NVIDIA provides two Azure MQTT protocol libraries installed with DeepStream under /opt/nvidia/deepstream/deepstream/lib
libnvds_azure_proto.so- a device client protocol for sending messages from the device to an Azure IoT Hub instance. Requires minimal setup.libnvds_azure_edge_proto.so- a module client protocol for bidirectional device-server messaging. Applications linking with DSL must be run in a deployed Docker container.
The protocol adapter libraries are used by the DSL Message Sink and Message Broker components.
sudo apt-get install -y libcurl4-openssl-dev libssl-dev uuid-dev libglib2.0 libglib2.0-dev
Follow the directions at https://docs.microsoft.com/en-us/azure/iot-hub/tutorial-connectivity#create-an-iot-hub.
There are three methods to register your device as outlined at https://docs.microsoft.com/en-us/azure/iot-edge/how-to-provision-single-device-linux-symmetric?view=iotedge-2020-11&tabs=azure-portal%2Cubuntu#register-your-device - Azure portal, Visual Studio Code, and Azure CLI - with Azure portal being the simplest.
The following section details the installation process and steps to use the Azure CLI.
The Azure CLI team maintains a script to run all installation commands in one step. This script is downloaded via curl and piped directly to bash to install the CLI.
curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bashCurrently, the only way to use Azure CLI on ARM64 is to install from PyPI (https://pypi.org/project/azure-cli/):
sudo pip3 install azure-cli
If the install fails, see Failure installing azure-cli on Jetson under Trouble Shooting.
Important: once the installation is complete you will need to reboot the device!
sudo rebootVerify the installation with.
az --versionIf the verification command fails, see Failure verifying azure-cli install on Jetson under Trouble Shooting.
az extension add --name azure-iotLog into to your Azure IoT Hub on the cloud from your device terminal window. The following command will bring up the login screen in your browser.
az loginCreate an IoT edge device from your device terminal with the following command.
az iot hub device-identity create --device-id <device-id> --hub-name <hub-name> --edge-enabledwhere:
<device-id>= name (string) to identify the new device<hub-name>= the hub-name you used when you Setup an Azure IoT Hub Instance above.
Verify the device creation with the following command.
az iot hub device-identity list --hub-name <hub-name>Or check from your Azure IoT Hub instance on the cloud. From your hub dashboard, select the IoT Edge item in the left menu pane, you should then see your device listed in the main window.
Get the connection string for your new device
az iot hub device-identity connection-string show --device-id <device-id> --hub-name <hub-name>Or copy the Primary Connection String from your Azure IoT Hub instance by selecting the device name on IoT Edge main page (see image above).
You will need the connection-string to use the Message Sink and Message Broker API.
Run the settup script with the following commands.
sudo chmod u+x /opt/nvidia/deepstream/deepstream/sources/tools/nvds_logger/setup_nvds_logger.sh
sudo /opt/nvidia/deepstream/deepstream/sources/tools/nvds_logger/setup_nvds_logger.shLog messages will be written to /tmp/nvds/ds.log.
Note: when using the Message Sink, setup errors will result in the Pipeline failing to play. The reason for the failure may be found in the nvds log file, for example
Mar 24 16:37:57 prominenceai1-desktop dsl-test-app.exe: DSLOG:NVDS_AZURE_PROTO: Error. Azure connection string not provided#012
Mar 24 16:37:57 prominenceai1-desktop dsl-test-app.exe: DSLOG:NVDS_AZURE_PROTO: nvds_msgapi_connect: Failure in fetching azure connection string
Your device setup is now sufficient to use the Device Client libnvds_azure_proto.so with the following examples.
Follow the below steps from the instructions here. https://docs.microsoft.com/en-us/azure/iot-edge/how-to-install-iot-edge-linux. Specifically:
- Install IoT Edge.
- Caution: make sure to skip the section Install a container engine.
- Install the IoT Edge runtime.
