Nvp model service crashes after installing JP6.0 Rev1 on Orin Nano 8GB developer kit

After flashing my (new) Jetson Orin Nano with Jetpack 6 (rev1) with the sdkmanager the nvpmodel service fails during boot (among a few other services) and Ubuntu starts throwing errors. A quick systemctl status check confirms that the service didn’t start. I have tried different NVMe’s with all the same result.

Rolling/flashing back to Jetpack 5.1.3. solves the problem, so it is not the board and it seems to be a Jetpack 6 issue.

  • Orin Nano 8Gb development kit (serial 1421424601925)
  • NVMe 1: Samsung SSD 970 EVO Plus, 2 Tb
  • NVMe 2: Samsung SSD 990 Pro, 1 Tb

Steps taken:

I flashed both JP 5.1.3. and JP 6.0 (rev 1) on both NVMe’s and the 2 configs with JP 6.0 (rev 1) give the same issue on the nvpmodel service.

They were all flashed in recovery mode with SDK manager on a Linux host.

I include the syslog from the last flash with JP 5.1.3., which shows succesfull loading of the nvpmodel service and the syslog after flash with Jetpack 6.0 (rev1) showing the service failes.

Any help is much appreciated.
syslog.txt (1.1 MB)

syslog_JP513.txt (1.4 MB)

1 Like

Please move to rel-36.3. Jetpack6.0 rev1 is probably rel-36.2 which is developer preview version.

Thanks for your reply, will try and get back.

Hi,

I have flashe again with Jetpack 6.0 rev2 thru the sdkmanager. Although LT-36.3 is installed, I get the same problem: nvpmodel crashes. Here is my sys info:

Output from systemctl status nvpmodel.service:

Jun 17 17:42:21 ubuntu systemd[1]: Starting nvpmodel service…
Jun 17 17:42:21 ubuntu nvpmodel.sh[676]: /etc/systemd/nvpmodel.sh: line 13: 677 Segmentation fault (core dumped) >
Jun 17 17:42:21 ubuntu systemd[1]: nvpmodel.service: Main process exited, code=exited, status=139/n/a
Jun 17 17:42:22 ubuntu systemd[1]: nvpmodel.service: Failed with result ‘exit-code’.
Jun 17 17:42:22 ubuntu systemd[1]: Failed to start nvpmodel service.

output from jetsonUtilities:jetsonInfo:

NVIDIA NVIDIA Jetson Orin Nano Developer Kit
L4T 36.3.0 [ JetPack UNKNOWN ]
Ubuntu 22.04.4 LTS
Kernel Version: 5.15.136-tegra
CUDA NOT_INSTALLED
CUDA Architecture: 8.7
OpenCV version: 4.5.4
OpenCV Cuda: NO
CUDNN: 8.9.4.25
TensorRT: NOT_INSTALLED
Vision Works: NOT_INSTALLED
VPI: 3.1.5
Vulcan: 1.3.204

Any thoughts are welcome

Could you share the full dmesg?

Also, any line after this? Looks like something should be there after core dumped error.

And what would happen if you run this script manually later? Will it show error too?

Sorry, missed a bit: here is the complete line after (core dumped):

nvpmodel.service - nvpmodel service
Loaded: loaded (/etc/systemd/system/nvpmodel.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Thu 1970-01-01 01:00:33 CET; 54 years 5 months ago
Process: 675 ExecStart=/etc/systemd/nvpmodel.sh (code=exited, status=139)
Main PID: 675 (code=exited, status=139)
CPU: 67ms

Jun 17 19:02:37 ubuntu systemd[1]: Starting nvpmodel service…
Jan 01 01:00:33 ubuntu nvpmodel.sh[675]: /etc/systemd/nvpmodel.sh: line 13: 677 Segmentation fault (core dumped) /usr/sbin/nvpmodel -f /etc/nvpmodel.conf
Jan 01 01:00:33 ubuntu systemd[1]: nvpmodel.service: Main process exited, code=exited, status=139/n/a
Jan 01 01:00:33 ubuntu systemd[1]: nvpmodel.service: Failed with result ‘exit-code’.
Jan 01 01:00:33 ubuntu systemd[1]: Failed to start nvpmodel service

