Mass flash

Robotics & Edge Computing Jetson Systems Jetson Orin NX
1

Hello,

I have a large number of Jetson Orin NX Module devices, paired with JETSON-ORIN-IO-BASE, in my possession. Some of these devices have not been flashed and remain in their initial state, while others have been updated to the latest version using NVIDIA’s SDK Manager . I’ve successfully set up one of the Jetson devices with the necessary applications and configurations. Now, I’m faced with the task of replicating this setup across an additional 50-100 Jetson devices.

I have tried creating backups using the command:

./tools/backup_restore/l4t_backup_restore.sh -b jetson-orin-nano-devkit-nvme

and then restoring with:

./tools/backup_restore/l4t_backup_restore.sh -r jetson-orin-nano-devkit-nvme

However, I encountered an issue where the boot order does not automatically configure. I’d like to emphasize that it’s crucial for me to avoid manually setting the boot order on each device, as connecting a monitor to each of the 50-100 devices isn’t feasible.

Additionally, I attempted cloning with:

sudo ./flash.sh -r -k APP -G img_name p3509-a02+p3767-0000 nvme0n1p1

but this process seems to take considerably longer than expected – well over two hours. I decided to terminate the process without waiting for it to complete due to its prolonged duration.

I’m reaching out to ask for your advice: what best practices and solutions would you recommend for my task? I’m looking for the most efficient and reliable way to mass duplicate the setup across a large number of devices.

Thank you in advance for your assistance and advice.

2

Jetpack version 35.4

4

Hi,

So what’s your boot order after restoring?
If you only have one NVMe drive as the booting device, the device should supposedly boot into it regardless of the booting order.

5

when the boot order is not set, it goes through boot options (network boot, etc.) and this boot lasts 5 minutes

Start HTTP Boot over IPv6.
Error: Could not retrieve NBP file size from HTTP server.

Error: Unexpected network error.

Start HTTP Boot over IPv4.
Error: Could not retrieve NBP file size from HTTP server.

Error: Server response timeout.

Start PXE over IPv6.
PXE-E16: NO valid offer received.

6

image
image1643×47 11.9 KB

these are the files the backup team made for me

7

Hi,

Do you have an Ethernet cable or PCIe devices plugged in during boot up?
If yes, please try removing them as they may be recognized as bootable devices, hence changing the boot order.
If not, you may try re-build the UEFI binary after removing network stack in the config:

Sharing serial console log during boot up may also help debugging.

8

The problem is exactly this, as I indicated above. But the original Jetson does not have this problem.

You don’t have the means to copy the entire image along with the bios and restore it to many other jetsons?

9

So is it solved this way?

Boot order is not related to backup/restore.
The image is cloned completely, but the behavior of UEFI may change upon boot.

10

No, this is not a solution. I need the jetson on which I restored the firmware from my backup to also turn on normally. But on jetson after recovery it does not have boot order settings

11

Then please try this.

12

understood you correctly. You suggest that I modify the bios on the Jetson source and remove downloading from the network from there, and then clone this system using ./tools/backup_restore/l4t_backup_restore.sh?

13

YES. Exactly.