-
Story
-
Resolution: Done
-
Normal
-
None
-
None
-
False
-
False
-
Passed
-
Undefined
-
Reference slack thread: https://coreos.slack.com/archives/CFN0XJEPR/p1624868823435300
Customer request came in for more information how to configure a generic S3 bucket as their replication repository (which was Netapp), referencing this section of the product documentation:
MCG is going to be the same as a generic S3 bucket as far as MTC is concerned, but the documentation doesn't do a great job describing how exactly you configure it within the MTC web console.
For generic S3 (effectively anything that is non-AWS S3), we expect the following:
Bucket Endpoint: probably in the form of "https://<my_bucket_endpoint", note the protocol (https://,http://, is important)
Access key, secret access key, both of which should be fairly obvious
Region: This is the tricky one, some backends require this, and some backends do not. We cannot tell the user what value to provide because we don't know the specific backend they're going to be using. The first recommendation should be to confirm with your s3 product documentation as to what the expected client value should be. Then we can mention that some/most s3 providers will accept an empty region. For example, Netapp accepts region as you'd expect, but if you leave it empty it will default to us-east-1.
- links to
ernelson@redhat.com
The PR has not yet been merged because of the docs freeze, so I updated it to include an "Additional resources" section in the section about configuring the repository. The Additional resources list has links to MTV workflow, "About data copy methods", and "Adding replication repository to web console'. I think that should provide sufficient context.