Problem with Installation of jetson-inference / cmake

Issue Overview

Users are experiencing difficulties installing the jetson-inference library on the Nvidia Jetson Orin Nano Dev board, specifically during the CMake configuration step. The main symptoms include:

• Errors during Docker execution: Users report receiving a "command not found" error when trying to run Docker commands.
• CMake warnings and failures: Users encounter warnings related to the CUDA_TOOLKIT_INCLUDE-NOTFOUND message, indicating issues with CUDA paths during the CMake configuration step.
• Unmet dependencies: Attempts to install CUDA lead to dependency issues, with users receiving messages about conflicting packages and unmet dependencies.

The issue typically arises after following installation instructions from the GitHub repository for jetson-inference, particularly after setting up JetPack and running Docker containers. The users are utilizing an Nvidia Jetson Orin Nano with 8 GB of RAM and a 2 TB SSD. The frequency of this issue appears to be consistent among new users attempting similar setups, significantly impacting their ability to utilize AI image evaluation features.

Possible Causes

• Docker Not Installed: Users may not have Docker installed, which is necessary for running the jetson-inference setup.

• CUDA Not Properly Installed: The CUDA toolkit may not have been installed correctly as part of the JetPack installation, leading to missing paths during configuration.

• Configuration Errors: Incorrect environment variables or paths set for CUDA can result in CMake failing to locate necessary files.

• Driver Issues: Conflicts between installed drivers and the required libraries can lead to unmet dependencies during package installations.

• Environmental Factors: Using an SSD instead of a microSD card may introduce compatibility issues if not properly configured.

• User Misconfigurations: Users might be following outdated or incorrect installation instructions that do not align with their specific system configurations.

Troubleshooting Steps, Solutions & Fixes

1. Install Docker:

   • Ensure Docker is installed by running:

     sudo apt-get install docker.io
     

   • Verify installation by checking the Docker version:

     docker --version
     

2. Set Up CUDA Environment Variables:

   • If CUDA is installed, set the following environment variables:

     export PATH=/usr/local/cuda-12.2/bin:$PATH
     export LD_LIBRARY_PATH=/usr/local/cuda-12.2/lib64:$LD_LIBRARY_PATH
     

3. Check CUDA Installation:

   • Confirm that CUDA is included with JetPack by checking:

     ls /usr/local/cuda/
     

   • If it’s not present, reinstall JetPack ensuring that CUDA components are selected.

4. Resolve Unmet Dependencies:

   • Run the following command to fix broken installations:

     sudo apt --fix-broken install
     

   • If conflicts persist, consider removing incompatible packages using:

     sudo apt-get remove <package-name>
     

5. Reinstall jetson-inference:

   • After resolving any dependency issues, clone the repository again and follow these commands:

     git clone --recursive --depth=1 https://github.com/dusty-nv/jetson-inference
     cd jetson-inference
     mkdir build
     cd build
     cmake .. 
     make
     

6. Consult Documentation:

   • Refer to updated installation guides or forums for any changes in installation procedures or requirements.

7. Best Practices for Future Installations:

   • Always ensure that your system is updated before starting installations:

     sudo apt-get update && sudo apt-get upgrade
     

   • Follow the official installation instructions closely and verify compatibility with your specific hardware setup.

8. Additional Resources:

   • For further troubleshooting regarding CUDA installation on Ubuntu, consult the official NVIDIA documentation or community forums.

By following these steps, users should be able to resolve their installation issues with jetson-inference on their Nvidia Jetson Orin Nano Dev board. Unresolved aspects may include specific compatibility issues related to using an SSD instead of a microSD card, which may require additional investigation or community support.