"Invalid argument - repository path is not a valid directory" when running local S3 on port 80

Description I’m trying to pull models from a local S3 using --model-repository=s3://cool.url.com:80/bucket_name/path/to/repo as described here but it gives the following error:

error: creating server: Invalid argument - repository path is not a valid directory

Changing the port from 80 to anything else gives access error, which makes sense

error: creating server: Internal - Could not get MetaData for bucket with name <bucket_name>

Removing 80 and using --model-repository=s3://cool.url.com/bucket_name/path/to/repo makes Triton think that the url is the bucket name:

error: creating server: Internal - Could not get MetaData for bucket with name cool.url.com

Triton Information What version of Triton are you using? nvcr.io/nvidia/tritonserver:20.08-py3

Are you using the Triton container or did you build it yourself? I pulled the image

To Reproduce Steps to reproduce the behavior. Get a local S3 running on port 80 Add your models pull them from there following here

Describe the models (framework, inputs, outputs), ideally include the model configuration file (if using an ensemble include the model configuration file for that as well). No relevant

Expected behavior I expect models to be loaded into Triton.

[CoderHam]

Thank you for bringing this to our attention. Could you repeat these steps to confirm it is not a configuration issue on your end?

1. Use S3 command line to confirm the credentials are set correctly.
2. Confirm that the path is a folder and not a file.
3. Use Tritonserver 20.10/20.09
4. Ensure you are following the S3 bucket name conventions.

[deadeyegoodwin]

Closing, please reopen if still an issue.