S3 Object Lambda Access Point: HTTP 301 errors with Ruby application

Question

I have a Ruby on Rails application with Active Storage, which uses S3 as the storage service. Active Storage obtains the bucket's information through a `storage.yml` file, which defines `service`, `region`, and `bucket` for S3. This has been working in production for several years now.

Currently, we are trying to add an **S3 Object Lambda Access Point**. As [this article](https://aws.amazon.com/blogs/aws/new-use-amazon-s3-object-lambda-with-amazon-cloudfront-to-tailor-content-for-end-users/) mentions, Access Point's Alias can be used interchangeably with the bucket name, so we are replacing the bucket name for the AP alias in the `bucket` field of our storage.yml file.

However, we are getting HTTP 301 errors (we've confirmed that the key exists in the bucket, and that the Lambda function is not being invoked):

```
2.5.5 :017 > variant.service.exist?(variant.key)
  S3 Storage (115.6ms) Checked if file exists at key:  (no)
Traceback (most recent call last):
        1: from (irb):17
Aws::S3::Errors::Http301Error (Aws::S3::Errors::Http301Error)
```

When using the bucket name (instead of the alias), the typical S3 URL for the objects, once presigned, looks like:
`https://.s3..amazonaws.com//`

We are using the latest version of the `aws-sdk-s3` gem, and we have checked that the access point’s alias is correct.

Any help you could provide? Maybe Object Lambda Aliases can't be used with the `aws-sdk-s3` gem, even when the article mentions that they can be used as bucket names?

Answer

Hello,

I have not specifically used Active Storage for S3, but when working with Dynamo or other databases I have found that the URL for end points vs access can differ. Given that HTTP 301 errors usually indicate a redirection this may be a good place to start. In the context of AWS S3, this may mean that the request is being made to the wrong endpoint. When using S3 Access Points, I believe the endpoint URL structure is different from the standard S3 bucket URL structure. The standard structure is https://bucket-name.s3.region.amazonaws.com, but for access points, the structure is https://access-point-name-123456789012.s3-accesspoint.region.amazonaws.com.

It's possible that the AWS SDK for Ruby is not correctly handling the Access Point alias and is trying to use the standard bucket URL structure, leading to the 301 redirect error. You might need to manually specify the correct endpoint URL when creating the S3 client in your Ruby code.

Maybe adding a line like:

```
s3 = Aws::S3::Resource.new(region: 'us-west-2', endpoint: 'https://access-point-name-123456789012.s3-accesspoint.region.amazonaws.com')

```

Hopefully this can add a starting point!

S3 Object Lambda Access Point: HTTP 301 errors with Ruby application

Relevant content