"Bucket name already exists", but it is not listed on the S3 Mgmt Console

Pull up a web browser and go to http://example-bucket.s3.amazonaws.com, substituting your desired bucket name for example-bucket. If you get any response other than an XML error that says NoSuchBucket, then the bucket name is already taken.

The bucket namespace is global (not per-account), so if someone else already has a bucket by that name, you won't be able to create it.

The workaround is to create a bucket with a different name, configure it the way you want for the redirect, make a note of the bucket's web site hosting endpoint, create a CloudFront distribution, configure the origin domain name as the web site hosting endpoint for the new bucket, configure the CloudFront Alternate Domain Name as your original name (the hostname you want to be able to redirect), and then point the DNS to the assigned CloudFront endpoint. This configuration allows you to have a bucket whose name does not need to match the CNAME record, because CloudFront handles the remapping automatically. (These steps probably sound complicated, but once you've done it, it all makes sense).

An advantage of this solution is that you can also associate a certificate from Amazon Certificate Manager so that https access to the hostname you want to redirect works, as well as http. (When you use S3 directly, the redirection feature only supports http).

@AWS-staff: This global namespace thing is absurd. a) Remove it; Why should we have to compete with the entire world for names for our internal products? We don't have to compete for EC2 instance names, or IAM usernames, etc. This is a confusing exception. b) Until you remove it, please clearly state this requirement in the error on the bucket create page. It's unnecessarily time-consuming for every admin who uses AWS to have to research this issue. Please remedy.

What does global mean? If someone in the world creates a bucket with the name "csv-files", no one in the world can create a bucket with the same name?