This av-service av-rtmp-rtsp-light-sink for alpine is a container running nginx RTMP server which convert an incoming RTMP stream into RTSP stream. This av-service is used in IoT Edge device with Azure Video Analyzer to convert an incoming RTMP stream into a RTSP stream to feed Azure Video Analyzer AI components. When you will deploy this component with avtool.sh, it will deploy a complete AVA infrastructure with IoT Edge Hub, Azure Container Registry, Azure Media Services (so far mandatory to deploy Azure Video Analyzer), Azure Storage, Azure Virtual Machine acting as IoT Edge device and running docker. When you will start, stop this component with avtool.sh, it will start, stop the rtmpsource container in the IoT Edge device. When you will test this component with avtool.sh, it will test automatically the following scenarios:
- RTMP to RTSP adaptor
- AVA Motion Detection
At least, when the rtmpsource will be fed with a Live RTMP stream, you could consume the following streams with VLC:
RTSP URL: rtsp://<IoTEdgeVMName>.<REGION>.cloudapp.azure.com:8554/rtsp/stream
SSH command: ssh <VMAdmin>@<IoTEdgeVMName>.<REGION>.cloudapp.azure.com
It's recommended to use and manage the av-rtmp-rtsp-light-sink alpine service with the avtool.sh command line tool.
As avtool.sh is a Linux bash file, you could run this tool from a machine or virtual machine running Ubuntu 20.04 LTS.
- Ensure git is installed running the following command
sudo apt-get install git- Clone the av-services repository on your machine
mkdir $HOME/git
cd $HOME/git
git clone https://github.com/flecoqui/av-services.git
cd av-services/envs/container/docker/av-rtmp-rtsp-light-sink/alpine - Run avtool.sh -a install to install the pre-requisite ffmpeg, Azure Client, ...
./avtool.sh -a installOnce the pre-requisites are installed, you can deploy the Live Analytics infrastructure (IoT Edge Hub, Azure Container Registry, Azure Media Services, Azure Storage, Azure Virtual Machine acting as IoT Edge device and running docker) and build the av-rtmp-rtsp-light-sink alpine container.
- Run the following command to build and run the container
./avtool.sh -a deployWhen you run avtool.sh for the first time, it creates a file called .avtoolconfig to store the av-rtmp-rtsp-light-sink in AVA configuration. By default, the file contains these parameters:
AV_RESOURCE_GROUP=av-rtmp-rtsp-ava-rg
AV_RESOURCE_REGION=eastus2
AV_IMAGE_NAME=av-rtmp-rtsp-light-sink-alpine
AV_IMAGE_FOLDER=av-services
AV_CONTAINER_NAME=av-rtmp-rtsp-light-sink-alpine-container
AV_EDGE_DEVICE=rtmp-rtsp-ava-device
AV_PORT_RTMP=1935
AV_PREFIXNAME=rtmprtspava
AV_VMNAME=rtmprtspavavm
AV_CONTAINERNAME=avchunks
AV_LOGIN=avvmadmin
AV_PASSWORD={YourPassword}
AV_HOSTNAME=rtmprtspavavm.eastus2.cloudapp.azure.com
AV_PORT_RTMP=1935
AV_PORT_RTSP=8554
AV_TEMPDIR=
AV_STORAGENAME=
AV_SASTOKEN=
AV_IOTHUB=
AV_IOTHUB_CONNECTION_STRING=
AV_DEVICE_CONNECTION_STRING=
AV_CONTAINER_REGISTRY=
AV_CONTAINER_REGISTRY_DNS_NAME=
AV_CONTAINER_REGISTRY_USERNAME=
AV_CONTAINER_REGISTRY_PASSWORD=
AV_SUBSCRIPTION_ID=Below further information about the input parameters in the file .avtoolconfig: It's important before running ./avtool.sh -a deploy to define the following input parameters:
- AV_PREFIXNAME: The name prefix used for all the Azure resources. As the Azure Container Registry name and the virtual machine name depends on this prefix, if a deployment as already used the same prefix, your deployment will fail because of a name conflict in Azure (by default rtmprtspava)
- AV_PASSWORD: The password for the Virtual Machine running IoT Edge.
