Orin Nano Devkit Automounting block device

Robotics & Edge Computing Jetson Systems Jetson Orin Nano
1

Hello I have a NVIDIA Jetson Orin Nano (8GB ram) - Jetpack 6.1 [L4T 36.4.0]. 699-level Part Number: 699-13767-0003-300 T.1.

I have programed the NVMe to be the main boot drive and a USB eMMC device as a fail back boot device.
Both boot fine and the NVMe is the top of the boot order. All is fine with booting.

However sometimes on a boot something keeps mounting my USB eMMC APP partition to /mnt and sometimes it mounts the esp partition to /mnt as well:

image
image428×683 8.76 KB

I cant find any systemd service or anything similar that’s mounting my eMMC APP or esp partitions and I need to find the root cause. Here is the eMMC and the NVMe PARTLABLES:

image
image800×683 9.89 KB

Im unsure if there is some NVIDIA code somewhere that is mounting my sda partitions because its looking at the PARTLABLES even though I’m not booted to the SDA device. Any help would be much appreciated.

Thanks, Matt

2

*** Please note that this reply is generated by LLM automatically ***

The issue you’re experiencing with your NVIDIA Jetson Orin Nano is related to the mounting of the USB eMMC device’s APP and ESP partitions to /mnt during boot, even though the NVMe drive is set as the primary boot device.

To troubleshoot this issue, let’s break down the possible causes and potential solutions:

  1. PARTLABELS: The PARTLABELS on your eMMC device might be causing the system to mount the partitions automatically. The PARTLABELS are used by the Linux kernel to identify and mount partitions. In your case, the PARTLABELS on the eMMC device are set to APP and ESP, which might be triggering the mounting of these partitions.

    To verify this, you can check the PARTLABELS on your eMMC device using the lsblk command:

lsblk -d -o name,label

   This will show you the device names and their corresponding labels.

2. **Systemd services**: Although you mentioned that you couldn't find any systemd services that might be mounting the eMMC partitions, it's still worth checking the systemd configuration files. You can use the `systemd-analyze` command to analyze the systemd configuration and identify any potential services that might be mounting the partitions.

   Run the following command to analyze the systemd configuration:
   ```bash
systemd-analyze critical-chain

This will show you the critical chain of services that are started during boot.

  1. NVIDIA-specific code: As you mentioned, it’s possible that NVIDIA-specific code is mounting the eMMC partitions. NVIDIA provides a set of scripts and tools for managing the boot process on their devices. You can check the NVIDIA documentation and the Jetson Orin Nano’s boot scripts to see if there’s any code that might be mounting the eMMC partitions.

    One possible location to check is the /etc/init.d directory, which contains the init scripts for the system. You can also check the /etc/systemd directory, which contains the systemd configuration files.

  2. Mount configuration: Another possibility is that the mount configuration is set to mount the eMMC partitions automatically during boot. You can check the /etc/fstab file to see if there are any entries that might be mounting the eMMC partitions.

    Run the following command to check the /etc/fstab file:

cat /etc/fstab

   This will show you the file system table, which contains the mount points and options for the file systems.

To resolve the issue, you can try the following:

1. **Update the PARTLABELS**: You can update the PARTLABELS on the eMMC device to remove the `APP` and `ESP` labels. This might prevent the system from mounting the partitions automatically.

2. **Disable systemd services**: If you find any systemd services that are mounting the eMMC partitions, you can disable them to prevent the mounting from happening.

3. **Modify the NVIDIA boot scripts**: If you find any NVIDIA-specific code that is mounting the eMMC partitions, you can modify the scripts to prevent the mounting from happening.

4. **Update the mount configuration**: You can update the `/etc/fstab` file to remove any entries that might be mounting the eMMC partitions.

By following these steps, you should be able to identify and resolve the issue that's causing the eMMC partitions to be mounted automatically during boot.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***
5

Hi,

Please provide the fstab and /boot/extlinux/extlinux.conf for us to review.

Thanks

6

Here is my /etc/fstab:

image
image909×180 5.98 KB

And here is my /boot/extlinux/extlinux.conf

image
image2006×430 16.5 KB

Also here is an example of double /mnt issue that I have now captured:

image
image433×294 8.72 KB

7

Hi,

We’ll set up the developer kit to check if we can reproduce the issue.

Thanks

8

Hi,

On our side, we were unable to reproduce the issue.

$ lsblk  -o NAME,LABEL,PARTLABEL
NAME         LABEL PARTLABEL
loop0              
sda                
├─sda1             APP
├─sda2             A_kernel
├─sda3             A_kernel-dtb
├─sda4             A_reserved_on_user
├─sda5             B_kernel
├─sda6             B_kernel-dtb
├─sda7             B_reserved_on_user
├─sda8             recovery
├─sda9             recovery-dtb
├─sda10            esp
├─sda11            recovery_alt
├─sda12            recovery-dtb_alt
├─sda13            esp_alt
├─sda14            UDA
└─sda15            reserved
nvme0n1            
├─nvme0n1p1        APP
├─nvme0n1p2        A_kernel
├─nvme0n1p3        A_kernel-dtb
├─nvme0n1p4        A_reserved_on_user
├─nvme0n1p5        B_kernel
├─nvme0n1p6        B_kernel-dtb
├─nvme0n1p7        B_reserved_on_user
├─nvme0n1p8        recovery
├─nvme0n1p9        recovery-dtb
├─nvme0n1p10       esp
├─nvme0n1p11       recovery_alt
├─nvme0n1p12       recovery-dtb_alt
├─nvme0n1p13       esp_alt
├─nvme0n1p14       UDA
└─nvme0n1p15       reserved

Please review your /etc/udev/rules.d configuration to check if there are any rules that automatically mount to /mnt.

Thanks

9

Hi @DavidDDD , Thanks for looking into this. I have checked the udev rules and haven’t found anything that’s mounting /mnt. I also checked in /etc/init.d and did not find anything in there. I did find on my system that this issue didn’t happen on every boot and was irregular.

image
image1907×299 25.8 KB

Could it be some code in the /boot/initrd gzip image?

10

Hi,

Please refer to official sections Modifying the RAM Disk to check whether any auto mount in initrd

Thanks