KHB: Documentation about S3 Storage "buggy" - Bucket names must be unique

Feedback & Ideas
,
1

Hi,

while dwelving in the docs I noticed a possible documentation bug that can have severe issues depending whenever or not the bucket-names are hard coded.

S3 Storage

Amazon S3 Buckets

KNIME Business Hub can use four dedicated AWS S3 buckets for different data types:

Bucket Name Purpose Access Level
catalog-service Workflows and metadata Private
avatars User profile pictures Public read
execution-jobs Job execution data and results Private
customization-profiles User customization settings Private

An S3 bucket name must always be unique. There cannot be two or more KHB installations, even in different accounts and regions, that try to create the same bucket name:

image
image1007×197 24.7 KB

What happens if I delete a bucket and want to reuse the name?

If you delete an Amazon S3 bucket and want to create a new bucket with the same name, you must wait for the name to become available again. It can take 48–72 hours for the changes to take effect. It’s a best practice to wait at least 48 hours before creating a new bucket that reuses a previous name. Amazon S3 bucket names are globally unique, and all AWS accounts share the same namespace.

Can anyone check that please? If the bucket names are hardcoded, that might be a solvable but still concerning issue.

PS: Checking the deplyoment settings of the recent KHB I found the section which allows to configure the bucket names. That still raises the question if the documentation is not quite accurate.

image
image789×1105 56.7 KB

Best
Mike

2

Thanks for raising this, Mike. Great catch!

The bucket names in the KNIME Business Hub AWS setup are not hard-coded but fully customizable. By configuring the bucket names in the Replicated portal or ArgoCD configuration, you can use whatever names you like for your buckets. This means you won’t run into global S3 name conflicts as long as you choose unique names during setup.

We’ll update the documentation to make this clearer — specifically that:

  • the default bucket names shown in the docs are just examples, and

  • users must customize them to avoid global uniqueness constraints imposed by AWS S3.

Thanks again for pointing this out and helping improve our documentation!

2 Likes
3

For internal reference, this is tracked in:
DOCS-917