Server indicated failure while installing cuda10.0

Autonomous Machines Jetson & Embedded Systems Jetson AGX Xavier
1

I am trying to install cuda10.0 on my Jetson AGX running Ubuntu18.04.
When I execute the commands as outlined here,
https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html#installcuda
I get the following error:

ubuntu@jetson4:~$ sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
[sudo] password for ubuntu:
Executing: /tmp/apt-key-gpghome.UlvkYiwvW3/gpg.1.sh --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
gpg: requesting key from ‘http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub’
gpg: WARNING: unable to fetch URI http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub: Server indicated a failure

I would appreciate help installing the cuda10.0

Thanks

Esterina

2

Esterina, are you trying to install CUDA on Host PC or Jetson device?
What is the output of the execution of the command below at the target device?

nvcc -V

You may get it installed with sdkmanager when you are doing flashing OS to Jetson

3

I have two systems. In this system, Jetson4, I am trying to install on the Jetson AGX.
When I do nvcc -V, the response is
-bash: nvcc: command not found

Is it possible to install Cuda etc without the SDK manager?

4

If you are trying to install to Jetson AGX, please use sdkmanager;
If you will be targeting x86_64, then corresponding documentation will work out;
did you run nvcc -V at Jetson or at Host PC? At both devices?
cuda could be installed manually, but that would be more complicated, in my opinion: Jetson/Installing CUDA - eLinux.org
what is the output of

nvcc -V

if you execute it at Jetson device?

5

Hi,

On my Jetson4 AGX, I would like to install using this suggestions although it may be more complicated.

Can you please point me to the deb file that I need for the Jetson AGX.
I am following the instructions which states “Download the .deb file for the CUDA Toolkit for L4T either using a web browser on the device,”

Thanks

6

Hi,

You will need to download the required package from sdkmanager.
The package on the website is for desktop environment.

You can download it with sdkmanager and install them manually.The default directly is ${HOME}/Downloads/nvidia/sdkm_downloads/.

Thanks.