docker-nginx-rtmp fork
A Dockerfile installing NGINX, nginx-rtmp-module and FFmpeg from source with default settings for HLS live streaming. Built on Alpine Linux (Ubuntu for cuda).
- Nginx 1.23.2 (Mainline version compiled from source)
- nginx-rtmp-module 1.2.2 (compiled from source)
- ffmpeg 6.1 (compiled from source, support HEVC,VP9,AV1 codec in enhanced flv format & rtmp protocol)
- Default HLS settings (See: nginx.conf)
Usage
Server
- Build and run container from source:
docker build -t nginx-rtmp .
docker run -it -p 1935:1935 -p 8080:80 --rm nginx-rtmp
- Stream live content to:
rtmp://localhost:1935/stream/$STREAM_NAME
SSL
To enable SSL, see nginx.conf and uncomment the lines:
listen 443 ssl;
ssl_certificate /opt/certs/example.com.crt;
ssl_certificate_key /opt/certs/example.com.key;
This will enable HTTPS using a self-signed certificate supplied in /certs. If you wish to use HTTPS, it is highly recommended to obtain your own certificates and update the ssl_certificate
and ssl_certificate_key
paths.
I recommend using Certbot from Let's Encrypt.
nginx.conf
Custom If you wish to use your own nginx.conf
, mount it as a volume in your docker-compose
or docker
command as nginx.conf.template
:
volumes:
- ./nginx.conf:/etc/nginx/nginx.conf
OBS Configuration
- Stream Type:
Custom Streaming Server
- URL:
rtmp://localhost:1935/stream
- Stream Key:
hello
Watch Stream
- Load up the example hls.js player in your browser:
http://localhost:8080/player.html?url=http://localhost:8080/live/hello.m3u8
- Or in Safari, VLC or any HLS player, open:
http://localhost:8080/live/$STREAM_NAME.m3u8
- Example Playlist:
http://localhost:8080/live/hello.m3u8
- HLS.js Player
- FFplay:
ffplay -fflags nobuffer rtmp://localhost:1935/stream/hello
Restream to multiple destination
H264
example nginx.conf
rtmp {
access_log /var/log/nginx/access.log;
server {
listen 1935; # Listen on standard RTMP port
chunk_size 4000;
# This application is to accept incoming stream
application push_live_60fps_m {
allow publish 127.0.0.1;
allow publish 192.168.1.0/24;
deny publish all;
meta copy;
live on; # Allows live input
drop_idle_publisher 15s;
exec_push /usr/local/bin/ffmpeg -i rtmp://localhost:1935/$app/$name
-c:v libx264 -preset medium -b:v 6000k -maxrate 6000k -bufsize 12000k
-r 60 -profile:v high -level 5.1
-x264opts nal-hrd=cbr:bframes=2:keyint=120:no-scenecut
-c:a aac -b:a 192k
-f flv rtmp://localhost:1935/restream/$name;
}
application restream {
allow publish 127.0.0.1;
allow publish 192.168.1.0/24;
deny publish all;
live on;
meta copy;
record off;
push_reconnect 500ms;
push rtmp://fra05.contribute.live-video.net/app/my_stream_key;
push rtmp://a.rtmp.youtube.com/live2/my_stream_key;
push rtmp://b.rtmp.youtube.com/live2?backup=1/my_stream_key;
}
}
}
h265
example nginx.conf
rtmp {
access_log /var/log/nginx/access.log;
server {
listen 1935; # Listen on standard RTMP port
chunk_size 4000;
# This application is to accept incoming stream
application push_live_60fps_h265 {
allow publish 127.0.0.1;
allow publish 192.168.1.0/24;
deny publish all;
meta copy;
live on; # Allows live input
drop_idle_publisher 15s;
exec_push /usr/local/bin/ffmpeg -i rtmp://localhost:1935/$app/$name
-c:v libx265 -preset fast
-r 60 -profile:v main -level 5.1
-x265-params bitrate=4500:vbv-maxrate=4500:vbv-bufsize=9000:strict-cbr:bframes=2:keyint=120:scenecut=0
-c:a aac -b:a 192k
-f flv rtmp://localhost:1936/restream/$name;
exec_push /usr/local/bin/ffmpeg -listen 1 -i rtmp://localhost:1936/restream/$name
-c copy -f flv rtmp://a.rtmp.youtube.com/live2/my_stream_key
-c copy -f flv rtmp://a.rtmp.youtube.com/live2/my_stream_key2;
}
}
}
or you can use tee muxer, but i didn't tested it
exec_push /usr/local/bin/ffmpeg -i rtmp://localhost:1935/$app/$name
-c:v libx265 -preset fast
-r 60 -profile:v main -level 5.1
-x265-params bitrate=4500:vbv-maxrate=4500:vbv-bufsize=9000:strict-cbr:bframes=2:keyint=120:scenecut=0
-c:a aac -b:a 192k
-map 0 -f tee '[f=flv:onfail=ignore]rtmp://a.rtmp.youtube.com/live2/my_stream_key|[f=flv:onfail=ignore]rtmp://a.rtmp.youtube.com/live2/my_stream_key2;
FFmpeg Build
$ ffmpeg -buildconf
ffmpeg version 6.1 Copyright (c) 2000-2023 the FFmpeg developers
built with gcc 12.2.1 (Alpine 12.2.1_git20220924-r10) 20220924
configuration: --prefix=/usr/local --enable-version3 --enable-gpl --enable-nonfree --enable-libmp3lame --enable-libx264 --enable-libx265 --enable-libsvtav1 --enable-libaom --enable-libdav1d --enable-libvpx --enable-libtheora --enable-libvorbis --enable-libopus --enable-libfdk-aac --enable-libass --enable-libwebp --enable-postproc --enable-libfreetype --enable-openssl --disable-debug --disable-doc --disable-ffplay --extra-libs='-lpthread -lm'
libavutil 58. 29.100 / 58. 29.100
libavcodec 60. 31.102 / 60. 31.102
libavformat 60. 16.100 / 60. 16.100
libavdevice 60. 3.100 / 60. 3.100
libavfilter 9. 12.100 / 9. 12.100
libswscale 7. 5.100 / 7. 5.100
libswresample 4. 12.100 / 4. 12.100
libpostproc 57. 3.100 / 57. 3.100
configuration:
--prefix=/usr/local
--enable-version3
--enable-gpl
--enable-nonfree
--enable-libmp3lame
--enable-libx264
--enable-libx265
--enable-libsvtav1
--enable-libaom
--enable-libdav1d
--enable-libvpx
--enable-libtheora
--enable-libvorbis
--enable-libopus
--enable-libfdk-aac
--enable-libass
--enable-libwebp
--enable-postproc
--enable-libfreetype
--enable-openssl
--disable-debug
--disable-doc
--disable-ffplay
--extra-libs='-lpthread -lm'
FFmpeg Hardware Acceleration
A Dockerfile.cuda
image is available to enable FFmpeg hardware acceleration via the NVIDIA's CUDA.
docker build -t nginx-ffmpeg-cuda -f .\Dockerfile.cuda .
docker run -it -p 1935:1935 -p 8080:80 --rm nginx-ffmpeg-cuda
change codec in nginx.conf
for example: libx264
to h264_nvenc
You must have a supported platform and driver to run this image.
- https://github.com/NVIDIA/nvidia-docker
- https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker
- https://docs.docker.com/docker-for-windows/wsl/
- https://trac.ffmpeg.org/wiki/HWAccelIntro#CUDANVENCNVDEC