Galacticraft Performance on Dedicated Server

Lenk

Member
Feb 3, 2019
2
0
1
44
Hey All,

I am running a dedicated forge server for a small group and have Galacticraft included as one of the mods. However, I am seeing pretty significant performance issues with Galacticraft loaded with constant "Can't keep up" messages and even a full stall report every once in a while. With the Galacticraft JARs removed, everything works great. I thought it might be realted to a specific GC block or item so cleaned up the map of most everything but am still seeing issues. Grabbing a snapshot via Sampler I am seeing a significant amount of time spent in "micdoodle8.mods.galacticraft.core.tile.TileEntityCrafting" class with the TPS hanging around 12.8.

I don't know enough about Java programming to really diagnose the issue here. Judging by the method names it seems like it has something to do with recipe registration with the base Minecraft engine and other mods. Would really love to get some assistance on this one as I love this mod but need to get my server running better. I have a pretty beefy system running the server; it should really be able to scream running this workload but even it is struggling.

Details
Chassis: Physical System (Not a VM\VPS)
CPU: Intel Xeon E3-1276 v3 @ 3.60GHz (Quad Core, 8 logical cores w/HT)
RAM: 32 GB
Storage: SSD
OS: Windows Server 2016 x64 1607
Java: 1.8.0_201
Minecraft: 1.12.2
Forge: 14.23.5.2810
Galacticraft: 4.0.2.190 (though 181 also tried with same result)
Net: 1Gbps symmetric fiber
Java Args: -Xms20G -XX:+UseG1GC -XX:+UnlockExperimentalVMOptions -XX:MaxGCPauseMillis=50 -XX:+DisableExplicitGC -XX:TargetSurvivorRatio=90 -XX:G1NewSizePercent=50 -XX:G1MaxNewSizePercent=80 -XX:InitiatingHeapOccupancyPercent=10 -XX:G1MixedGCLiveThresholdPercent=50 -XX:+AggressiveOpts -Dfml.queryResult=confirm

Sampler Snapshot:
sampler.PNG
 

Lenk

Member
Feb 3, 2019
2
0
1
44
Determined this was being caused by the magnetic crafting table. We only had two on the entire map with hoppers feeding them (two input hoppers and one output). As soon as we removed the magnetic crafting tables the TPS shot up from 6 right back to 20 and all the server overloaded messages stopped.

Prior to removing the blocks, the server thread seem to stall on ever other tick. I would get a warning about a tick taking about 60s and then one taking 5s and then another about 60s again repeating. Also, as time went on it progressively got worse though memory usage stayed pretty constant so probably not a memory leak.

Code:
02/12/2019 22:03:54
22:03:54 Can't keep up! Did the system time change, or is the server overloaded? Running 58886ms behind, skipping 1177 tick(s)
22:04:00 Can't keep up! Did the system time change, or is the server overloaded? Running 6242ms behind, skipping 124 tick(s)

I don't have a GitHub account, can someone get this over to their issue tracker?
 

Share this page