- Provision the device with its cloud identity
- Verify successful configuration
See the instructions and Docker file under the deepstream-services-library-docker GitHub repository.
As a privileged user (root), append the following lines to file /etc/profile to allow access to the local X-server.
if [ "$DISPLAY" != "" ]
then
xhost +
fiMake the file executable
sudo chmod u+x /etc/profileThen execute the file.
sudo /etc/profileYou should see confirmation similar to below.
access control disabled, clients can connect from any host.
Note: how to correctly setup specific or limited access is still under investigation.
The following instructions detail the steps to create a new IoT Edge Custom Module to run the Docker Image created in the previous section Build and deploy a Docker Image.
From your Azure portal, select IoT Edge from the left menu pane, then select your device by its id. You should see the two existing IoT Edge System Modules that were deployed when you Setup the Azure IoT Edge runtime on the edge device.
To setup a new IoT Edge Custom Module to run the message_broker_azure_module_client.py example, select the Set modules from the upper menu bar on the device page, as show below.
Select the + Add button under the IoT Edge Modules section on the Set modules page and select the first item + IoT Edge Module as shown below.
Define the Module Settings by specifying the IoT Edge Module Name and Image URI to the Docker Image created in the previous section Build and deploy a Docker Image as shown in the image below.
Select the Environment Variables tab and add the DISPLAY variable as shown below. Set the value to :0 for local or :1 for remote.
Select the Container Create Options tab
Then add the below JSON code and select Add
{
"Entrypoint": [
"python3",
"/opt/prominenceai/deepstream-services-library/examples/python/message_broker_azure_module_client.py"
],
"HostConfig": {
"runtime": "nvidia",
"NetworkMode": "host",
"Binds": [
"/tmp/argus_socket:/tmp/argus_socket",
"/tmp/.X11-unix/:/tmp/.X11-unix/",
"/tmp/.dsl/:/tmp/.dsl/"
],
"IpcMode": "host"
},
"NetworkingConfig": {
"EndpointsConfig": {
"host": {}
}
},
"WorkingDir": "/opt/prominenceai/deepstream-services-library/examples/python/"
}Once the code has been added, select the Review + create button from the Set modules main page as show below:
Make sure that the validation passed - as show in the upper left corner in the image below - and then select Create
All three IoT Edge Modules should now be in a running state as shown below.
The example script directs the DSL INFO logs and Python application logs to files located under /tmp/.dsl/ which is viewable from outside of the Module container.
Use the trouble shooting page in the Azure portal to monitor IoT Edge devices and modules.
Create a single database in Azure SQL Database using either the Azure portal, a PowerShell script, or an Azure CLI script. You then query the database using Query editor in the Azure portal.
If the command to install azure-cli using pip3 fails with the following module dependency errors
No package 'libffi' found
c/_cffi_backend.c:15:10: fatal error: ffi.h: No such file or directory
#include <ffi.h>
^~~~~~~
compilation terminated.
Install the dev suite of libffi as follows:
sudo apt-get install libffi6 libffi-devIf the command to install fails with the following error
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/tmp/pip-build-jshgucrb/cryptography/setup.py", line 14, in <module>
from setuptools_rust import RustExtension
ModuleNotFoundError: No module named 'setuptools_rust'
=============================DEBUG ASSISTANCE==========================
If you are seeing an error here please try the following to
successfully install cryptography:
Upgrade to the latest pip and try again. This will fix errors for most
users. See: https://pip.pypa.io/en/stable/installing/#upgrading-pip
=============================DEBUG ASSISTANCE==========================
Upgrade to the latest version of setuptools and pip with the following commands
sudo apt-get install python3-setuptools
sudo pip3 install --upgrade pipIf the command az --version fails with
/usr/bin/az: line 2: /opt/az/bin/python3: cannot execute binary file: Exec format error
You have an invalid Debian version installed which can be removed with.
sudo apt-get remove -y azure-cliOpen a new terminal and re-verify.