Twilight Forest Crashing
Mysticlxl opened this issue ยท 5 comments
Bug Description
When I am traveling around the Twilight Forest, I am usually fine for a bit, but then my server decides to crash.Did this behavior use to work in the previous version?
N/ASteps to Reproduce (for bugs)
- Walk around the twilight
- Continue walking
- Wait for a crash
Logs
- Client/Server Log:https://pastebin.com/7XL8eDEH
- Crash Log:
Client Information
- Modpack Version: 3.1.2 Hotfix-1
- Java Version: 1.8.0.51
- Launcher Used: Twitch
- Memory Allocated: 7GB
- Server/LAN/Single Player: Server
- Optifine Installed: N/A
- Shaders Enabled: N/A
World Information
- Modpack Version world created in: 3.1.2 Hotfix-1
- Additional Content Installed: N/A
Server Information
- Java Version: 1.8.0_252
- Operating System: Linux (amd64) version 3.10.0-1127.el7.x86_64
- Hoster/Hosting Solution: MC Pro Hosting
- Sponge (Non-Vanilla Forge) Server: N/A
This is a performance issue. A ServerHangWatchdog crash is caused by a server skipping more than a certain amount of ticks (in this case, 60 seconds), due to it being overloaded. It's recommended to go to server.properties and set max-tick-time to -1 to disable the watchdog entirely.
If your able too. try adding -Xss1M before your -Xmx java arguments.
I had a similar issue when using OpenJDK on my VM and it stopped occurring after adding this, From what i was able to gather after getting a HeapDump and following the stack-trace it stemmed from -> (In case this might be helpful to someone)
"Chunk I/O Executor Thread-1" Id=<#id> WAITING on java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject@256b3495 at sun.misc.Unsafe.park(Native Method)
The thread is waiting on a ReadLock (as well as others) however the same thread held a WriteLock, Could only assume the thread was ignoring or swallowing its own SOE so the WriteLock is never released.
