Jetpack SDK manager 2.1.0 with Jetpack 5.1.3 fails from badly formatted command

Firstly, Logging in 35 times to an internet login, to flash an embedded device is outrageous. Please, code this so it uses the already authorized browser cookie and stop wasting hours of paying customers time.

i have very limited time to provide details as I’ve mostly been relogging into nvidia.com for this absurd UX.

NVIDIA SDK MANAGER v2.1.0.11660

docker run -it --rm sdkmanager:2.1.0.11660-Ubuntu_20.04 --cli --action install --login-type devzone --product Jetson --target-os Linux --version 5.1.3 --host --target JETSON_XAVIER_NX_TARGETS --select 'Jetson Linux' --select 'Jetson Runtime Components' --select 'Jetson SDK Components' --select 'Developer Tools' --flash

info: Installing QEMU binary in rootfs                                                                                                                                                                                                                                                     │
│info: /home/nvidia/nvidia/nvidia_sdk/JetPack_5.1.3_Linux_JETSON_XAVIER_NX_TARGETS/Linux_for_Tegra/rootfs /home/nvidia/nvidia/nvidia_sdk/JetPack_5.1.3_Linux_JETSON_XAVIER_NX_TARGETS/Linux_for_Tegra                                                                                       │
│info: Installing BSP Debian packages in /home/nvidia/nvidia/nvidia_sdk/JetPack_5.1.3_Linux_JETSON_XAVIER_NX_TARGETS/Linux_for_Tegra/rootfs                                                                                                                                                 │
│error: chroot: failed to run command 'dpkg': Exec format error                                                                                                                                                                                                                             │
│error: [exec_command]: /bin/bash -c /tmp/tmp_NV_L4T_FILE_SYSTEM_AND_OS_COMP..sh; [error]: chroot: failed to run command 'dpkg': Exec format error                                                                                                                                          │
│info:                                                                                                                                                                                                                                                                                      │
│info:                                                                                                                                                                                                                                                                                      │
│info: [ Package Install Finished with Error ]                                                                                                              

===== INSTALLATION FAILED =====
- CUDA on Host: Up-to-date
- CUDA Cross Compile Package on Host: Up-to-date
- NvSci: Up-to-date
- VPI on Host: Up-to-date
- Nsight Systems: Up-to-date
- Nsight Graphics: Up-to-date
- Nsight Compute: Up-to-date
- Nsight Perf SDK: Up-to-date
- Drivers for Jetson: Installed
- File System and OS: Failed
- Flash Jetson Linux: DependencyFailure
- DateTime Target Setup: DependencyFailure
- CUDA Runtime: DependencyFailure
- CuDNN Runtime: DependencyFailure
- TensorRT Runtime: DependencyFailure
- OpenCV Runtime: DependencyFailure
- CuPVA Runtime: DependencyFailure
- VPI Runtime: DependencyFailure
- NVIDIA Container Runtime with Docker integration (Beta): DependencyFailure
- Multimedia API: DependencyFailure
- CUDA Toolkit for L4T: DependencyFailure
- cuDNN on Target: DependencyFailure
- TensorRT on Target: DependencyFailure
- OpenCV: DependencyFailure
- VPI on Target: DependencyFailure
- Nsight Systems: DependencyFailure
- Nsight Graphics: DependencyFailure
- Nsight Compute: DependencyFailure

===== Installation failed - Total 28 components =====
===== 1 succeeded, 1 failed, 8 up-to-date, 18 skipped =====

Do you have QEMU installed on the host?

1 Like

The host is the Ubuntu 22.04 sdk manager container provided by NVIDIA.

The command shown calls sdkmanager --cli in the container, so whatever deps nvidia configured would determine if qemu is installed or not.

I followed the user guide exactly.

Please answer my question first.
I’m talking about the host OS, not the docker container.

1 Like

You need qemu-user-static to be installed.

1 Like

Ahh right, the container host. The doc keeps referring the manager software as a “host”… :D understood.

Thanks for the help, I will try that.

1 Like

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