The biggest problem is the confusing documentation.
If I dive into the other problems this post will get lengthy.
I would appreciate if we could solve this problem, given the security issues involved.
There are 3 official docs from Nvidia and they have conflicting information between each, some typos and syntax issues and deprecated stuff.
- First documentation
Following the Nucleus Enterprise doc “Configuring SSL/TLS with NGINX”. On the Section NGINX Configuration from steps 1 to step 3 the documentation tells me to Copy the nginx.ingress.router.conf file from your Nucleus Server to your NGINX Server and place it into the conf.d folder within the NGINX path.
So, basically put the nginx.ingress.router.conf on /etc/nginx/conf.d
https://docs.omniverse.nvidia.com/prod_nucleus/prod_nucleus/enterprise/installation/ssl_nginx.html
- Second documentation
On the Official Nvidia blog post at AWS, which is even cited at Nvidia Docs, at the Section Configure The Reverse Proxy Server steps 2 to step 9 we are told to copy the nginx.ingress.router.conf into S3 named as nginx.conf and then download it at /etc/nginx named as nginx.conf.
And it fails to address all the rest of the required NGINX configuration and the errors within nginx.ingress.router.conf
- This documentation is the best one …
The guide on Nucleus Enterprise to deploy on AWS.
It not only repeat the problems in the AWS post by telling us to update nginx.ingress.router.conf and place directly at /etc/nginx/nginx.conf but it has syntax errors by telling us to copy events { worker_connections 1024; } into the SSL/TLS server and there is a lose note about wrapping the config with http { } but it ignores all the other issues. To complement the CDK stack is has a few errors and the boto3 script is incomplete
https://docs.omniverse.nvidia.com/prod_nucleus/prod_nucleus/enterprise/cloud_aws_ec2.html