Nvmflash hangs


I’m working on to automate the installation process Jetson Nanos using the massflash script (now Im using a devkit to test the installation). But after I start it just hangs with and keeps repeating the “Ongoing processes” messages.

I basically executed these steps:

wget https://developer.nvidia.com/embedded/L4T/r32_Release_v4.4/r32_Release_v4.4-GMC3/T210/Tegra210_Linux_R32.4.4_aarch64.tbz2
wget https://developer.nvidia.com/embedded/L4T/r32_Release_v4.4/r32_Release_v4.4-GMC3/T210/Tegra_Linux_Sample-Root-Filesystem_R32.4.4_aarch64.tbz2
tar jxvf Tegra210_Linux_R32.4.4_aarch64.tbz2
sudo tar jxvf  Tegra_Linux_Sample-Root-Filesystem_R32.4.4_aarch64.tbz2 -C Tegra210_Linux_R32.4.4_aarch64/Linux_for_Tegra/rootfs/

Changed the root password, created a new user, set hostname and timezone as it was described in this post:
TX2 L4T 32.1 Automatic Booting

Created the massflash blob (the nanos will be installed in a trusted environment so I would not need signing and encryption fo now):

    ./nvmassflashgen.sh \
        jetson-nano-devkit-emmc mmcblk0p1

and then moved the tar to the host machine and

tar xvjf mfi_jetson-nano-devkit-emmc.tbz2
cd mfi_jetson-nano-devkit-emmc

put the the Nano dev kit to recovery (it was showing up using lsusb), and execute the follwing command:

Start flashing device: 1-3, PID: 11218
Ongoing processes: 11218
Ongoing processes: 11218

And after that nothing happens, I cannot see anything interesting in the logs, the last message I get is (I will attach the logs: mfi.log (5.1 KB) 20210621-071547_11182_flash_1-3.log (3.3 KB) 20210620-210938_22424_flash_1-3.log (3.3 KB)):

*** Retrieving storage infomation
/home/ubnt/sources/dc_linux/tmp/unpack/massflash/mfi_jetson-nano-devkit-emmc/tegradevflash --instance 1-3 --oem platformdetails storage 11218_storage_info.bin
Cboot version 00.01.0000
Saved platform info in 11218_storage_info.bin

And after I timeout (I let it run overnight) I hit Ctrl-C and I got this error message in the logs right away:
*** Error: Retrieving storage infomation failed.

I tried to use the flash.sh, using that the Nano is installed properly with the new users and stuff.

Do you have any idea why could this happen?

Thank you!


Please move to rel-32.5.1 and try again. Just read the readme file(README_Massflash.txt) in Linux_for_Tegra.

Unfortunately we need to stick with 4.4 since our cameras only have support for that release :(

Does the massflash has this limitation to only support the 5.1 release? I mean the installed version should be transparent from the massflash point of view.

I cannot remember the nvmassflash status of every release. Can we check the board info first?

  1. Is it a custom board or devkit?
  2. Is it a emmc module or sdcard module?
  3. Do you have devkit on your side?


  1. This is a standard B01 version devkit with 4GB of RAM: Buy the Latest Jetson Products | NVIDIA Developer
  2. its an SD card version and I’m using it with a 64 GB SD card
  3. both the host (Ubuntu 18.04) and the Jetson Nano devkit is with me on my table




The massflash does not support sdcard module. For nano, massflash only supports production module (emmc)

The original design for the sdcard module should only work on NV devkit, but not some mass production use.

1 Like

ahhh I see, thats a bummer. Then the only way to test it is with production modules (at least in case of Nanos)?

Also it is a bit misleading that the README_Massflash.txt referencing the Nano as ‘jetson-nano-devkit-emmc’, I have fallen into the trap :D

The name already told you it is for emmc.

Honestly, If you are not sensitive enough to notice that board config is not for your board. I would suggest you start from flash.sh script first. Just get familiar with common flash process first before doing those massflash work.

I know that most jetson nano sd module users rely on the sdcard image a lot so it will take time to learn how to use those tools.

yeah maybe my “reading between the line” skills is not sophisticated enough :D , (also Im pretty sure that Im not relly familiar with the naming conventions) but as the name also contains devkit I assumed it will work with an actual devkit as it was not stated otherwise and all of the devkits I used so far are coming with the SD slot.

Using the simple flash method is working fine, so to move forward with this is to test it with production modules. However it would have be nice if we could test it beforehand, as the process itself should not be much more complicated as a low level copy like dd.

Anyway I guess we can workaround this limitation with using production modules with eMMC in the Dev carry board.

Thanks for the clarifications!

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.