• Hardware Platform (Jetson / GPU) Laptop with NVIDIA GPU RTX 4060 • DeepStream Version 7.0 • TensorRT Version 8.6.1.6-1+cuda12.0 • NVIDIA GPU Driver Version Driver Version: 550.107.02 CUDA Version: 12.4 • Issue Type Questions
Greetings.
Faced an interesting behavior of rest server (nvds_rest_server) used inside nvmultiurisrcbin, in conjunction with nvds_obj_enc and glog.
Here is some sample code that reproduces the problem:
If we run the code as is, it causes a segmentation fault inside libcivetweb.so when receiving/parsing a rest request.
The version of the glog library I’m using is 0.29.2, installed via apt install in ubuntu 22.04.
If I comment out the glog or nvds_obj_encode API calls, rest server starts processing requests correctly and the error does not occur.
What is the cause of this behavior and how to resolve the conflict?
Since it’s a problem of using nvds_obj_enc and glog together, could you try to reproduce that with our sources\apps\sample_apps\deepstream-image-meta-test?
If it can be reproduced in our demo, we can analyze the problem much faster.
It looks like you are right about a problem of using nvds_obj_enc and glog together.
Interestingly, the problem did not manifest itself until some time later when we decided to add nvmultiurisrcbin to our pipeline.
The problem is very easy to reproduce in any code.
Here is an example of a patch that reproduces the problem in deepstream-image-meta-test.
Also, don’t forget to install glog via apt install libgoogle-glog-dev and add it to your Makefile. You should also add -lnvds_rest_server to the Makefile.
I have tried that with our deepstream-server demo. The code does not compile properly.
Could you provide us the Makefile and diffs of our code based on deepstream-server?
Since Gstreamer has a perfect log mechanism, why would you want to use the GoogleLog?
There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks