mirror of
https://github.com/games-on-whales/wolf.git
synced 2026-10-15 13:02:03 +00:00
[GH-ISSUE #246] Need help with custom config and startup options #170
Labels
No labels
🧠 1 - 🤖 999
AMD
bug
bug
cannot reproduce
documentation
duplicate
easy
enhancement
enhancement
enhancement
fixed?
fixed?
good first issue
hard
help wanted
Intel
invalid
LXC
medium
moonlight-protocol
Multi GPU
nixos
not a bug
nvidia
pull-request
question
wontfix
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
games-on-whales/wolf#170
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @zeedif on GitHub (Aug 23, 2025).
Original GitHub issue: https://github.com/games-on-whales/wolf/issues/246
Title: Wolf fails to start on Intel iGPU (Alder Lake-N) with encoder detection and
std::bad_allocerrorsI'm trying to set up Wolf on my Debian server which has an integrated Intel N97 GPU (UHD Graphics, Alder Lake-N). The Wolf container gets stuck in a crash loop and restarts continuously.
Depending on the GStreamer configuration in
config.toml, I encounter two different fatal errors.System Details
vainfooutput:Steps to Reproduce
docker-compose.ymlas provided below.config.tomlwith a simple RetroArch application added.docker compose up -d.docker logs wolf.Observed Behavior
The container fails to start and enters a restart loop. The error depends on the configuration:
Scenario 1: Default
config.tomlWith the default GStreamer encoder order (which prioritizes
nvcodec), the application fails immediately.Relevant Log Snippet:
It seems Wolf is unable to find the NVIDIA encoders (which is expected on an Intel system) and fails before it can try the Intel-compatible ones like
qsvorva.Scenario 2: Modified
config.tomlto prioritize Intel encodersAs a troubleshooting step, I modified the
config.tomlto move theqsvandvaencoder definitions to the top of the[[gstreamer.video.h264_encoders]]list. After this change, the error is different.Relevant Log Snippet:
In this case, Wolf correctly selects the
qsvencoder but then immediately crashes with astd::bad_allocerror. I'm not sure why it's falling back to the "legacy pipeline" or what could be causing the memory allocation error, as the system has plenty of free RAM. It seems to be an issue in how the GStreamer pipeline is constructed.Expected Behavior
Wolf should successfully initialize on an Intel-based system, select the appropriate
qsvorvaencoder, start all its services (HTTP, RTSP, etc.), and remain running to accept Moonlight connections.Configuration Files
docker-compose.ymlconfig.toml(usado en el Escenario 1)@ABeltramo commented on GitHub (Aug 23, 2025):
Sounds like you have a different issue, Wolf definitely runs on Intel iGPUs.
Could you please share the full Wolf logs without edits/cuts and more info about the HW? Have you got any other GPU plugged on your system?
@ABeltramo commented on GitHub (Aug 23, 2025):
Also, make sure that you are running the latest image with:
@zeedif commented on GitHub (Aug 23, 2025):
I shared the
docker-compose.ymlI'm using, where you can see that I'm pullingghcr.io/games-on-whales/wolf:stable.These are the full logs when I don't customize the gstreamer configuration:
When I add this to the end of my
config.toml:I get the following logs:
@ABeltramo commented on GitHub (Aug 23, 2025):
The first
config.tomlthat you've posted isn't valid, I'm not sure why you are manually writing it instead of using the default, but that's what's causing all your exceptions.Delete your
config.tomlso that Wolf creates a valid default file or if you really want to do that yourself just copy and paste the valid default file from https://github.com/games-on-whales/wolf/blob/stable/src/moonlight-server/state/default/config.v5.toml@zeedif commented on GitHub (Aug 23, 2025):
Thanks, I hadn't found that file. After adjusting my config file like this:
config.toml
It finally starts up! The logs now show that it correctly identifies my Intel iGPU and selects the
qsvencoder. The only problem now seems to be with PulseAudio, as it fails to connect even after trying to start its own container:Logs WolfPulseAudio:
@ABeltramo commented on GitHub (Aug 23, 2025):
You are missing
XDG_RUNTIME_DIRand the relative mount.Please just run the defaults from the quickstart guide, you are making a bit of a mess by manually editing stuff without the right knowledge 😅
@zeedif commented on GitHub (Aug 23, 2025):
Thanks, it seems to be running fine now. Clarifying the documentation would help a lot.
@ABeltramo commented on GitHub (Aug 23, 2025):
Pull requests are always welcome 😉