System breaks after filesystem copy from emmc to External storage

Hi Everyone.

Firstly thanks to the community for being the go to place for the endless supply of knowledge and support.

I’ve faced an issue since i bought my device a few months back. (Device type, Software version and steps taken included at the bottom)

My issue is that after flashing my device and being left with not even enough space to build the jetson-inference hello ai repo on the device, I am then forced to copy the emmc file system to an external storage device.

All goes well as far as my knowledge goes with the copy seeing as I’ve had to do the process plenty of times(lol). I can succesfully boot onto the external storage with no errors popping up on the OS’s side.

Where i run into issues is when cloning the dusty-nv/jetson-inference github repo im faced with the 2 options to build the repo.

Option 1
docker/run.sh

Option 2
Build from source

This is where the issues begin. Both options fail regardless of what i do, giving me all sorts of errors ranging from my C Compiler being broken, to major errors with goroutine being broken when using docker to run.

My Ultimate question is.

Is there some dependancies and modules that cannot be copied like i do from the original flashed emmc image - External USB device? The main reason that i believe this is the case, that when i have just flashed to the emmc memory i have no errors when trying to build the repo with docker or the build from source option. But once i have moved the root file system to the external storage i am unable to succesfully build the repo.

I can provide any info needed, and will appreciate any suggestions.

Cheers

Device Details

Jetson nano Dev Kit Model:P3448 (180-13448-DAAA-B01)
Flash method = SDK Manager (I flash both the OS image and SDK components onto the emmc storage)
nvidia-l4t-core = 32.7.4-20230608212426
nvidia-jetpack = 4.6.4-b39
R32 (release), REVISION: 7.4, GCID: 33514132, BOARD: t210ref, EABI: aarch64,

Steps taken.

  1. Flash the OS image and SDK components to the Device via the provided cable, Choosing the preconfig option in the SDK manager.
  2. Boot the device up
  3. Ive tried both options of updating and upgrading ubuntu with apt before copying the file system to the external disk, or to not update and upgrade ubuntu before copying the file system to the external disk (Didnt notice much difference between the 2)
  4. after the data has copied over i modify the extlinux.conf file to point to the USB (dev/sda1)
  5. reboot the machine
  6. upon bootup confirm it has infact booted on to the USB
  7. update and upgrade ubuntu
  8. clone the hello ai inference repo off github
    git clone --recursive --depth=1 GitHub - dusty-nv/jetson-inference: Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson.
    cd jetson-inference
    docker/run.sh

OR

  1. clone the hello ai inference repo off github
    git clone --recursive --depth=1 GitHub - dusty-nv/jetson-inference: Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson.
    cd jetson-inference
    mkdir build
    cd build
    cmake …/

Hi,

Would you mind sharing the error log so we can know more about the issue you face?

Please note that you might need to check out a specific branch to get the source that is compatible with your environment.
For example, it’s recommended to checkout L4T-R32.7.1 and try it again.

Thanks.

1 Like

Thank you so much for the speedy response!

This seems to have done the trick. I will spend some more time on it the next few days and report my findings back to you.

After cloning the reccomended fork last night and running $docker/run.sh , i was presented with the download-models.sh screen as shown by Dusty in the first video, which had never happened. So its off to a great start!

Hi @AastaLLL

Here is the CMakeError.log and CMakeOutput.log after the failed attempt to build the repo from source.

CMakeError.log (4.5 KB)
CMakeOutput.log (449 Bytes)

here is the docker/run.sh error.

errordocker.txt (1.3 KB)

Hi,

Thanks for sharing the log.
Will update more info with you later.

1 Like