[Bug] Memory leak and OOM crash when running in headless Docker (Webtop) after ~20 hours #70

Closed
opened 2026-04-05 16:15:55 +02:00 by MrUnknownDE · 0 comments
Owner

Originally created by @potaxo on 2/14/2026

Describe the bug
When running VRCX inside a headless Docker container (specifically using the linuxserver/webtop Ubuntu XFCE image), the application suffers from a slow memory leak. Even when sitting completely idle in the background tracking friend logs, the RAM usage steadily climbs. After roughly 20 hours of continuous uptime, the container hits a 2GB memory limit and is forcefully terminated by the Docker engine ("OOMKilled": true).
(Note: Without the --disable-gpu flag, it also suffers from an aggressive crash loop where it endlessly tries to respawn GPU helpers, spiking CPU load to 25+).

To Reproduce
Steps to reproduce the behavior:

  1. Deploy VRCX (via .AppImage extraction) inside a linuxserver/webtop:ubuntu-xfce Docker container.
  2. Launch VRCX using the command: ./vrcx --no-sandbox --disable-gpu --disable-software-rasterizer --disable-dev-shm-usage
  3. Set a Docker container memory limit to 2GB (mem_limit: 2048m).
  4. Leave VRCX running continuously in the background.
  5. Wait approximately 20 hours. Docker will OOM-kill the container once the memory leak hits the 2GB ceiling.

Expected behavior
The application should manage its memory footprint and garbage collection properly so it can run 24/7 as a background tracking tool without eventually exhausting system RAM and triggering an OOM kill.

Screenshots
Image

Image Image

What version you are running
Version: VRCX 2026.01.28 (Linux AppImage)
Environment: OpenWrt (Intel J1900 CPU), Docker Host Mode, linuxserver/webtop:ubuntu-xfce image.

*Originally created by @potaxo on 2/14/2026* **Describe the bug** When running VRCX inside a headless Docker container (specifically using the linuxserver/webtop Ubuntu XFCE image), the application suffers from a slow memory leak. Even when sitting completely idle in the background tracking friend logs, the RAM usage steadily climbs. After roughly 20 hours of continuous uptime, the container hits a 2GB memory limit and is forcefully terminated by the Docker engine ("OOMKilled": true). (Note: Without the --disable-gpu flag, it also suffers from an aggressive crash loop where it endlessly tries to respawn GPU helpers, spiking CPU load to 25+). **To Reproduce** Steps to reproduce the behavior: 1. Deploy VRCX (via .AppImage extraction) inside a linuxserver/webtop:ubuntu-xfce Docker container. 2. Launch VRCX using the command: ./vrcx --no-sandbox --disable-gpu --disable-software-rasterizer --disable-dev-shm-usage 3. Set a Docker container memory limit to 2GB (mem_limit: 2048m). 4. Leave VRCX running continuously in the background. 5. Wait approximately 20 hours. Docker will OOM-kill the container once the memory leak hits the 2GB ceiling. **Expected behavior** The application should manage its memory footprint and garbage collection properly so it can run 24/7 as a background tracking tool without eventually exhausting system RAM and triggering an OOM kill. **Screenshots** <img width="2317" height="1544" alt="Image" src="https://github.com/user-attachments/assets/3fe67f51-de71-42d8-9252-31ccf055f28c" /> <img width="1971" height="1001" alt="Image" src="https://github.com/user-attachments/assets/2eac64ab-4982-42dd-90b8-1f60c248a745" /> <img width="1102" height="546" alt="Image" src="https://github.com/user-attachments/assets/d000635b-1195-4580-bb27-e389c8c56123" /> **What version you are running** Version: VRCX 2026.01.28 (Linux AppImage) Environment: OpenWrt (Intel J1900 CPU), Docker Host Mode, linuxserver/webtop:ubuntu-xfce image.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github/VRCX#70