Pipeline freeze?

This is the command: python3 flavionew.py -o screen -i rtsp://xxx:xxx@10.21.45.19 rtsp://xxx:xxx@10.21.45.19 rtsp://xxx:xxx@10.21.45.19 rtsp://xxx:xxx@10.21.45.19

These are additional facts I collected so far:

  1. nvgstplayer-1.0 is still running for almost 4 days with no problem to get the RTSP stream
  2. The warning gst-resource-error-quark: Could not read from resource. lead me to this post Error with gst-resource-error-quark, but I already changed the batched-push-timeout to fit my 25fps RTSP stream, that is, batched-push-timeout=1000000/25
  3. I changed the camera by another and the freezing problem persists.
  4. I also removed all appliances from lab network, and created a physically segregated network, just in case. Now I have the machine device with the GPU using a static IP, the camera with static IP, and both of them connected to a hub. The freezing still occurs.
  5. At this post Adjusting muxer's batch push timeout based on FPS of fastest source 40000 - #6 by bpheng it is suggested to make streammux.set_property("sync-inputs", False). This property is not described at the manual (Gst-nvstreammux — DeepStream 6.3 Release documentation) but is describer in a new nvstreammux (Gst-nvstreammux New — DeepStream 6.3 Release documentation). Since there was no complain on such property, I suppose I am using this new nvstreammux. However, the freezing persists.