To whom it may concern,
I am an owner of an Australian server using Spongeforge and have come across a certain issue that generates no crash report. The issue occurs every 10 to 20 minutes consecutively after the server has started and lowers the tps of the server to below 1/20 and starts generating an ms reading of over 10,000. These readings were taken using the commands “forge tps” and “gc”.
Directly after this event the server freezes all players and actions on the server for 5-10 seconds then disconnects all players straight after. The server then proceeds to return to normal after 30 seconds and players can log back in again as normal.
I have replicated this issue on spongeforge versions “1.10.2-2202-5.1.0-BETA-2045”, “1.10.2-2202-5.1.0-BETA-2042”, “1.10.2-2202-5.1.0-BETA-2039” and “1.10.2-2202-5.1.0-BETA-2020”. This issue does not seem to occur with spongeforge version “1.10.2-2171-5.1.0-BETA-2010” or prior.
Current modpack: Mech and Magic [ATLauncher]
Server Ram: 3.85 Gb
Server CPU: Intel® Xeon® CPU E5-2620 v2 @ 2.10GHz, 2 cores
Operating System: Ubuntu Linux 16.04
The other plugins the we are using are grief prevention, nucleus, luckyperms and catclearlagg. We did disable these plugins while trying to narrow down the cause of the issue and found out that it was spongeforge.
I have also contacted Y0ungSandwich the creator of the Mech and Magic and they seem to be experiencing the same issue aswell.
If you request anymore information please do not hesitate to contact me.
Kind regards,
Quarter_Master.
Thanks for the detailed bug report!
The version information you provided should be helpful in tracking down what’s gone awry. There is something else you could do to help figure out what’s dragging things down… there is a Timings Module built into Sponge implementations, and if you manage to capture the lag event in a timings capture, it may point to whatever is going berserk. If you manage to do so, post a link to that here and it might help expedite things.
EDIT: Try the latest SpongeForge (2046), it should have this fixed.
Thank you for the reply,
I have been trying to run the command but the tps gets so low that I cant even retrieve an answer from command line. Ill keep on trying though. I am currently testing the issue on spongeforge “1.10.2-2202-5.1.0-BETA-2045”.
Finally got the report:
https://timings.aikar.co/?id=9c3c4f3e625a41a28f733eebbe7f89e4&raw=1
I hope this help. This report was generated with “spongeforge-1.10.2-2171-5.1.0-BETA-2010.jar”. This was the version that was working but doesnt appear to be working anymore.
Rawdom my Co-Owner also found this on the console when running spongeforge version “1.10.2-2202-5.1.0-BETA-2049” as the error was occurring.
“[mixin]: Method overwrite conflict for getSlotProvider in mixins.common.core.json:item.inventory.TraitInventoryAdapter, previously written by org.spongepowered.common.mixin.core.inventory.MixinInventoryLargeChest. Skipping method”.
It’s very strange that you had no problems on the earlier version (SF 2010), and yet now you do. It could possibly be some form of world corruption. Does this happen with a fresh world (with a fresh sponge config?)
The Tile Entity tick is pretty high in that timings slice, but it doesn’t look like the main issue. It appears the server times out due to some type of lag, which causes the watchdog to force end it. You could raise the max-tick-time in server.properties if it occurs on startup.
The only other approach you could take is to strip mods out systematically to see if the culprit(s) can be identified.
That’s an internal Sponge message - you can safely ignore it, as it has no effect on the game itself.
Had a similar issue with builds 2049 to 2053, updating to 2056 solved it. Try running the latest available.