-
Notifications
You must be signed in to change notification settings - Fork 10
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Error when using cross-compiled plugin inside Nvidia DeepStream container in Jetson #111
Comments
Currently working on exactly what you did. |
The above errors are also raised when I install and use the plugin natively on AGX Orin Jetpack 6.0 (L4T r36.3.0) with pylon SDK 7.4.0 installed. Thus they do not seem to be related to docker. Also, the version of the plugin that raises the above errors in my setup is |
UPDATE While the plugin works after I downgraded the version to 0.7.0 it does not work when running inside a container using the setup and the Dockerfile I described above. I am getting a segmentation fault. I provide the stacktrace below:
Is the fault raised because it cannot find the source file Also, I keep getting two messages |
I managed to solve the above issue by copying all the contents of |
Hi @kostastsing, we now have binary debian packages for multiple debian and ubuntu versions: https://github.com/basler/gst-plugin-pylon/releases/tag/v0.7.3 and an explanation to build the plugin debian packages for NVIDIA deepstream including nvmm support: https://github.com/basler/gst-plugin-pylon?tab=readme-ov-file#debian-nvidia-packaging |
I have built the
gst-plugin-pylon
insidedeepstream-l4t:6.4-triton-multiarch
in a x86_64 workstation using qemu machine emulator. As the plugin depends onlibnvbufsurface.so
library for enabling NVMM feature, I download the L4T BSP inside the container, extract it and copy the corresponding library in/opt/nvidia/deepstream/deepstream/lib
. I provide the code snippet in the dockerfile below:Also, I have previously installed the pylon SDK for arm64 in the container in
/opt/pylon
. The plugin is then built successfully with NVMM support inside the container as you can see below:As, I want to deploy my application to a Jetson device, I build a docker image based on
deepstream-l4t:6.4-samples-multiarch
that contains only the runtime libraries of deepstream. I copy the plugin librarylibgstpylon.so
and all the shared libraries it depends on at runtime into the container from the image I built previously (base). I provide the corresponding dockerfile code snippet below:When, I run the final production image on Jetson device I am taking the below errors:
The plugin is successfully loaded and inspected inside the container, but there are two messages
globbing failed
printed:Regarding the hardware setup, I use two basler GigE cameras each connected to a different network port in a basler GigE interface card that is connected on a Jetson AGX Orin via the PCIe bus.
I have also installed the pylon SDK on Jetson (out of the container) and I have verified that the cameras operate successfully. In addition, to verify that the cameras can be accessed from inside the container I have installed
pypylon
and I have run thestartup.py
sample successfully.Has anyone with a similar setup experienced this problem? Can anyone give a direction in resolving this issue?
The text was updated successfully, but these errors were encountered: