Frequent boot failure in Jetson Orin

Robotics & Edge Computing Jetson & Embedded Systems Jetson Orin NX
, ,
1

Hi all,
we are incurring in a boot failure as in the following screenshot:
Jetson System firmware version v36.3.0 date 2024-02-09T18:35:29+00:00
ESC to enter Setup.
F11 to enter Boot Manager Menu.
Enter to continue boot.
…
*** L4TLauncher: Attempting Recovery Boot***
Android image header not seen
Failed to boot recovery:0 partition

>>Start HTTP Boot over IPv4…
From here the system is unable to continue and we need to reset the UEFI partition.
Any ideas about why this is happening? When we return to the same partition and reboot, the system boots normally, so it does not seem to be a data corruption.
We incurred in this issue with both Jetpack 5.2 and using Yocto/meta-tegra with both Linux 5.10 and Linux 6.6

3

Hi albert.burbea,

Are you using the devkit or custom board for Orin NX?

It seems there’s no available boot device recognized in UEFI so that it is trying to boot from network.
What’s your boot device?

May I know what’s your customization?

4

Hi
it’s a custom design.
What makes it even stranger is that when we tell specifically the UEFI to boot from the failed partition, it boots! So, it doesn’t look like a disk corruption.

I didn’t understand

5

How did you perform this?
Is your issue about boot-order?
The priority of booting from network is higher than from NVMe SSD.
If so, please try to re-configure the boot order in UEFI menu.

6

Hi,
this is my current boot manager configuration

/------------------------------------------------------------------------------
| Boot Manager |
------------------------------------------------------------------------------/

                                                     Device Path :

Boot Manager Menu VenHw(1E5A432C-0466-4D
31-B009-D4D9239271D3)/
UEFI WD IX SN530 SDBPTPZ-170G-XI 222720800027 1 MemoryMapped(0xB,0x141
UEFI HTTPv4 (MAC:48B02DF7A9D7) 00000,0x1411FFFF)/PciR
UEFI PXEv6 (MAC:48B02DF7A9D7) oot(0x0)/Pci(0x0,0x0)/
UEFI PXEv4 (MAC:48B02DF7A9D7) Pci(0x0,0x0)/NVMe(0x1,
UEFI HTTPv6 (MAC:48B02DF7A9D7) 69-05-72-4B-8B-44-1B-0
UEFI Shell 0)

Use the <^> and keys to choose a boot option,
the key to select a boot option, and the
key to exit the Boot Manager Menu.

                                                                            �I/TC: WARNING: Test UEFI variable auth key is being used !

/------------------------------------------------------------------------------
| |
| ^v=Move Highlight =Select Entry Esc=Exit |
------------------------------------------------------------------------------/

7

As you can see, the first boot device is always the local SDD.
What we are not understanding is : why the Orion thinking that the partition is corrupt (and we know it is not, since we can reboot manually from it)
This happens quite often. Is there any log/additional data we can look at?

8

No customization here. We simply use the UEFI boot manager to force it to reboot from the same device

9

If it can not boot from the first boot device, then it will try to boot from the following boot device.

How did you flash the NVMe SSD?
Is there the issue on the devkit or the issue is specific to your custom carrier board?

10

we are using meta-tegra for bitbake and its flashing tools to flash the device’s SSD.
We haven’t tested it extensively on the dev kit. The thing we are not understanding is why we are able to reuse the very same partition after we force UEFI to use it, using its config menu.

11

Are you using Yocto rather than L4T on Orin NX?

Sorry that I’m not quite understand what you mean here.
Could you elaborate on this and state your issue?

You can also remove the unexpected boot devices like IPv4, PXE…etc in your case.

13

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