Below the list of input parameters:
| Variables | Description |
|---|---|
| AV_RESOURCE_GROUP | The name of the resource group where AVA infrastructure will be deployed (av-rtmp-rtsp-ava-rg by default) |
| AV_RESOURCE_REGION | The Azure region where AVA infrastructure will be deployed (eastus2 by default) |
| AV_SERVICE | The name of the service (by default av-rtmp-rtsp-light-sink) |
| AV_FLAVOR | The flavor of this service (by default alpine) |
| AV_IMAGE_NAME | The suffix of the image name (by default ${AV_SERVICE}-${AV_FLAVOR}) |
| AV_IMAGE_FOLDER | The image folder, the image name will be ${AV_IMAGE_FOLDER}/${AV_IMAGE_NAME} |
| AV_CONTAINER_NAME | The name of the container (by default av-rtmp-rtsp-light-sink-alpine-container) |
| AV_EDGE_DEVICE | The name of the Edge device (by default rtmp-rtsp-ava-device) |
| AV_PREFIXNAME | The name prefix used for all the Azure resources (by default rtmprtspava) |
| AV_VMNAME | The name of the virtual machien running IoT Edge (by default "$AV_PREFIXNAME"vm) |
| AV_HOSTNAME | The host name of the container. Default value: "$AV_VMNAME"."$AV_RESOURCE_REGION".cloudapp.azure.com |
| AV_CONTAINERNAME | The name of the container in Azure Storage Account where the video and audio chunks will be stored (by default avchunks) |
| AV_LOGIN | The login for the Virtual Machine running IoT Edge. Default value: avvmadmin |
| AV_PASSWORD | The password for the Virtual Machine running IoT Edge. Default value: |
| AV_HOSTNAME | The host name of the container. Default value: localhost |
| AV_PORT_RTMP | The RTMP port. Default value: 1935 |
| AV_PORT_RTSP | The RTSP port. Default value: 8554 |
| AV_TEMPDIR | The directory on the host machine used to store MKV and MP4 files for the tests |
When the service is running and fed with a RTMP stream, the following urls could be used for the tests:
RTMP URL: rtmp://<IoTEdgeVMName>.<REGION>.cloudapp.azure.com:1935/live/stream
RTSP URL: rtsp://<IoTEdgeVMName>.<REGION>.cloudapp.azure.com:8554/rtsp/stream
SSH command: ssh <VMAdmin>@<IoTEdgeVMName>.<REGION>.cloudapp.azure.com
Below the output parameters:
| Variables | Description |
|---|---|
| AV_RESOURCE_GROUP | The name of the resource group where AVA infrastructure will be deployed (av-rtmp-rtsp-ava-rg by default) |
| AV_STORAGENAME | The name of the sotrage account created |
| AV_SASTOKEN | The Shared Access Signature for the storage account |
| AV_IOTHUB | The IoT Hub name |
| AV_IOTHUB_CONNECTION_STRING | The IoT Hub connection string |
| AV_DEVICE_CONNECTION_STRING | The IoT Device connection string |
| AV_CONTAINER_REGISTRY | The Azure Container Registry name |
| AV_CONTAINER_REGISTRY_DNS_NAME | The Azure Container Registry login server name |
| AV_CONTAINER_REGISTRY_USERNAME | The Azure Container Registry user name |
| AV_CONTAINER_REGISTRY_PASSWORD | The Azure Container Registry password |
| AV_SUBSCRIPTION_ID | The Azure Subscription ID |
Once the rtmpsource service is built and deployed you can start and stop the container .
- Run the following command to start the container in IoT Edge Device
./avtool.sh -a start- If the container is still running, you can run the following command to stop the container in IoT Edge Device
./avtool.sh -a stop- If the container is still running, you can run the following command to get the status of the container in IoT Edge Device
./avtool.sh -a statusOnce the image is built you can test if the container is fully functionning.
- Run the following command to test the container
./avtool.sh -a testFor this container, the test feature will check if the output MP4 files have been created in the temporary folder from output RTSP url.
By default for the tests, it will test automatically the following scenarios:
- RTMP to RTSP adaptor
- AVA Motion Detection
For this test, we use an incoming Live RTMP stream created from a MKV file using the following ffmpeg command:
ffmpeg -hide_banner -loglevel error -re -stream_loop -1 -i "${AV_TEMPDIR}"/camera-300s.mkv -codec copy -bsf:v h264_mp4toannexb -f flv rtmp://${AV_HOSTNAME}:${AV_PORT_RTMP}/live/streamIf on the host machine a Webcam is installed, you can use this webcam to generate the incoming RTMP stream using the following ffmpeg command:
ffmpeg.exe -v verbose -f dshow -i video="Integrated Webcam":audio="Microphone (Realtek(R) Audio)" -video_size 1280x720 -strict -2 -c:a aac -b:a 192k -ar 44100 -r 30 -g 60 -keyint_min 60 -b:v 2000000 -c:v libx264 -preset veryfast -profile main -level 3.0 -pix_fmt yuv420p -bufsize 1800k -maxrate 400k -f flv rtmp://localhost:1935/live/stream