NvBufSurfTransform failed with error -3 while converting buffer

[nrsw5c06]

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
GPU

• DeepStream Version
6.1.1

• JetPack Version (valid for Jetson only)
• TensorRT Version

• NVIDIA GPU Driver Version (valid for GPU only)
515.86.01
• Issue Type( questions, new requirements, bugs)
Error in deepstream pipeline

• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)

• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)
nv-infer

Hello,

We are getting the following error after running for about 50 frames in a video.

nvinfer gstnvinfer.cpp:1389:convert_batch_and_push_to_input_thread: error: NvBufSurfTransform failed with error -3 while converting buffer

Our deepstream pipeline has multiple secondary models, and the addition of another model, breaks the pipeline. We are dynamically adding obj meta to the frames to be used as area of interest for the AI models. The dynamic setting of AOI is where we can isolate the error.

To me it seems like adding obj meta to the frame where the AOI might change in between frames tries to access unallocated memory. We use PYDS to add the metadata.

There are no resources on this error online.

Could you please inform us what is means and potentially how to fix it?

[yuweiw]

The error means parameter configuration error:

  /** Specifies invalid input parameters. */
  NvBufSurfTransformError_Invalid_Params = -3,

So it’s werid that you got the error after running for about 50 frames in a video. Theoretically, the first frame should report the error. Could you share your method to add the metadata with PYDS and attach your video source? Also could you use our demo to duplicate the problem?deepstream-test2. It has multiple secondary models too.

[nrsw5c06]

Thank you. We did indeed have a parameter configuration error.

We managed to debug the issue down to the frame where it happened. However, what was weird was that the frame which produced the error was different in different environments.

We did solve it. We were dynamically setting PGIE bboxes for a secondary model. Those PGIE boxes were sometimes assigned to coordinates outside of the frame, which led to the error. So a parameter configuration error indeed.

Maybe this post will lead someone else down the right path. Thank you for getting back to us.