Metropolis microservices Platform Support and Setup

From RidgeRun Developer Wiki



Preferred Partner Logo 3




As of January of 2024, Metropolis Microservices support two targets:

The targets need to be flashed with Jetpack 6.0+.

Here, you can find a quick start guide to work with Metropolis Microservices for Jetson.

DeepStream install

You will need to install Deepstream on the Jetson, you can do so by using the SDK manager. First open sdkmanager. And make the selection similar to the next figure:


Deepstream install guide step 1


Make sure to have selected DeepStream and click next. On the next step if you dont want to flash the Jetson, untick the Jetson Linux box to only install the SDK components, as seen on figure 2. Then hit continue and the install should start.

Deepstream install guide step 2


Services Install

After that you can install the microservices on the Jetson. Log in into the target Jetson and send the following commands:

sudo apt update
sudo apt install nvidia-jetson-services

Post-install

It's recommended to use the max performace mode:

sudo nvpmodel -m 0

Reboot, and after the platform boots:

sudo jetson_clocks

Another recomendation is to increase the kernel limits on network sockets:

sudo sysctl -w net.core.rmem_default=2129920
sudo sysctl -w net.core.rmem_max=10000000

Finally, NVIDIA runtime should be used by default and docker should be used without sudo. To do so:

1. Modify the /etc/docker/daemon.json file on your system and add "default-runtime": "nvidia"

{
    "runtimes": {
        "nvidia": {
            "path": "nvidia-container-runtime",
            "runtimeArgs": []
        }
    },

    "default-runtime": "nvidia"
}

2. Add your user to the docker group to allow docker command execution without sudo.

sudo groupadd docker
sudo usermod -aG docker $USER

3. Restart docker

sudo systemctl restart docker

At this point, you are ready to start testing sample applications. Move to the next section if you want to learn how.

Sample Applications

If you prefer to start right away testing the available sample applications here is a list of some of them, including RidgeRun's SmartSeek360 demo (learn more later).


SmartSeek360

Smart PTZ system. It captures and analyzes 360-degree video streams using an AI model, and then, based on the model detections it performs PTZ operations and triggers video recordings.

Zero Shot Detection