Disconnecting after reaching the reconnection count, there is still information about this stream in printing

Accelerated Computing Intelligent Video Analytics Metropolis Microservices for Jetson
1

Using the RTSP protocol to disconnect the stream after reaching the reconnection count, and then disconnecting after reaching the reconnection count, there is still information about this stream in printing, which can also be obtained through the REST API
config.txt (5.6 KB)
deepstream-person-detect-dla0-v100.zip (241.8 KB)

2

Below seeting in your config.txt means rtstsrc will try to reconnect 60 times, each reconnection need wait 60 seconds. Can you please set rtsp-reconnect-attempts=3? So the stream will be removed after 3x60 seconds.

[source-attr-all]
enable=1
type=3
num-sources=1
gpu-id=0
cudadec-memtype=0
latency=100
rtsp-reconnect-interval-sec=60
select-rtp-protocol=4
rtsp-reconnect-attempts=60

3

• Hardware Platform (Jetson / GPU) Jetson AGX
• DeepStream Version 7.0
**• JetPack Version (valid for Jetson only) 6.0 GA
**• TensorRT Version 8.6
**• NVIDIA GPU Driver Version (valid for GPU only) 540.3.0
**• Issue Type( questions, new requirements, bugs) questions

4

This is the morning log. I would like to ask why there is no output after * * PERF:, only uri:/pai/v1/stream/getstream info method: GET is printed. Did the PERF program exit, and now
Is Deepstream normal?
deepstream-person-detect-dla0-v100.zip (7.8 KB)

捕获
捕获1431×774 24.4 KB

5

The phenomenon in the morning was reproduced, and a flow was manually added afterwards, but it remained the same.
deepstream-person-detect-dla0-v100.zip (24.8 KB)
config.txt (5.6 KB)

6

I will try to reproduce the issue in my side and feedback later.

7

Thanks

8

If you find it difficult, you can first identify the cause of the problem and see if it is related to the network

9

I will simulate unstable RTSP with NVStreamer. I will stop NVStreamer and check the result of DeepStream.

10

I tested the feature in my side. It works fine in my side:

  1. Add on video in NVStreamer.
  2. Add the NVStreamer RTSP to VST.

$ curl -XGET ‘http://localhost:9010/api/v1/stream/get-stream-info’
{
“reason” : “GET_LIVE_STREAM_INFO_SUCCESS”,
“status” : “HTTP/1.1 200 OK”,
“stream-info” :
{
“stream-count” : 1,
“stream-info” :
[
{
“camera_id” : “00”,
“camera_name” : “17862d01-5506-4524-9dd3-8996e740cba9”
}
]
}
}
4. Stop NVStreamer: $ docker stop nvstreamer
5. Waiting for 3 x 60 seconds and get this log in DeepStream “Resetting source 0, attempts: 3”
6. DeepStream will print below log:

nvstreammux: Successfully handled EOS for source_id=3
new stream removed [3:00]
Removed stream : sub-batch = 1 seq index = 0 pad index = 3

  1. Check DeepStream stream info:
    $ curl -XGET ‘http://localhost:9010/api/v1/stream/get-stream-info’
    {
    “reason” : “GET_LIVE_STREAM_INFO_SUCCESS”,
    “status” : “HTTP/1.1 200 OK”,
    “stream-info” :
    {
    “stream-count” : 0,
    “stream-info” :
    }
    }
11

deepstream-person-detect-dla0-v100.zip (42.0 KB)

12
  1. Starting deepstream docker compose-p 33000-f compose_agx.yaml up-d based on Docker
  2. Add/stream based on RESTful API (can add multiple streams with random intervals)
13

deepstream-person-detect-dla0-v100.zip (68.8 KB)

14

Sorry, yesterday’s test was for the entire system. Now we are running deepstream test app directly on the machine
This can reduce the impact on other factors
The operation steps are as follows, and this log can reproduce the phenomenon of the real question I wanted to ask back then.
deepstream-person-detect-dla0-v100.zip (4.6 KB)

May I ask if the deepstreamtest5 app has an automatic deletion function? If there is any reason why it has not been triggered,

15

操作步骤1.启动deepstream-app程序
2、基于restfulapi 添加
curl -XPOST ‘http://localhost:33002/api/v1/stream/add’ -d '{
“key”: “sensor”,
“value”: {
“camera_id”: “1701”,
“camera_name”: “front_door”,
“camera_url”: “rtsp://192.168.31.50:1935/29/386103317”,
“change”: “camera_add”
}
}

16

I remember you said DeepStream will remove camera source with above setting. Can you share your configure file? Are you set above parameters?

17

Can you add some log in apps/deepstream/sample_apps/deepstream-app/deepstream_app.c to debug why pause_perf_measurement() has been called for DeepStream no perf log issue?

18

config.txt (5.6 KB)

20

I don’t see any issue with the configure file. Can we focus on the entire system test and try to debug the no perf log issue?

21

Why we should focus on the entire system test? If we test the entire system .
There are other factors have affect on this probelm, This can not reduce the impact on other factors