After using nvmultiurisrcbin and using nvvideoconvert and capsfilter in the pipeline, the program Segmentation fault (core dumped)

Accelerated Computing Intelligent Video Analytics DeepStream SDK
, ,
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) GPU
• DeepStream Version docker-ds6.3
• JetPack Version (valid for Jetson only)
• TensorRT Version 8.5
• NVIDIA GPU Driver Version (valid for GPU only) 550.142
• Issue Type( questions, new requirements, bugs)

Here’s how my pipes are connected:

        self.nvmultiurisrcbin.link(self.queue1)
        self.queue1.link(self.preprocess)
        self.preprocess.link(self.queue2)
        self.queue2.link(self.pgie)
        self.pgie.link(self.queue3)
        self.queue3.link(self.converter)
        self.converter.link(self.queue4)
        self.queue4.link(self.capsfilter)
        self.capsfilter.link(self.queue5)


        if self.nvdslogger:
            self.queue5.link(self.nvdslogger)
            self.nvdslogger.link(self.tiler)
        else:
            self.queue5.link(self.tiler)
        self.tiler.link(self.queue6)
        self.queue6.link(self.nvvidconv)
        self.nvvidconv.link(self.queue7)
        self.queue7.link(self.nvosd)
        self.nvosd.link(self.queue8)
        self.queue8.link(self.sink)

The self.converter and self.capsfilter plugins look like this:

self.converter = Gst.ElementFactory.make("nvvideoconvert", f"converter-2RBGA")
        self.converter.set_property("gpu_id", GPU_ID)
        self.capsfilter = Gst.ElementFactory.make("capsfilter", f"capsfilter-2RGBA")
        caps = Gst.Caps.from_string("video/x-raw(memory:NVMM), format=RGBA")
        # self.converter.set_property("nvbuf-memory-type", 0)
        mem_type = int(pyds.NVBUF_MEM_CUDA_UNIFIED)
        self.converter.set_property("nvbuf-memory-type", mem_type)
        self.tiler.set_property("nvbuf-memory-type", mem_type)
        self.capsfilter.set_property("caps", caps)

The reason I added these two plugins in the pipeline is that I need to convert the raw images to opencv format images in the probe function

When I used these two plugins, the video stream would crash after it finished. Segmentation fault (core dumped), but not when I comment out these two plugins

Here’s my code
code.zip (16.6 KB)

3

Please follow the platform compatibility first. Installation — DeepStream documentation

4

My graphics card is 4080s, it doesn’t seem to fit R525.125.06

5
  1. The DeepStream 6.3 nvdsmultiurisrcbin is just alpha version, there may be some bug in it.
  2. Have you debug with your code to find out whether the segmentation fault is related to the DeepStream APIs? Can the sample /opt/nvidia/deepstream/deepstream/sources/apps/sample_apps/deepstream-server work for your case?
6

Seems you need to upgrade to the latest DeepStream version.

7

I encountered another phenomenon on the basis of this environment and program, when I used nvmultiurisrcbin plugin to open the code stream of two cameras, I manually unplugged the network cable of one camera, after a period of time, the program crashed, I wanted to test whether it was a problem with the graphics card driver, I was on the 3060 graphics card this phenomenon occurred, I can’t find the R525.125.06 driver on the official website, can you give me a connection, or can you check if there is this problem before

image
image1196×252 111 KB

8

The 525.125.06 driver can be downloaded here Linux x64 (AMD64/EM64T) Display Driver 525.125.06 | Linux 64-bit | NVIDIA

Can the sample /opt/nvidia/deepstream/deepstream/sources/apps/sample_apps/deepstream-server work for your case?

9

I just tried it, the environment didn’t change, and when I unplugged one of the cameras, the phenomenon was the same

image
image1002×913 102 KB

10

image
image1483×879 163 KB

This is the configuration file

11

Please upgrade to the latest DeepStream version.

13

Is this a problem with properly installing docker-ds6.3? My 3060 uninstall 550.142 driver install 525.125 driver will give an error.

14

No. The latest DeepStream fixed many bugs. It is better to use the latest DeepStream version.

15

May I ask if this is an answer to my question about the program exception after the abnormal closure of the code stream above? If ds6.3 is normally installed, this problem will not be encountered, I want to use it on 6.3 as much as possible, and reinstall the operating system to try

16

The answer is for your question of “Is this a problem with properly installing docker-ds6.3?”

There is no clear clue to show what is the root cause of the program exception problem.