How to increase the robustness of deepstream

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) GPU
• DeepStream VersionDeepStreamSDK 6.1.0
• JetPack Version (valid for Jetson only)
• TensorRT Version8.4
**• NVIDIA GPU Driver Version (valid for GPU only)**510.47.03
**• Issue Type( questions, new requirements, bugs)**bugs
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)

I use Deepstream to connect multiple RTSP, if one of the connections has network fluctuations, the whole program will be interrupted, I am in this link ( app-more-robust-in-multi-sources/237593) saw your reply, but the current upgrade environment is more complicated, and it is temporarily impossible to upgrade to 6.1.1, can it be achieved through python coding? For example, when an rtsp has an exception When connecting, block the exception directly, and other streams continue to run, or catch this exception and remove it from the pipeline. Is it handled in the bus_call method? But I am new to deepstream, and I don’t know how to do it, please help me
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

As to the case you mentioned, if the rtsp exception can trigger gstreamer bus error, you can handle the error in bus callback. The relevant elements can be stopped and disconnected from the pipeline and reconncted and restarted by the basic GStreamer operations. States (

thank you for your reply. After evaluation, I finally decided to upgrade to 6.1.1, but I followed the 6.1.1 official document to install CUDA Toolkit 11.7.1 (CUDA 11.7 Update 1) and NVIDIA driver 515.65.01. But why is cuda12.1 installed in the end? And the graphics card driver is R530.30.02 at this time, but I read the official document, the cuda version requirement of DS 6.1.1 is 11.7, and the graphics card driver is R515.65.01, what should I do?

Please uninstall the TensorRT and CUDA before installing the GPU driver.

Since your GPU is RTX 3090, you can download driver from Linux x64 (AMD64/EM64T) Display Driver | 515.86.01 | Linux 64-bit | NVIDIA

After the diver is installed correctly, you can following the instructions Quickstart Guide — DeepStream 6.1.1 Release documentation to install CUDA and TensorRT.

ok thanks for your suggestion

I completely uninstalled cuda and tensorRT, and then through [Linux x64 (AMD64/EM64T) Display Driver | 515.86.01 | Linux 64-bit | NVIDIA] (Linux x64 (AMD64/EM64T) Display Driver | 515.86.01 | Linux 64-bit | NVIDIA /en-us/)
Download the driver. At this time, I found that the driver is 515.86.01 through nvidia-smi, and then I executed the following command in the 6.1.1 document, and then I executed nvidia-smi. At this time, the cuda version is 12.1, and the graphics card version is 530.30.02. Why is this? The graphics card version is automatically upgraded? Is it caused by the following command? Or is it automatically upgraded by ubuntu?

$ wget
$ sudo dpkg -i cuda-repo-debian11-11-7-local_11.7.1-515.65.01-1_amd64.deb
$ sudo rm /etc/apt/sources.list.d/cuda
$ sudo apt-key adv --fetch-keys
$ sudo add-apt-repository “deb Index of /compute/cuda/repos/ubuntu2004/x86_64 /”
$ sudo apt-get update
$ sudo apt-get -y install cuda

The last command should be “sudo apt-get install cuda-toolkit-11-7”.

The version is finally correct, thank you for your help

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.