System Information
| Field | Value |
|---|---|
| Operating System | Linux - Debian GNU/Linux 13 on x86_64 |
| Product | AMP ‘Proteus’ v2.8.0.8 (Mainline) |
| Virtualization | VMware |
| Application | Rust Dedicated Server |
| Module | RustModule |
| Running in Container | No |
| Current State | Ready |
Problem Description
Issue
Rust starts successfully without uMod, but reproducibly crashes during world startup when uMod is installed through AMP. No third-party plugins are installed.
Environment:
- Debian GNU/Linux 13, x86_64
- AMP 2.8.0.8, build 20260927.1, Mainline
- 12 vCPUs, approximately 64 GB RAM, no configured instance RAM limit
- Rust version reported by “status”: 2634
- Oxide Rust extension: 2.0.7810
- Oxide Core: 2.0.4143
- Unity version recorded by Oxide: 6000.3.15x1-13
- Procedural Map, size 4500, seed 2019194420, PvE enabled
The failing startup ends after:
Generated ocean patrol path with node count: 399
[25.6s] Ocean Patrol Paths
AMP then reports:
The application stopped unexpectedly. Exit code 139
Tests already performed:
- Created a fresh instance: vanilla starts and allows players to connect.
- Enabled uMod and updated through AMP: startup crashes.
- Disabled uMod and validated the server through AMP: vanilla starts again.
- Renamed UMod-Rust.zip and updated with uMod enabled. AMP created a new archive, but the crash persisted.
- Oxide’s log confirms that its Rust extension and built-in plugins load without a recorded error.
- The checked kernel log contained no OOM-kill entry. Approximately 53 GiB RAM was available after a failure.
- An earlier Rust startup failure produced a kernel segfault entry in lib_burst_generated.so in a Job.Worker thread. I cannot confirm that this entry corresponds to the fresh-instance failure.
A separate RCON issue occurs after server updates: localhost connection attempts report an incorrect password and trigger repeated 300-second bans. During one affected vanilla run, the “status” command still returned server information while failed connection attempts continued approximately every six seconds.
Restarting the AMP management process for this instance, followed by starting Rust, cleared the password errors in repeated vanilla tests. The “status” command then worked without further password errors in the observed output.
Importantly, the uMod crash log records “Web RCON connected” and “RCON connection successful” before exit code 139, so the uMod crash also occurs with a successful RCON connection.
Currently the instance is running successfully as vanilla with uMod disabled. No custom RCON password is intentionally configured. No rcon.password entry was found in serverauto.cfg, and no server.cfg was present in the checked configuration directories.
Core dumps are disabled (soft limit 0), coredumpctl is not installed, and I do not want to enable large core dumps.
Please advise how to diagnose and fix the uMod startup crash using AMP’s managed installation workflow, and whether the recurring RCON password issue is a known AMP defect.
Reproduction Steps
- Create a new Rust Dedicated Server instance in AMP with uMod disabled. Install the server, enable PvE, and start it. Vanilla starts successfully and players can connect.
- Stop the Rust server, enable “Install UMod” in Configuration, and run Update. Wait for the update to complete successfully.
- Start the server with uMod enabled and no third-party plugins installed. The console output stops after “Ocean Patrol Paths”, and AMP reports that the application stopped unexpectedly with exit code 139.