NvidiaDeepStreamSDK module got error with Azure IoTEdge

Question

I followed this tutorial to get to this step"Start monitoring the messages sent from the device to the cloud".But I didn't succeed, and I found that NvidiaDeepStreamSDK module cannnot work.

** ERROR: <create_primary_gie_bin:107>: Failed to create 'primary_gie_classifier' ** ERROR: <create_primary_gie_bin:161>: create_primary_gie_bin failed ** ERROR: <create_pipeline:1045>: create_pipeline failed ** ERROR: main:1096: Failed to create pipeline Quitting

How to solve this?

Answer 1

I understand that you’re encountering issues with the NVIDIA DeepStream SDK module while trying to monitor messages from your RTSP camera to the cloud via Azure IoT Edge.

The error messages you’ve encountered provide valuable clues. Let’s break down the issues and explore potential solutions: Failed to create ‘primary_gie_classifier’:

• This error indicates that the primary GIE (Generalized Inference Engine) classifier could not be initialized.
  • Possible reasons include incorrect configuration, missing model files, or issues with the model itself.
  create_primary_gie_bin failed: - The create_primary_gie_bin function failed to create the necessary components for the primary GIE.

     - Verify that the required model files (such as **`.etlt`** or **`.engine`** files) are present and accessible. Ensure that the paths specified in your configuration are correct.
  
     **create_pipeline failed**:
        - The overall pipeline creation failed. This could be due to issues with any of the components (source, sinks, plugins, etc.).
  
           - Double-check your pipeline configuration files (**`config_infer_primary.txt`**, **`config_infer_secondary.txt`**, etc.). Look for any typos or inconsistencies.
  

Possible Solutions:

• Model Files: Ensure that you have the correct model files (for object detection, classification, etc.) in the specified paths. If you’re using custom models, make sure they are properly converted and optimized.
• Configuration: Double-check your configuration files. Pay attention to model paths, labels, and other parameters.
• Permissions: Confirm that the user running the application has the necessary permissions to access files and resources.
• Dependencies: Check if all required dependencies (such as CUDA, TensorRT, etc.) are installed and compatible.
• Logs: Examine detailed logs to identify specific issues. Look for additional error messages or warnings.
• Environment Variables: Set any required environment variables (e.g., LD_LIBRARY_PATH, GST_PLUGIN_PATH) correctly.
  • Reboot: Sometimes, a system reboot can resolve issues related to environment variables or cached configurations.

Answer 2

• Double-check the Azure IoT Edge configuration for the NVIDIA DeepStream SDK module. Verify that the module is correctly provisioned and connected to the IoT Hub.
• Review the security daemon file at /etc/iotedge/config.yaml to ensure it contains the correct settings.

---

• If your RTSP camera is using a hostname (e.g., mycamera.local), ensure that DNS resolution works correctly. You can add the hostname and its corresponding IP address to the /etc/hosts file on the Jetson Nano.
• Alternatively, you can use the IP address directly in the RTSP URL.

Answer 3

• If your RTSP camera is using a hostname (e.g., mycamera.local), ensure that DNS resolution works correctly. You can add the hostname and its corresponding IP address to the /etc/hosts file on the Jetson Nano.
• Alternatively, you can use the IP address directly in the RTSP URL.

FYI - Apologies, but I only understand English.

Answer 4

Hi, I can see that Desired status is Running, can you please share the screen where it is showing that it is not running.

Answer 5

Hi, is your problem resolved?