Flashing Failure - Jetson Orin Nano Super Dev Kit - JP 6.2.1

Robotics & Edge Computing Jetson Systems Jetson Orin Nano
1

Hi,

I’ve tried both SDKManager and CLI to flash a super nano dev kit using a 256GB SSD (NVMe), but it’s falling over at this point:

***************************************
*                                     *
*  Step 3: Start the flashing process *
*                                     *
***************************************
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for target to boot-up...
Waiting for device to expose ssh ......Waiting for device to expose ssh ...Run command: flash on fc00:1:1:0::2
08:56:28.252 - Debug: Debug log saved to /tmp/tmp.4GXm9UXz6C.
SSH ready
mount.nfs: access denied by server while mounting [fc00:1:1:0::1]:/home/user1/Downloads/nvidia/Linux_for_Tegra/rootfs
08:56:41.569 - Error: Flash failure.
08:56:41.571 - Error: Either the device cannot mount the NFS server on the host or a flash command has failed. Check your network setting (VPN, firewall,...) to make sure the device can mount NFS server. Debug log saved to /tmp/tmp.4GXm9UXz6C. You can access the target's terminal through "sshpass -p root ssh root@fc00:1:1:0::2" 
08:56:41.573 - Debug: The last 5 lines of the debug log are:
SSH ready
mount.nfs: access denied by server while mounting [fc00:1:1:0::1]:/home/user1/Downloads/nvidia/Linux_for_Tegra/rootfs
Cleaning up...

I’ve already tried the following to rectify, but they are all failing in the same way:

  • Wipefs and mkfs.ext4 on the SSD.
  • Removing /etc/exports.
  • Ensuirng nfs-server-kernel is installed via apt:
    • sudo apt install nfs-kernel-server
  • Disabling ufw.
  • Allowing nfs on ufw:
    • sudo ufw allow from fc00:1:1:0::2 to any port nfs
    • sudo ufw allow from fc00:1:1::/48 to any port nfs
  • Different USB-C ports on host
  • Different USB-C data cables

My last attempt which I’m about to try is using a USB-A to USB-C port and cable,

Are there any other suggestions on how to rectify this?

EDIT:

It also failed in the same manner with a now third cable on the USB-A port (the third port I have tried). For what it’s worth, this exact host, cable and usb port was used to successfully flash an AGX Orin many times, without issue,

The SDK terminal output for the failure is:

09:20:48 ERROR: 09:20:48.857 - Error: Flash failure.
09:20:48 ERROR: 09:20:48.859 - Error: Either the device cannot mount the NFS server on the host or a flash command has failed. Check your network setting (VPN, firewall,...) to make sure the device can mount NFS server. Debug log saved to /tmp/tmp.h2u4xbNu9g. You can access the target's terminal through "sshpass -p root ssh root@fc00:1:1:0::2" 
09:20:48 INFO: 09:20:48.860 - Debug: The last 5 lines of the debug log are:
09:20:48 INFO: SSH ready
09:20:48 INFO: mount.nfs: access denied by server while mounting [fc00:1:1:0::1]:/home/user1/nvidia/nvidia_sdk/JetPack_6.2.1_Linux_JETSON_ORIN_NANO_TARGETS/Linux_for_Tegra/rootfs
09:20:48 INFO: Cleaning up...
09:20:49 ERROR: [exec_command]: /bin/bash -c /home/user1/.nvsdkm/replays/scripts/JetPack_6.2.1_Linux/NV_L4T_FLASH_JETSON_LINUX_COMP.sh; [error]: rpcbind: another rpcbind is already running. Aborting
09:20:49 INFO: [ Component Install Finished with Error ]
09:20:49 INFO: [host] [ 101.62 MB used. Disk Avail on Partition rpool/USERDATA/user1_4u0nl4: 3524.86 GB ]
09:20:49 INFO: [ NV_L4T_FLASH_JETSON_LINUX_COMP Install took 9m24s ]
09:20:49 ERROR: command error code: 11
09:20:49 ERROR: command terminated with error
09:20:49 SUMMARY: Installation failed.

EDIT2:
Looking at this bit:
/home/user1/.nvsdkm/replays

I’m going to try rm’ing that dir and everything in it and see if it is the problem?….

EDIT3

Nope, still no luck. I’ve got no further ideas at this point.

4

Well, now I’m really at a loss, as I even added the apparently failing nfs path (/home/user1/nvidia/nvidia_sdk/JetPack_6.2.1_Linux_JETSON_ORIN_NANO_TARGETS/Linux_for_Tegra/tools/kernel_flash/images) to:

/etc/exports:

home/user1/nvidia/nvidia_sdk/JetPack_6.2.1_Linux_JETSON_ORIN_NANO_TARGETS/Linux_for_Tegra/rootfs   fc00:1:1::1/48(rw,sync,no_root_squash,no_subtree_check)

