Need help to re-flash a jetpack on my orin agx 64 bit developper kit

[chakib.dj67]

Dear NVIDIA Support Team,

I am experiencing an issue while attempting to flash JetPack 6.0 on my Orin AGX 64-bit Developer Kit using the SDK Manager. The process terminates with an error. Below are the details of the error:

Previously, I successfully flashed JetPack on this board, but I encountered numerous dependency issues and memory problems during the installation of CUDA and cuDNN. These issues resulted in missing dependencies and unresolved errors. Due to these problems, I decided to reflash JetPack on the board to start afresh.

Could you please assist me in resolving this issue?

here is the log files that i get via the sdk manager :
SDKM_logs_JetPack_6.0_(rev._2)_Linux_for_Jetson_AGX_Orin_modules_2024-07-26_15-03-11.zip (57.4 KB)

[chakib.dj67]

Ps : I have already yaml 6.0 installed on my host c

[nagesh_accord]

@chakib.dj67

but I encountered numerous dependency issues and memory problems during the installation of CUDA and cuDNN

To. successfully install CUDa and other frameworks use this command to install after you successfully flash the unit
$sudo apt-get install jetpack

[chakib.dj67]

The problem is that now my jetson couldn’t boot so i tried to re-flash the jetpack via thé sdk manager and get those errors in my log from the sdk

[AastaLLL]

Hi,

Based on your log, the device is not in recovery mode which causes the flashing failure.

15:03:30.591 - Error: [exec_command]: /bin/bash -c /tmp/tmp_NV_L4T_FLASH_JETSON_LINUX_COMP.chakib.sh; [error]: --- Error: Reading board information failed.

As you might have already tried, the recovery mode is set by the below command:

15:03:30.536 - Info: "/home/chakib/nvidia/nvidia_sdk/JetPack_6.0_Linux_JETSON_AGX_ORIN_TARGETS/Linux_for_Tegra/bootloader/tegraflash.py" --chip 0x23 --applet "/home/chakib/nvidia/nvidia_sdk/JetPack_6.0_Linux_JETSON_AGX_ORIN_TARGETS/Linux_for_Tegra/bootloader/mb1_t234_prod.bin" --cfg readinfo_t234_min_prod.xml --dev_params tegra234-br-bct-diag-boot.dts --device_config tegra234-mb1-bct-device-p3701-0000.dts --misc_config tegra234-mb1-bct-misc-p3701-0000.dts --bins "mb2_applet applet_t234.bin" --skipuid --cmd "dump eeprom cvm cvm.bin; reboot recovery" 

Which fails with the ModuleNotFoundError: No module named 'yaml' error.

How do you verify the yaml installation on your host PC?
Could you try the below command to see if the same error appears?

$ cd /home/chakib/nvidia/nvidia_sdk/JetPack_6.0_Linux_JETSON_AGX_ORIN_TARGETS/Linux_for_Tegra/bootloader
$ python3
# from tegrasign_v3_util import *

Thanks.

[chakib.dj67]

Hello, i try your command and now i have this log :
SDKM_logs_JetPack_6.0_(rev._2)_Linux_for_Jetson_AGX_Orin_modules_2024-07-29_09-49-21.zip (79.5 KB)
i dont understand cause my jetson is recognized via a manually recovery mode boot so i selected it after that via the sdk manager and even with the importation of all modules in tegrasign_v3_util there is the problem with the yaml

[chakib.dj67]

@AastaLLL ?

[chakib.dj67]

@KevinFFF can you help me on this please ?

[AastaLLL]

Hi,

Is there any error when running the test shared above?

The command is not a fix but to identify if any requirement is missing.
It is the similar command that applying in the SDKmanager.

In your 0729 log, it looks like the command still fails with the yaml library.

09:49:32.186 - Info:   File "/home/chakib/nvidia/nvidia_sdk/JetPack_6.0_Linux_JETSON_AGX_ORIN_TARGETS/Linux_for_Tegra/bootloader/tegrasign_v3_util.py", line 23, in <module>
09:49:32.186 - Info:     import yaml
09:49:32.186 - Info: ModuleNotFoundError: No module named 'yaml'

Thanks.

[chakib.dj67]

@AastaLLL I have no error when testing the command cause i have i have yaml installed to perform thé requirement but can be find by thé sdk manager

[AastaLLL]

Hi,

Could you try to flash the device manually?

https://docs.nvidia.com/jetson/archives/r34.1/DeveloperGuide/text/IN/QuickStart.html

$ cd /home/chakib/nvidia/nvidia_sdk/JetPack_6.0_Linux_JETSON_AGX_ORIN_TARGETS/Linux_for_Tegra/rootfs/
$ sudo tar xpf ../../${SAMPLE_FS_PACKAGE}
$ cd ..
$ ./tools/l4t_create_default_user.sh -u nvidia -p nvidia -a
$ sudo ./apply_binaries.sh
$ sudo tools/board_automation/boardctl -t topo recovery
$ sudo ./flash.sh ${BOARD} mmcblk0p1

Thanks.

[chakib.dj67]

Hello @AastaLLL , i just get this error :

[AastaLLL]

Hi,

Please check the link above to set up the environment variables accordingly.

• ${SAMPLE_FS_PACKAGE} contains the sample file system package’s filename: Tegra_Linux_Sample-Root-Filesystem_L4TRELEASEVERSION.tbz2 .

Thanks.