If I sudo bash nvpmodel.sh, no (error) message, but same result on the service:

jetson@ubuntu:/etc/systemd$ sudo bash nvpmodel.sh
jetson@ubuntu:/etc/systemd$ systemctl status nvpmodel.service

nvpmodel.service - nvpmodel service
Loaded: loaded (/etc/systemd/system/nvpmodel.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Thu 1970-01-01 01:00:33 CET; 54 years 5 months ago
Process: 675 ExecStart=/etc/systemd/nvpmodel.sh (code=exited, status=139)
Main PID: 675 (code=exited, status=139)
CPU: 67ms

Jun 17 19:02:37 ubuntu systemd[1]: Starting nvpmodel service…
Jan 01 01:00:33 ubuntu nvpmodel.sh[675]: /etc/systemd/nvpmodel.sh: line 13: 677 Segmentation fault (cor>
Jan 01 01:00:33 ubuntu systemd[1]: nvpmodel.service: Main process exited, code=exited, status=139/n/a
Jan 01 01:00:33 ubuntu systemd[1]: nvpmodel.service: Failed with result ‘exit-code’.
Jan 01 01:00:33 ubuntu systemd[1]: Failed to start nvpmodel service.
lines 1-12/12 (END)

Further I include the dump from dmesg.
dmesg.txt (59.4 KB)

I see. Looks like we got similar issue reported.

We will check.

Many thanks, looking forward to your findings.

Sorry that we are still debugging this issue.
Could you help to share the result of following commands on your board?

$ cat /etc/nv_boot_control.conf
$ cat /etc/nv_tegra_release

Thanks for getting back. I know it doesn’t help but to answer your question, I did a fresh flash with Jetpack 6.0 rev2 and this time (after the last 10 flashes) the nvpmodel loads without a problem. Here is the result of your commands:
etson@ubuntu:~$ cat /etc/nv_boot_control.conf

TNSPEC 3767-300-0005-N.1-1-1-jetson-orin-nano-devkit-
COMPATIBLE_SPEC 3767–0005–1–jetson-orin-nano-devkit-
TEGRA_BOOT_STORAGE nvme0n1
TEGRA_CHIPID 0x23
TEGRA_OTA_BOOT_DEVICE /dev/mtdblock0
TEGRA_OTA_GPT_DEVICE /dev/mtdblock0

jetson@ubuntu:~$ cat /etc/nv_tegra_release

R36 (release), REVISION: 3.0, GCID: 36191598, BOARD: generic, EABI: aarch64, DATE: Mon May 6 17:34:21 UTC 2024

KERNEL_VARIANT: oot

TARGET_USERSPACE_LIB_DIR=nvidia
TARGET_USERSPACE_LIB_DIR_PATH=usr/lib/aarch64-linux-gnu/nvidia

I’ll do another flash to see if the problem comes back and let you know.

regards.

Hi,
I have just done two flashes, one without installing the Jetson SDK Components (only runtimes on CUDA and CUDA-X-AI) and a flash with all the SDK components installed. (SDK components are not checked as standard install components in de sdkmanager)

Conclusion:
-When the SDK components are NOT installed, the problem with the nvpmodel comes up and the service crashes.

-When the SDK components ARE installed the service starts as expected during boot.

Hope this will give you some direction.

regards

It seems the issue is specific to Orin Nano 8G module (SKU 0005).
We are still finding the root cause.

Is it 100% reproducible in your case?

To be sure I did two other flashes today resp. with and without the SDK components and the behavior is 100% reproducible: With only the runtimes installed the nvpmodel crashes. If the full SDK components are installed the service starts and runs without a problem.

I have another Jetson Orin Nano 8GB dev board running here on which I can do the same test. Let me know if this would be helpful, it will take me a week or two to free it up.

Sorry that I’m not clear about your status.

=> It seems not relating to SDK components

=> It seems relating to SDK components.

The above 2 statements sounds conflict to me.

Hi, sorry if I wasn’t clear. I tried to tell you that I verified again the behavior by doing another flash with only the runtimes installed and in that config the nvpmodel fails. In the next flash I checked all the sdk components in the sdk manager and in that config the nvpmodel loads without a problem. So I reproduced the problem.

.

For the nvpmodel loaded failed issue in JP6.0GA, please refer to Segfault in /usr/sbin/nvpmodel - #23 by KevinFFF for the fix.

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