agentleFS
Seeed-ProjectsSeeed-Projects/Seeed-Jetson-DevelopTool/skills/codex/backup-restore/AGENTS.md

Seeed-Jetson-DevelopTool / backup-restore

AGENTS.md57 starsChanged 4 months agoMITView on GitHub

Backup and restore a reComputer Jetson system image using the L4T backup/restore script. Requires Ubuntu host, USB-C cable, and the matching JetPack BSP package.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this AGENTS.md into the rules it gives an agent

6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run one phase at a time. After each phase:

    line 14
  2. Very specific:

    If output contains [STOP] → stop immediately, consult the failure decision tree below.

    line 16
  3. Very specific:

    If output ends with [OK] → tell the user "Phase N complete" and proceed to the next phase.

    line 17
  4. Distinctive:

    Do these steps in order:

    line 35
  5. Distinctive:

    Verify on the host:

    line 43
  6. Distinctive:

    Ensure the device is in force recovery mode (redo Phase 1), then:

    line 110

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Backup and Restore reComputer Jetson System Image

Hardware required: Ubuntu host PC, USB-C cable, reComputer J3011 (JetPack 5.1.3). The backup and restore operations use the same script with different flags — complete Phases 1–4 for both operations.


Execution model

Run one phase at a time. After each phase:

  • Relay all command output to the user.
  • If output contains [STOP] → stop immediately, consult the failure decision tree below.
  • If output ends with [OK] → tell the user "Phase N complete" and proceed to the next phase.

Prerequisites

RequirementDetail
Host OSUbuntu 20.04 or 22.04 (native, not VM)
CableUSB-C data cable ≤1.5 m
BSP packageJetson_Linux_R35.5.0_aarch64.tbz2 (JetPack 5.1.3)
Flashing packagemfi_recomputer-orin (Seeed massflash image)
Disk space≥64 GB free on host for backup image

Phase 1 — Enter force recovery mode (physical steps, no commands)

Do these steps in order:

  1. Power off the reComputer J3011 completely.
  2. Locate the REC (recovery) button and GND pin on the carrier board.
  3. Hold the REC button (or short REC to GND with a jumper).
  4. While holding REC, connect the USB-C cable from the Jetson to the Ubuntu host.
  5. Apply power to the Jetson; wait 2–3 seconds, then release REC / remove the jumper.

Verify on the host:

lsusb | grep -i nvidia
# Expected: Bus 00x Device 00x: ID 0955:7045 NVIDIA Corp. APX

[OK] when the NVIDIA APX device appears. [STOP] if it does not — see failure decision tree.


Phase 2 — Download and extract BSP

Download Jetson_Linux_R35.5.0_aarch64.tbz2 from the NVIDIA L4T archive, then extract:

tar -xvzf Jetson_Linux_R35.5.0_aarch64.tbz2
# Produces: Linux_for_Tegra/

[OK] when extraction completes without errors and Linux_for_Tegra/ directory exists.


Phase 3 — Merge BSP into flashing package

Copy BSP files into the Seeed massflash package without overwriting existing files:

sudo cp -rn Linux_for_Tegra/* mfi_recomputer-orin/

[OK] when the command completes with no errors.


Phase 4 — Confirm board name

Find the correct board config name before running backup or restore:

cd mfi_recomputer-orin
ls | grep "\.conf"
# Note the board name, e.g. recomputer-orin-j3011.conf → board name is recomputer-orin-j3011

[OK] when you have identified the .conf filename to use as the -b argument.


Phase 5a — Run backup

# Run from inside mfi_recomputer-orin/
sudo ./tools/backup_restore/l4t_backup_restore.sh -e nvme0n1 -b recomputer-orin

Replace recomputer-orin with the board name identified in Phase 4 (without .conf).

Expected: script prints progress and ends with a success message. The backup image is written to the backup/ subdirectory.

[OK] when the script exits cleanly. [STOP] on any error — see failure decision tree.


Phase 5b — Run restore (optional, only if restoring)

Ensure the device is in force recovery mode (redo Phase 1), then:

# Run from inside mfi_recomputer-orin/
sudo ./tools/backup_restore/l4t_backup_restore.sh -e nvme0n1 -r recomputer-orin

Replace recomputer-orin with the same board name used during backup.

[OK] when the script exits cleanly and the device reboots. [STOP] on any error — see failure decision tree.


Failure decision tree

SymptomAction
lsusb shows no NVIDIA APX deviceDevice not in recovery mode. Redo Phase 1. Try a different USB port. Confirm cable is data-rated (not charge-only).
Recovery mode entered but device disappears mid-operationInsufficient power or bad cable. Use the official power adapter and a cable ≤1.5 m.
l4t_backup_restore.sh: No such file or directoryBSP merge incomplete. Rerun Phase 3 and confirm tools/backup_restore/ exists inside mfi_recomputer-orin/.
Script fails with "board not found" or config errorWrong board name passed to -b. Redo Phase 4 and use the exact filename stem from ls *.conf.
Backup script stalls or times outUSB instability. Flash from bare-metal Ubuntu (not a VM). Reconnect USB and retry from Phase 1.
Restore completes but device won't bootBackup image may be corrupt or from a mismatched board. Verify the backup was taken from the same device and board config.

Reference files

  • references/source.body.md — full Seeed Wiki tutorial with screenshots and additional context (reference only)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.