CUDA L4T Toolkit Failure

I have flashed an SD card with 4.4 on my NX.

Everything seems to be running find and I can SSH into it via my wifi router using a local LAN address. I can also access if via a public IP address from my host PC.

I am attempting to load Jetson SDK Components from a virtual machine (that was used to flash NX) via the local IP address. It connects OK but cuda always fails. Please see attached for the exported logs from SDKM. (234.1 KB)

Here are the filtered cuda errors

18:21:56 ERROR: CUDA Toolkit for L4T: E: Unable to correct problems, you have held broken packages.
18:21:56 ERROR: CUDA Toolkit for L4T: [error]: dpkg install on target failed; [exec_command]: safe_apt_install “sudo dpkg -i /opt/nvidia/deb_repos/cuda-repo-l4t-10-2-local-10.2.89_1.0-1_arm64.deb” 5; sleep 5; ; safe_apt_install "sudo apt-get -y install cuda-toolkit-10-2 " 5; safe_apt_install “sudo apt-get purge -y cuda-repo-l4t-10-2-local-10.2.89” 5; [deb_path]: /home/ubuntu/Downloads/nvidia/sdkm_downloads/cuda-repo-l4t-10-2-local-10.2.89_1.0-1_arm64.deb
18:21:56 ERROR: CUDA Toolkit for L4T: command terminated with error

Hi alex15,

I couldn’t find the flashing log.
Each version of JetPack is shipped with a bundle of compatible softwares. Target components, like CUDA, OpenCV, etc, shipped in JetPack version A, may be incompatible to L4T OS shipped in JetPack version B. To install target components on it, please make sure the L4T OS on target is compatible. If not sure, please flash Jetson device with the same version of JetPack.

Despite my best efforts to flash things using a VM, I think that something was just always getting missed during the installation process.

I instead moved to running native ubuntu from an external hdd and everything about the flash and SDK loading was a whole lot smoother