No access to CUDA repo from CUDA L4t docker container

This ticket is a near duplicate of this ticket.

I can do whatever I want if I use a regular ubuntu:latest container as the base container. If I want to use cuda, however, I get errors. This is the start of my docker container.

FROM nvcr.io/nvidia/l4t-cuda:10.2.460-runtime
SHELL ["/bin/bash", "-c"]

I get the following error (I am running on a Jetson Orin):

W: Failed to fetch http://cuda-internal.nvidia.com/release-candidates/kitpicks/cuda-r10-2-tegra/10.2.460/006/repos/ubuntu1804/arm64/InRelease  Could not resolve 'cuda-internal.nvidia.com'

I am not modifying the original container - I am simply adding files to the new container in the user directories.
Here is the full error of the start of the image build process:

docker build dockerfile/ --tag cuda_ros
Sending build context to Docker daemon  7.168kB
Step 1/22 : FROM nvcr.io/nvidia/l4t-cuda:10.2.460-runtime
 ---> ac921ae1d2ac
Step 2/22 : SHELL ["/bin/bash", "-c"]
 ---> Running in 5b1eabcd4f67
Removing intermediate container 5b1eabcd4f67
 ---> 3ecd75e20bfd
Step 3/22 : RUN DEBIAN_FRONTEND=noninteractive apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -qq --yes --no-install-recommends     apt-file     apt-utils     apt-transport-https
 ---> Running in 0fd5ff381a02
Err:1 http://cuda-internal.nvidia.com/release-candidates/kitpicks/cuda-r10-2-tegra/10.2.460/006/repos/ubuntu1804/arm64  InRelease
  Could not resolve 'cuda-internal.nvidia.com'
Get:2 http://ports.ubuntu.com/ubuntu-ports bionic InRelease [242 kB]
Get:3 http://ports.ubuntu.com/ubuntu-ports bionic-updates InRelease [88.7 kB]
Get:4 http://ports.ubuntu.com/ubuntu-ports bionic-backports InRelease [83.3 kB]
Get:5 http://ports.ubuntu.com/ubuntu-ports bionic-security InRelease [88.7 kB]
Get:6 http://ports.ubuntu.com/ubuntu-ports bionic/multiverse arm64 Packages [153 kB]
Get:7 http://ports.ubuntu.com/ubuntu-ports bionic/restricted arm64 Packages [572 B]
Get:8 http://ports.ubuntu.com/ubuntu-ports bionic/main arm64 Packages [1285 kB]
Get:9 http://ports.ubuntu.com/ubuntu-ports bionic/universe arm64 Packages [11.0 MB]
Get:10 http://ports.ubuntu.com/ubuntu-ports bionic-updates/restricted arm64 Packages [5682 B]
Get:11 http://ports.ubuntu.com/ubuntu-ports bionic-updates/main arm64 Packages [2093 kB]
Get:12 http://ports.ubuntu.com/ubuntu-ports bionic-updates/universe arm64 Packages [2083 kB]
Get:13 http://ports.ubuntu.com/ubuntu-ports bionic-updates/multiverse arm64 Packages [5351 B]
Get:14 http://ports.ubuntu.com/ubuntu-ports bionic-backports/universe arm64 Packages [20.1 kB]
Get:15 http://ports.ubuntu.com/ubuntu-ports bionic-backports/main arm64 Packages [64.0 kB]
Get:16 http://ports.ubuntu.com/ubuntu-ports bionic-security/universe arm64 Packages [1385 kB]
Get:17 http://ports.ubuntu.com/ubuntu-ports bionic-security/restricted arm64 Packages [5008 B]
Get:18 http://ports.ubuntu.com/ubuntu-ports bionic-security/multiverse arm64 Packages [2761 B]
Get:19 http://ports.ubuntu.com/ubuntu-ports bionic-security/main arm64 Packages [1713 kB]
Fetched 20.3 MB in 7s (3031 kB/s)
Reading package lists...
W: Failed to fetch http://cuda-internal.nvidia.com/release-candidates/kitpicks/cuda-r10-2-tegra/10.2.460/006/repos/ubuntu1804/arm64/InRelease  Could not resolve 'cuda-internal.nvidia.com'
W: Some index files failed to download. They have been ignored, or old ones used instead.

Is any more information required?

Hi @robotsgotmeso, can you disable that apt server by running this command in the container:

mv /etc/apt/sources.list.d/cuda.list /etc/apt/sources.list.d/cuda.disabled

After that, I’m able to run apt-get update in the container.

Hello,

That appears to have gotten rid of the error messages. Is there a more long-term solution to this problem? Doesn’t this change simply lock the cuda software version from any future updates? I can see how that would be considered a good thing. Are there any other implications for this change?

Thanks

That reference to that internal apt server will be removed from future versions of the container so that apt-get update works. It was used for us to build the container and not for upgrades to JetPack (for that, you could add the JetPack apt servers like l4t-jetpack container has).

That makes sense - only engineers inside the internal NVidia network would gain access to those apt servers. Thanks for the help!