Help in setting up Nucleus Server

I’m following these instructions: Installing an Enterprise Nucleus Server — Omniverse Nucleus on a freshly-installed Ubuntu 22.04 LTS box,. and have successfully done the “docker login nvcr.io” command, with a “Login Succeeded” response.

When I issue this command, however:
/opt/ove/base_stack# docker compose --env-file /opt/ove/base_stack/nucleus-stack.env -f /opt/ove/base_stack/nucleus-stack-no-ssl.yml pull

I get this response:
unable to get image ‘nvcr.io/nvidia/omniverse/nucleus-auth:1.5.8’: Error response from daemon: client version 1.53 is too new. Maximum supported API version is 1.41

Next steps? I’m unclear where to start looking; this isn’t covered in the documentation. Thanks.

What Nucleus stack version and docker version are you using in this deployment?

Using the recommended Docker version from that documentation, and I downloaded this stack file:

nucleus-stack-2023.2.9+tag-2023.2.9.gitlab.31778301.bd3f484c.tar.gz

1 Like

i have exactly the same problem.

I was advised to do this before the “pull” command:

export DOCKER_API_VERSION=1.41

and that allowed me to successfully pull from the nvcr.io repository, and to run the Nucleus Server.

Thanks @spencer7 - can you let me know where you got that info from? I’d like to make sure we address this appropriately.

I opened a ticket, yesterday, with NVIDIA Enterprise Support. That was their suggestion.

1 Like

Thanks everyone for the info. We are discussing this issue with the software distribution team and will report back once we have additional information.

In the interim, using the export command @spencer7 mentioned in the thread is the correct workaround.

Florian, perhaps opening a ticket with NVIDIA Enterprise Support would be a good next step for you.

I did not see the warning that you see.