Using OverlayFS on Jetson Orin Nano (Ubuntu 20.04)

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

Hi everyone,

I’m trying to set up OverlayFS on my Jetson Orin Nano, but I haven’t been able to find clear instructions that work. I’ve attempted several approaches, including modifying /boot/extlinux/extlinux.conf and updating /etc/fstab, but these changes have only resulted in the Jetson becoming unbootable.

The closest reference I found is this forum post:
OverlayFS mount fail R36.3.0.
However, it doesn’t provide specific guidance for implementing OverlayFS on JetPack 6.

Could anyone who has successfully configured OverlayFS on a Jetson Orin Nano running Ubuntu 20.04 share their experience or point me to detailed instructions?

Thank you!

3

Hi JanserKcc,

Are you using the devkit or custom board for Orin Nano?
What’s the Jetpack version in use?

Are you using Jetpack 5.x (R35.x)?
Please share the result of $ cat /etc/nv_tegra_release on your board.

Please refer to Overlayfs on Root File System for this feature.

4

Hi @KevinFFF,

Thank you for your quick response!

I successfully configured OverlayFS with JetPack 6.1 following your suggestion. Here’s a summary of what I did, in case it helps others facing a similar issue.

My previous setup:

  1. Host PC: Ubuntu 20.04, Jetson OS: Ubuntu 20.04
  2. JetPack: 4.5.1
    (As mentioned in my earlier post, there isn’t an easy way to enable OverlayFS with JetPack 4.5.1.)

Solution:

  1. Host PC: Ubuntu 20.04, Jetson OS: Ubuntu 22.04
  2. Used SDK Manager to upgrade the Jetson Orin Nano to JetPack 6.1.
  3. Followed the instructions in [1] and ran sudo nv_overlayfs_config -e, then rebooted to enable OverlayFS.
  4. To test if OverlayFS is working:
    After setting up OverlayFS and rebooting, create a file (e.g., ~/1.txt).
    Reboot again and verify if the file (~/1.txt) is gone—this confirms that OverlayFS is functioning correctly.

Caution:

Upgrading JetPack from 4.5.1 to 6.1 will also upgrade your CUDA version from 12.1 to 12.6. Note that as of 2024-12-03, the officially supported CUDA version for PyTorch is 12.4. You may need to check compatibility before proceeding.

Also the Jetson OS will be Ubuntu 22.04, if you are using ROS1 noetic before, this will be an issue.

Reference:
[1] OverlayFS on Root File System (NVIDIA Developer Guide)
[2] Pytorch installation on JetPack6.1

1 Like
5

Hi @KevinFFF,

Another issue I encountered when switching to JetPack 6.1 is that the Intel RealSense D455 appears to be incompatible with JetPack 6.1 at this time.

If possible, I would prefer to maintain the old configuration while implementing a method to protect the operating system from corruption (e.g., using overlayFS or another suitable solution).

  • JetPack version: 5.1.2-b104
  • OS: Ubuntu 20.04
  • CUDA 11.4
  • Camera: Intel Realsense D455
version detail 
$ sudo apt-cache show nvidia-jetpack | grep "Version"
Version: 5.1.2-b104

$ lsb_release -a
No LSB modules are available.
Distributor ID:	Ubuntu
Description:	Ubuntu 20.04.5 LTS
Release:	20.04
[nv_overlayfs_config.txt|attachment](upload://ggVcrmjnQP1HzrKdUuQgtDQLpz8.txt) (3.4 KB)

Codename:	focal

nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Sun_Oct_23_22:16:07_PDT_2022
Cuda compilation tools, release 11.4, V11.4.315
Build cuda_11.4.r11.4/compiler.31964100_0

I also reviewed the nv_overlayfs_config script on JetPack 6 to see if it could be adapted for JetPack 5.1.2. However, it appears that implementing this requires a deeper understanding of kernel rebuilding, which is quite challenging for me.
nv_overlayfs_config.txt (3.4 KB)

6

OverlayFS feature is not supported in JP5.x.

For this issue, I would suggest you opening another topic to discuss in details.
You may also request the vendor for the porting guide.

8

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