Deepstream Library linking issue in Docker

[devansharma92]

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU): Jetson and GPU
• DeepStream Version: 6.3
• JetPack Version (valid for Jetson only): 5.1.2-b104
• TensorRT Version: 8.5.2-1+cuda11.4 (Jetson) / 8.5.3-1+cuda11.8 (GPU)
• NVIDIA GPU Driver Version (valid for GPU only): 535.171.04

While building a custom app in nvidia deepstream 6.3 docker image provided by nvidia, getting a library linking error, but if some command we are using inside docker it is working fine.

[junshengy]

Unlike X86, docker on jetson, some shared libraries are shared with the host.You can see them in the files below.

/etc/nvidia-container-runtime/host-files-for-container.d/drivers.csv

As a workaround, you can copy them from the host to docker, or just copy binary to docker instead of compiling them when building docker image.

[devansharma92]

where can I find the drivers.csv, it does not seem to be present on the system or the container

[junshengy]

Sorry, I seem to have misunderstood your question.

1.Did you encounter this problem on Jetson or dGPU? As I mentioned above, there are some differences

2.You encountered the problem when compiling the application in docker, right? So how did you start docker? Can you share the command line?

In addition, have you installed nvidia-container-toolkit?

https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html

@junshengy

/etc/nvidia-container-runtime/host-files-for-container.d/drivers.csv

If you don’t find this file on jetson, I suggest you use SDKManager to re-burn the system

[devansharma92]

1. I am facing the same issue on both Jetson and dGPU devices.

2. I am using this command to create the container:

docker create -it --restart=always --net=host --runtime nvidia --privileged -e DISPLAY=$DISPLAY -v /var/run/docker.sock:/var/run/docker.sock -v /usr/bin/docker:/usr/bin/docker -v /tmp/.X11-unix/:/tmp/.X11-unix <Image>

3. nvidia-container-toolkit is already installed.
   
   

[junshengy]

@devansharma92

<Image>

What is your base image? I’m on dGPU and Jetson with nvcr.io/nvidia/deepstream:7.0-triton-multiarch. Both of them work fine

sudo docker create -it --restart=always --net=host --runtime nvidia --privileged -e DISPLAY=$DISPLAY -v /var/run/docker.sock:/var/run/docker.sock -v /usr/bin/docker:/usr/bin/docker -v /tmp/.X11-unix/:/tmp/.X11-unix nvcr.io/nvidia/deepstream:7.0-triton-multiarch

sudo docker start  "container_id"

sudo docker attach "container_id"

cd /opt/nvidia/deepstream/deepstream-7.0/sources/apps/sample_apps/deepstream-test1

make CUDA_VER=12.2

[yingliu]

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks