One RTSP stream stops after a day or two in production

[kd637]

• Hardware Platform: Jetson
• DeepStream Version 6.0
• JetPack Version 4.6
• TensorRT Version 8.2
• Issue Type: Question

I’ve a two-camera pipeline based on python bindings. I observe that after a day or two, one of the rtsp stream stops in production. Restarting the pipeline reconnects with the stream.
I changed the order of streams, the same stream stops, so my guess is that it is a network/camera issue.
I have two questions:

1. How do I log/catch exceptions[like no-further-possible-decoding] in a stream using the python pipeline’s uridecodebin on a specific stream? [do we have a similar reference app?]

2. How do I reconnect in during runtime [given that this is in python]
   Thanks

[Fiona.Chen]

@kd637

How do I log/catch exceptions[like no-further-possible-decoding] in a stream using the python pipeline’s uridecodebin on a specific stream? [do we have a similar reference app?]

If there is evidence for it is a network/camara issue, you can use “export GST_DEBUG=rtspsrc:7” to get more log for the RTSP client side. There is no way

@kd637

How do I reconnect in during runtime [given that this is in python]
Thanks

Reconnection has nothing to do with python. You can refer to the c/c++ deepstream-app sample for the correct reconnection logic and implement with python.

[kd637]

@Fiona.Chen

Reconnection has nothing to do with python. You can refer to the c/c++ deepstream-app sample for the correct reconnection logic

Yes, I can do that. I was thinking more on the lines of rtsp-reconnect-interval-sec property which is present in c/c++ but (in my understanding) not directly accessible via gst-python, right? So, the only option is to implement the logic?

Thanks for the debug flag.

[Fiona.Chen]

@kd637

So, the only option is to implement the logic?

Yes.