I then ssh’d to the Jetson orin per the notes in the terminal (password is ‘root’, per the sshpass command in the terminal log of sdkmanager):

ssh root@fc00:1:1:0::2

mkdir -p /nfs/nvidia

mount [fc00:1:1:0::1]:/home/olord/nvidia/nvidia_sdk/JetPack_6.2.1_Linux_JETSON_ORIN_NANO_TARGETS/Linux_for_Tegra/rootfs /nfs/nvidia

and…..the thing mounts just fine???
I even made the bob.temp on the host/server side just to be sure it really was seeing that, and there it is:

-bash-5.1# ls -la /nfs/nvidia                                                 
total 107
drwxr-xr-x  18 root    0  24 Aug 23  2025 .
drwxr-xr-x   5 root    0   0 Jan  1 00:06 ..
-rw-rw-r--   1 1000 1000  62 Jun 16  2025 README.txt
lrwxrwxrwx   1 root    0   7 Feb 17  2023 bin -> usr/bin
-rw-r--r--   1 root    0  26 Aug 23  2025 bob.temp
drwxr-xr-x   3 root    0  85 Aug 22  2025 boot
drwxr-xr-x   2 root    0   3 Aug 22  2025 dev
drwxr-xr-x 144 root    0 252 Aug 23  2025 etc
drwxr-xr-x   2 root    0   2 Apr 18  2022 home
lrwxrwxrwx   1 root    0   7 Feb 17  2023 lib -> usr/lib
drwxr-xr-x   2 root    0   2 Feb 17  2023 media
drwxr-xr-x   2 root    0   2 Feb 17  2023 mnt
-rw-r--r--   1 root    0 364 Aug 22  2025 nv_preseed.cfg
drwxr-xr-x   4 root    0   4 Aug 22  2025 opt
drwxr-xr-x   2 root    0   2 Apr 18  2022 proc
drwx------   3 root    0   5 Mar 22  2023 root
drwxr-xr-x  19 root    0  22 Oct 17  2024 run
lrwxrwxrwx   1 root    0   8 Feb 17  2023 sbin -> usr/sbin
drwxr-xr-x   2 root    0   2 Dec  1  2022 snap
drwxr-xr-x   2 root    0   2 Feb 17  2023 srv
drwxr-xr-x   2 root    0   2 Apr 18  2022 sys
drwxrwxrwt   2 root    0   2 Aug 22  2025 tmp
drwxr-xr-x  11 root    0  11 Feb 17  2023 usr
drwxr-xr-x  15 root    0  17 Aug 22  2025 var
5 
sudo ufw status
sudo ufw disable

You will have to completely shut off ufw until sdk manager is finished.

Also make sure your local internet gateway is not killing icmp requests, as in very high security mode.

7

@foxsquirrel1 Disabling ufw had already been tried. Possible on the icmp requests, but I doubt as I encountered the same issue whether i specified flashing using the IP interface over Ethernet, or the IP interface over the USB connection.

8

One thing I did notice was that I did have to fiddle around a couple of times. My case it was the internet gateway and if memory is correct UFW had to be disabled. Also make sure your nvme has been totally wiped. Go back to sdk manager and check just the jetson linux and forget about the other stuff. Also let it download locally first. This is a big download so install it on an hdd or ssd not your primary drive. It does work, sorry I did not take notes. I was in a hurry to get it up because we were waiting on the replacement for months.

(common) fred@orin-dev:~$ df 
Filesystem      1K-blocks     Used Available Use% Mounted on
/dev/nvme0n1p1  958803476 32595444 877429772   4% /
tmpfs             3901336      172   3901164   1% /dev/shm
tmpfs             1560536    51740   1508796   4% /run
tmpfs                5120        4      5116   1% /run/lock
/dev/nvme0n1p10     64511      110     64402   1% /boot/efi
tmpfs              780264      244    780020   1% /run/user/1000
9

All good mate, thanks for trying to help out (I’m in a similar boat to the on you were time and project progress wise). I’m using basically all the tips you have mentioned there already (they are all good pointers though for anyone that stumbles on this in future) as I too had very similar learnings from first experiences with flashes failing on the AGX Orin via SDK Manager. In the end the AGX became pretty reliable for flashing via SDK Manager, but this Orin Nano seems to be a complete n-go.

In this case though, even with all the tips and tricks, it kept falling over at that nfs mount failure message and I’ve given up due to time and just flashed an SD card and booted from that.

10

Keep in mind this thing is a beast to boot up, watch your monitor and see what is going on. I do recall some snags then it finally worked. Do you have the jumper on the correct pins?

11

Yep, was in recovery mode and detected as such in SDKM.

It’s booted okay off the SD card, now it’s failing miserably to build onnxruntime in a container with GPU support……gets to 96%, has some sort of random fault and reboots. Strange, it’s the same dockerfile that built just fine on the AGX Orin. Doesn’t seem to be thermal, tegrastats only reporting max temps of ~50c. The default config for volatile only logging of journald was a pain, so no I have worked around that and run the build a few more times trying to work out why it reboots…so painful on a jetson compared to x84_64 with dGPU.

12

I ripped out docker, it was interfering with some of the networking. Only vm we will use is venv, docker is like typing with gloves on….

1 Like
14

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.