S3 temporary redirect - intended handling of such redirects for v2?

[boaks]

Using "s3.amazonaws.com" may cause some temporary redirects after creating the bucket.
I read Why am I getting an HTTP 307 Temporary Redirect response from Amazon S3? and would like to redirect my S3 request during that period.

I found some examples for v1 using withRedirectLocation but I can't find, how that is intended to be done with v2.

Any hints? Examples?

[boaks]

Just some more details:

I use the S3AsyncClient (version 2.19.11, I may update that) and the global endpoint, "s3.amazonaws.com".
With that, I get that "TemporaryRedirect" for a day or so. If I use the region endpoint, "s3..amazonaws.com", it works.

With a couple of lines I can read the redirect address from the error response and it would be possible to setup a second S3AsyncClient with that redirected host.

But I'm wondering, if that code is required or if there is a smarter way, maybe even a configuration which does the same.

[debora-ito]

What withRedirectLocation does in Java SDK v1 is to add a "x-amz-website-redirect-location" header to the request. You can do the same thing in Java SDK v2 by setting a overrideConfiguration in the PutObject request builder.

        PutObjectRequest request = PutObjectRequest.builder()
                .bucket(BUCKET)
                .key(KEY)
                .overrideConfiguration(AwsRequestOverrideConfiguration.builder()
                        .putHeader("x-amz-website-redirect-location", URL)
                        .build())
                .build();

[debora-ito]

On a side note, try using the regional s3 endpoints whenever possible, instead of using the global s3.amazonaws.com endpoint. It is faster and you'll avoid all these region related errors.

[boaks]

Thanks for the response!

On a side note, try using the regional s3 endpoints whenever possible, instead of using the global s3.amazonaws.com endpoint. It is faster and you'll avoid all these region related errors.

That's somehow more a question of the tools and APIs.

E.g. the s3cmd, unfortunately, the configuration file is not that completely documented, see About the s3cmd configuration file.

That shows,

host_base = s3.amazonaws.com
host_bucket = %(bucket)s.s3.amazonaws.com 

But it's not too easy to see, if

host_base = s3.%(bucket_location)s.amazonaws.com
host_bucket = %(bucket)s.%(host_base)s 

or something similar works as well.

[debora-ito]

s3cmd is not a tool maintained by aws, so I can't answer for its documentation or help with questions about its usage.

[boaks]

s3cmd is not a tool maintained by aws

Sure, I just want to show, why following the recommendation " using the regional s3 endpoints whenever possible" isn't that easy. My question is answered, thanks a lot!

[github-actions[bot]]

Hello! Reopening this discussion to make it searchable.