GMS (Guac Media Server)

GMS is a simple, high performance and secure live media server written in pure Rust. It supports popular live protocols like RTMP/HLS/LL-HLS/HTTP-FLV. You can deploy it as a stand-alone server or a cluster using the relay feature.

Features

• Support multiple platforms(Linux/MacOS/Windows).
• Support RTMP as a stand-alone server or cluster(RTMP relay).
  • Support GOP cache which can be configured in the configuration file.
• Support HTTP-FLV/HLS protocols(Transferred from RTMP).
• Support configuring the service using command line or a configuration file.
• Support HTTP API/Notifications.
  • Support querying stream information.
  • Support notify stream status.

Preparation

Install Rust and Cargo

Document

Install and run

There are two ways to install gms:

• Using cargo to install
• Building from source

Install using cargo

Issue the following command to install gms:

cargo install gms

Start the service with the following command to get help:

gms -h
A secure and easy to use live media server

Usage: gms [OPTIONS] <--config <path>|--rtmp <port>>

Options:
  -c, --config <path>   Specify the gms server configuration file path.
  -r, --rtmp <port>     Specify the RTMP listening port(e.g.:1935).
  -f, --httpflv <port>  Specify the HTTP-FLV listening port(e.g.:8080).
  -s, --hls <port>      Specify the HLS listening port(e.g.:8081).
  -l, --log <level>     Specify the log level. [possible values: trace, debug, info, warn, error, debug]
  -h, --help            Print help
  -V, --version         Print version

Build from source

Clone gms

git clone https://github.com/guaclive/gms.git

Build

cd ./gms/application/gms
cargo build --release

Run

cd ./gms/target/release
./gms -h

CLI

Instructions

You can use command line to configure the gms server easily. You can specify to configure gms using configuration file or from the command lines.

Configure using file

gms -c configuration_file_path

Configure using command line

gms -r 1935 -f 8080 -s 8081 -l info

How to Configure the configuration file

RTMP

[rtmp]
enabled = true
port = 1935

# send webhook on publish and publish done
[rtmp.webhooks]
enabled = true
publish_url = "http://localhost:8080/api/v1/publish"
publish_done_url = "http://localhost:8080/api/v1/publish_done"

# pull streams from other server node.
[rtmp.pull]
enabled = false
address = "192.168.0.1"
port = 1935

# push streams to other server node.
[[rtmp.push]]
enabled = true
address = "localhost"
port = 1936
[[rtmp.push]]
enabled = true
address = "192.168.0.3"
port = 1935

HTTPFLV

[httpflv]
# true or false to enable or disable the feature
enabled = true
# listening port
port = 8081

HLS

[hls]
# true or false to enable or disable the feature
enabled = true
# listening port
port = 8080

Log

[log]
level = "info"
[log.file]
# write log to file or not（Writing logs to file or console cannot be satisfied at the same time）.
enabled = true
# set the rotate
rotate = "hour" #[day,hour,minute]
# set the path where the logs are saved
path = "./logs"

Configuration examples

I edit some configuration files under the following path which can be used directly:

gms/application/gms/src/config

It contains the following 4 files:

config_rtmp.toml //enable rtmp only
config_rtmp_hls.toml //enable rtmp and hls
config_rtmp_httpflv.toml //enable rtmp and httpflv
config_rtmp_httpflv_hls.toml //enable all the 3 protocols

Scenarios

Push

You can use two ways:

• Use OBS to push a live rtmp stream
• Or use FFmpeg to push a rtmp stream: ffmpeg -re -stream_loop -1 -i test.mp4 -c:a copy -c:v copy -f flv -flvflags no_duration_filesize rtmp://127.0.0.1:1935/live/test

Play

Use ffplay to play the rtmp/httpflv/hls live stream:

ffplay -i rtmp://localhost:1935/live/test
ffplay -i http://localhost:8081/live/test.flv
ffplay -i http://localhost:8080/live/test/test.m3u8

Relay - Static push

The configuration files are as follows:

The configuration file of Service 1 named config.toml:

[rtmp]
enabled = true
port = 1935
[[rtmp.push]]
enabled = true
address = "localhost"
port = 1936

The configuration file of Service 2 named config_push.toml:

[rtmp]
enabled = true
port = 1936

Run the 2 services:

./gms config.toml
./gms config_push.toml

Use the above methods to push rtmp live stream to service 1, then the stream can be pushed to service 2 automatically, you can play the same live stream from both the two services:

ffplay -i rtmp://localhost:1935/live/test
ffplay -i rtmp://localhost:1936/live/test

Relay - Static pull

The configuration file are as follows:

The configuration file of Service 1 named config.toml:

[rtmp]
enabled = true
port = 1935

The configuration file of Service 2 named config_pull.toml:

[rtmp]
enabled = true
port = 1936
[rtmp.pull]
enabled = false
address = "localhost"
port = 1935

Run the 2 services:

./gms config.toml
./gms config_pull.toml

Use the above methods to push live stream to service 1, when you play the stream from serivce 2, it will pull the stream from service 1:

ffplay -i rtmp://localhost:1935/live/test
ffplay -i rtmp://localhost:1936/live/test

Thanks

• media_server
• xiu - Codebase this was forked from.
• Phineas - LLHLS implementation.

Others

Open issues if you have any problems. Star and pull requests are welcomed. Your stars can make this project go faster and further.