Getting Started with reComputer Industrial | Seeed Studio Wiki

[giscus[bot]]

Getting Started with reComputer Industrial | Seeed Studio Wiki

reComputer Industrial Getting Started

https://wiki.seeedstudio.com/reComputer_Industrial_Getting_Started/

[NiemalsSebi]

Why do I need the custom Image? Some things are easier with the version installed via Nvidia SDK Manager e.g. pip, curl already installed.
What is the difference?

[thomaskleiven]

You need a custom image to ensure compatibility with your specific board's configuration files and DTB trees (among other things). While Nvidia SDK Manager's image has conveniences like pre-installed pip and curl, a custom image ensures everything works out of the box for your hardware.

[thomaskleiven]

Do you also have a Jetpack 6 image for the reComputer Industrial J4012 (16GB)? In the download section, it seems like only the 8GB image is available.

[josue-scheidegger]

Is there a specific reason why the J3010 (Jetson Orin Nano 4GB) custom image comes with CUDA-Toolkit 11.8, instead of the CUDA Toolkit 12.2 or 12.4?

[JorgeAlberto91MS]

¿Downloading the OS image takes forever and ever, could you please improve the link?

[rushi-kharade]

I want to make a NVMe for one jetson device and update the jetpack. I followed the instructions to flash the jetson device and create a golden NVMe. Now, I want to clone this golden NVMe to use on other jetson. When I insert the cloned NVMe, I get Invalid header detected on UEFI suppiler FDT,ignoring ... and it never boots up. Is there a way around it? I dont want to keep on flashin ubuntu on every jetson

[nfs0619]

Hello, flashing the system involves more than just writing the OS to the SSD or NVMe; it also includes updating the firmware (such as the bootloader and UEFI settings). Therefore, directly cloning the NVMe may lead to boot failure. Each Jetson device has different hardware and firmware configurations, and the cloned system may not be able to recognize and boot correctly.

[sosnus]

Hi, I just try flash Seedstudio Jetson recomputer industrial J4012 device. I add 1TB nvme disk, but it is not important for me where will be installed system: on Jetson or on nvme disk.

I try flash with Jetpack 5.1 using command:

sudo ./tools/kernel_flash/l4t_initrd_flash.sh --flash-only --massflash 1 --network usb0 --showlogs

and get output:

Writing gpt_secondary_3_0.bin (parittion: secondary_gpt) into /dev/mtd0
Sha1 checksum matched for /mnt/internal/gpt_secondary_3_0.bin
Writing /mnt/internal/gpt_secondary_3_0.bin (16896 bytes) into  /dev/mtd0:67091968
Copied 16896 bytes from /mnt/internal/gpt_secondary_3_0.bin to address 0x03ffbe00 in flash
[ 177]: l4t_flash_from_kernel: Successfully flash the qspi
[ 177]: l4t_flash_from_kernel: Failed to write to APP
[ 177]: l4t_flash_from_kernel: The device size indicated in the partition layout xml is smaller than the actual size. This utility will try to fix the GPT.
[ 177]: l4t_flash_from_kernel: Error flashing external device
Flash failure
Cleaning up...

When I try do same, but using Jetpack 6.0, i get another output.
Method to flash Jetpack 6.0:

 sudo ./tools/kernel_flash/l4t_initrd_flash.sh --flash-only --massflash 1 --network usb0 --showlogs

last lines of output:

Copied 109 bytes from /mnt/internal/qspi_bootblob_ver.txt to address 0x03fe0000 in flash
Writing gpt_secondary_3_0.bin (parittion: secondary_gpt) into /dev/mtd0
Sha1 checksum matched for /mnt/internal/gpt_secondary_3_0.bin
Writing /mnt/internal/gpt_secondary_3_0.bin (16896 bytes) into  /dev/mtd0:67091968
Copied 16896 bytes from /mnt/internal/gpt_secondary_3_0.bin to address 0x03ffbe00 in flash
[ 176]: l4t_flash_from_kernel: Successfully flash the qspi
[ 176]: l4t_flash_from_kernel: The device size indicated in the partition layout xml is smaller than the actual size. This utility will try to fix the GPT.
[ 176]: l4t_flash_from_kernel: Failed to write to APP
[ 176]: l4t_flash_from_kernel: Error flashing external device
Flash failure
Either the device cannot mount the NFS server on the host or a flash command has failed. Debug log saved to /tmp/tmp.NKjX8q6cJu. You can access the target's terminal through "sshpass -p root ssh root@fc00:1:1:0::2" 
Cleaning up...

How to fix it?

[nfs0619]

Hello, could you provide more details, such as your complete operation steps and full log output? Based on the information you have provided so far, it is difficult to accurately determine the issue.

If convenient, you can contact our technical support team at techsupport@seeed.io for further assistance. Our engineers will get in touch with you as soon as possible to help resolve the issue.

[mdrst]

Hi by any chance do you have a compilation of Tensorflow with GPU, that currently works with Orin NX 16GB: 0955:7323 NVidia Corp j4012 jetpack 6.1, tried but seems the build of your image crashes