Hello,
I am experiencing a reproducible issue with a Palworld Dedicated Server running under AMP.
Environment
- Windows Server 2025 Standard
- AMP 2.8.0.2
- latest Palworld Dedicated Server
- NTFS
- Microsoft Defender exclusions configured
- no third-party antivirus
Issue
The server starts and runs normally.
After some time (sometimes minutes, sometimes several hours):
- all connected players disconnect simultaneously
- the dedicated server process keeps running
- network ports remain open
- the server console reports:
Failed to save. Failed copy from backup.
After this point, no further successful saves are completed.
The SaveGames directory contains:
Level.sav.new_tmp
LevelMeta.sav.new_tmp
Players\*.sav.new_tmp
.atomic_save_update_manifest_world.json
world_save_temp
world_save_bak
This appears to indicate an interrupted atomic save transaction.
Already tested
- completely recreated AMP instance
- multiple new worlds
- fresh Palworld installation
- Defender exclusions
- verified NTFS permissions
- ProcMon trace
- upgraded AMP to 2.8.0.2
- disabled automatic backups
- increased AutoSaveSpan
- enabled/disabled RCON
- enabled/disabled Community Server
None of these changed the behaviour.
Additional observation
A completely separate SteamCMD installation outside of AMP remained stable for several hours.
Running the same server through AMP eventually reproduces the issue.
I cannot determine whether this is still related to AMP or whether it is a Palworld save-system issue, but I wanted to report the reproducible behaviour.
