Install Intel QSV GPU runtimes in amd64 container

This commit is contained in:
Technomancer702
2026-10-13 19:04:46 -07:00
parent 379971022d
commit 117ae86ee0
3 changed files with 3 additions and 2 deletions

View File

@ -26,6 +26,7 @@ jobs:
if docker exec nodecast-smoke node scripts/healthcheck.cjs; then
docker exec nodecast-smoke sh -c 'test ! -e /app/.env.container-sentinel && test ! -e /app/data/container-sentinel'
docker exec nodecast-smoke lspci --version
docker exec nodecast-smoke sh -c 'ldconfig -p | grep -q libmfxhw64.so.1 && ldconfig -p | grep -q libmfx-gen.so.1.2'
docker exec nodecast-smoke node -e "require('./server/services/hwDetect').detectQuickSync('ffmpeg').then(q => { if (q.available) process.exit(1); })"
exit 0
fi

View File

@ -19,7 +19,7 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
gnupg \
&& curl -fsSL https://deb.nodesource.com/setup_20.x | bash - \
&& if [ "$TARGETARCH" = "amd64" ]; then \
DRIVERS="mesa-va-drivers intel-media-va-driver vainfo"; \
DRIVERS="mesa-va-drivers intel-media-va-driver vainfo libmfx1 libmfx-gen1.2"; \
else \
DRIVERS=""; \
fi \

View File

@ -120,7 +120,7 @@ Ensure you have the [NVIDIA Container Toolkit](https://docs.nvidia.com/datacente
```
**Verify:**
On Linux, QSV detection checks an accessible Intel render device and verifies a short H.264 encode with the configured FFmpeg. If that check fails, QSV is not automatically recommended. Ensure the container has read/write permission on the mapped device. The image includes `pciutils` for systems where GPU vendor information is hidden in sysfs.
On Linux, QSV detection checks an accessible Intel render device and verifies a short H.264 encode with the configured FFmpeg. If that check fails, QSV is not automatically recommended. Ensure the container has read/write permission on the mapped device. The amd64 image includes the Intel Media SDK and VPL GPU runtimes for older and newer Intel GPUs. It also includes `pciutils` for systems where GPU vendor information is hidden in sysfs.
After restarting the container, go to **Settings -> Transcoding**. The **Hardware Detection** status should list your GPU (e.g., "NVIDIA GPU Detected" or "VAAPI Available").