**TL;DR Hardware Platform (Jetson / GPU)** Jetson AGX Orin Dev Kit DeepStream Version 7.1 JetPack Version (valid for Jetson only) 6.0 (L4T 36.3.0) TensorRT Version 10.7.0.23 NVIDIA GPU Driver Version (valid for GPU only) N/A (Jetson) Issue Type (questions, new requirements, bugs) Bug / s…

The probe function is a blocking function, if you do the NvBufSurfaceMap/UnMap correctly inside the probe function, there is no problem.

[image] renatomaynardetchepare: uridecodebin uri=rtsp://... (UDP, latency=60, drop-on-latency=true) ! nvv4l2decoder enable-max-performance=1 low-latency=1 disable-dpb=1 num-extra-surfaces=2 ! nvstreammux batch-size=1 live-source=1 sync-inputs=0 frame-num-latest=1 batched-push-timeout=5000 width=…

Hi Fiona, Thank you for your answer, Yes! :), this was the setting I am using:). This crash happens after some hours/ but the increase in latency happens after 5-10 min, really weird. Thank you so much for you time!

[image] renatomaynardetchepare: Surface mapping inside a pad probe (RGBA readback): I briefly map/unmap surfaces in a compositor-src probe to drive an ArUco-based layout switch. If the nvbufsurface_map/unmap usage is off for a given DeepStream/pyds build, it could be a slow-burn until a rare cod…

[image] renatomaynardetchepare: nvv4l2decoder.disable-dpb=true with B-frames: If sid=0 camera uses B-frames or longer reference chains, DPB disabling might cause jitter/instability and eventually trigger a downstream edge case. “disable-dpb” does not work with B frames, the B frames needs a l…

Thanks Fiona, I will do it and let you know how it goes! :)

Hi Fiona, My understanding is that “surface mapping inside a pad probe” refers to calling NvBufSurfaceMap/UnMap (plus NvBufSurfaceSyncForCpu) on an NvBufSurface inside the probe callback so the CPU can read pixels. That can work, but it’s easy to get subtle issues if the map flags are wrong, if UnM…

Thank you for your comment, Fiona. After reviewing my code again and again, I realized I was having a problem with the probe function, and after some time running the model I was accumulating memory and eventually crashing the RAM. Additionally, I changed my pipeline to use independent cameras, each…

DeepStream/GStreamer: One RTSP source’s latency drifts up (others stable) and pipeline crashes after ~7 hours. Decoder set to disable-dpb=true

Accelerated Computing Intelligent Video Analytics DeepStream SDK

Fiona.Chen November 11, 2025, 1:54am 5

Can you refer to the DeepStream SDK FAQ - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums to check whether the nvstreammux parameters and properties are set correctly according to your streams before we start the discussion of your issue?

Topic		Replies	Views
Deepstream-app crash with nvbufsurface: NvBufSurfaceSysToHWCopy error DeepStream SDK nvbugs	27	4945	September 1, 2021
Pipeline Crash DeepStream SDK deepstream	11	283	December 30, 2025
Jetson NX DeepStream problem DeepStream SDK	11	518	February 1, 2024
Jetson AGX Orin (DeepStream 7.1) – Smart Record / custom recording triggers NvBufSurface CUDA faults and NVENC crashes with 13 RTSP sources DeepStream SDK tensorrt , cuda , ubuntu , gstreamer , deepstream , jetson-orin	22	584	December 17, 2025
[deepstream-test.py] Error: Could not map EglImage from NvBufSurface for nvinfer on headless orin DeepStream SDK jetson-inference	6	1240	December 27, 2023
DeepStream. Jittered output DeepStream SDK camera , deepstream	11	1248	May 2, 2023
An understanding of the delay result produced by latency_measurement_buf_probe DeepStream SDK camera , cudnn , deepstream	39	686	December 24, 2024
Pipeline freeze? DeepStream SDK	14	1034	May 27, 2024
Pixelated Frames in Deepstream Pipeline DeepStream SDK	19	591	September 30, 2024
Deepstream 8 Failure DeepStream SDK deepstream	3	88	January 30, 2026

DeepStream/GStreamer: One RTSP source’s latency drifts up (others stable) and pipeline crashes after ~7 hours. Decoder set to disable-dpb=true

Related topics