1 is taking up a lot of the server thread multiple times

#

a significant amount of time is being spent in the dynamic linker/loader library which is responsible for loading shared libraries and linking them at runtime

#

it could be any number of things

#

but if you had to ask me its related to lithium

#

its definitely bottlenecking somewhere just not sure where

#

unfortunately since its related to the native library probably goes deeper like into how musl libc is interacting with the dependencies of the minecraft server

#

I would see if you can remove lithium and see if that helps but if it doesnt i would suggest moving off of Alpine linux

grave coral
#

Haven't been able to change much, just reduced 'Mon spawns and added Observable to see where some of the upticks were coming from and haven't gotten any closer to solving it. Not gonna be able to step away from Alpine but may try removing lithium. Thanks for the help!

rocky steeple
#

also having the same issue