Dedicated WindRose server on Linux/Wine rubberbands heavily, extremely poor performance when interacting with entities, chopping trees, etc; repeated ServerMove timestamp expirations, multi-second commit stalls, and internal check/callstack event
Environment
- WindRose dedicated server version: 0.10.0.3.104-256f9653
- Host OS: Ubuntu 24.04.3 LTS VM
- Host size: 12 vCPU, ~64 GB RAM
- Hosting: AMP container, host networking enabled
- Wine runtime: wine-staging 11.6
- Direct connection port: UDP 7777
- Max players: 4
- NetServerMaxTickRate: 60
Symptom
- Players experience severe lag/rubberbanding on the dedicated server.
- This was reproduced while a player was connected live on April 22, 2026.
- Note: timestamps inside R5.log are UTC. The host itself is running in PDT.
What I observed
-
At 2026-04-23 00:25:57 UTC, the server logged a large internal callstack in R5.log.
-
Immediately after that, the game status changed from Fine to R5CheckHappened.
-
After that point, the log repeatedly reports R5Error Report Calls TotalNum 1 / UniqueNum 1.
-
During live play, the log repeatedly emits LogNetPlayerMovement warnings like:
- ServerMove: TimeStamp expired
-
These come in bursts around
-
In the same general window, the server logs repeated slow commitT tasks and other hitch-style warnings:
- commitT slow tasks around 805 ms, 938 ms, 1375 ms, 1235 ms, 7646 ms, 1854 ms, 1817 ms, 1764 ms, 1285 ms, 1717 ms, 1352 ms, 1012 ms, 912 ms
- TaskGraph warnings up to about 170 ms
- intermittent GC lock warnings
-
Earlier in the same log, the game also reported memory pressure warnings climbing from 91% to 99% of 64302 MB.
Expected result
- Stable player movement and world interaction with a low-player-count dedicated session.
Actual result
- Movement updates appear to expire server-side during play, and players experience severe lag.