S3 to Google Cloud Storage using direct transfer (web interface)

Question

File transfer service seems to be rejecting valid credentials from AWS? I have an s3 bucket:: >>> s3://my-bucket-name/icd_to_ccs.csv Have tried to set up the transfer service using the access key and secret

The key and secret work fine using boto3 and the aws command line for accessing the same file on my local machine Can't come up with the right foo to make it work using the cloud console transfer facility. I have tried with and without the Filter setting with and without the file name in the original "bucket" window.

It keeps giving me invalid access key? any guidance or tricks to this?

Have checked this answer and verified that my bucket has permissions for list the Amazon S3 bucket, get the location of the bucket, and read the objects in the bucket.

As mentioned the same key(and an alternate set of fresh keys) works fine via the boto3/cli on my local but no go on the web console..

Answer 1

I'm an engineer on the Transfer Service team.

Are you writing out "s3://my-bucket-name/icd_to_ccs.csv" within the web UI? Make sure that you enter this as "my-bucket-name" only. The "s3://" part should be filled out for you (and will cause issues), and the tool does not transfer on an item-by-item basis (so don't include "/icd_to_ccs.csv").

If you want to restrict a transfer to only one item, use "Specify file filters" (from "Select source" in step 1) to provide an include prefix.

If none of the above is actually the problem, then I'll need a Transfer ID. You can obtain that by clicking on your failed job; the URL will now be ".../transfer/details/?project...". I just need the number specified by 'ID number'. With that, I'll be able to do more detailed diagnosis. (EDIT: You should email that Transfer ID to gs-team@google.com, rather than posting it here.)

EDIT: The client doesn't need a region to be specified. AWS uses a global namespace, and any needed region information can be determined by looking at the bucket itself.