Error building image in Orin developer kit as 4gb nano

Hello!

I will testing the capabilities to use nvblox on the jetson Orin Nano of 4GB, 8GB and Orin NX. I have an AGX Orin developer kit. I followed the instructions to use the AGX as a 4GB nano manually without SDK manager. Use the latest version to format. It worked fine and then I followed the instructions to create the images using realsense. I tried to compile it completely, with the ros2_humble.realsense configuration but it did not work, then I removed realsense and when compiling the problem continues. Compilation failed at [stage-0 17/22] RUN --mount=type=cache,target=/var/cache/apt mkdir -p /opt/ros/humble/src && cd /opt/ros/humble/src && git clone GitHub - moveit/moveit_task_constructor: A hierarchical multi-stage manipulation planner. The logs are the same in both situations, but i think the most important part is the following Error: 243.2 CMake Error at CMakeLists.txt:15 (find_package):
243.2 By not providing “Findpy_binding_tools.cmake” in CMAKE_MODULE_PATH this
243.2 project has asked CMake to find a package configuration file provided by
243.2 “py_binding_tools”, but CMake did not find one.
243.2
243.2 Could not find a package configuration file provided by “py_binding_tools”
243.2 with any of the following names:
243.2
243.2 py_binding_toolsConfig.cmake
243.2 py_binding_tools-config.cmake
243.2
243.2 Add the installation prefix of “py_binding_tools” to CMAKE_PREFIX_PATH or
243.2 set “py_binding_tools_DIR” to a directory containing one of the above
243.2 files. If “py_binding_tools” provides a separate development package or
243.2 SDK, be sure it has been installed.

I installed docker and nvidia-jetpack (6.0+b106) following all the documentation. It seems to be the same issue raised on github

Hi @rodrigo.delgado1

I have tested it and haven’t encountered this bug on my device.

Have you checked if all NVIDIA libraries are installed on your device? You can use the SDK manager just to complete the installation (without reinstall the firmware and OS)

Meanwhile, I will forward your issue to the engineering.

Let me know.

Best,
Raffaello

Hello @Raffaello,

I used the solution in the pr made on github and it worked fine (add to dockerfile the dependencies of python packages).

The new problem is that the 4 GB nano freezes when the Realsense packages are compiled within the container since the jetson runs out of memory space to compile (it uses the 4 available plus 2GB of swap). I’ll try to emulate an 8GB nano to see if it can compile well.

Hi,

I used the solution in the pr made on github

Can you please give a link to the PR?

It should be this one Add py_binding_tools as new dependency for MTC by rhaschke · Pull Request #136 · NVIDIA-ISAAC-ROS/isaac_ros_common · GitHub

3 Likes

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.