HDDS-10412. Prefix ACL check needs to resolve the bucket link

[ivandika3]

What changes were proposed in this pull request?

Following up on HDDS-4715, we also need to resolve the bucket link for prefix ACL.

Here are the changes included in this patch:

• Include prefix resource type to be BucketManagerImpl so that prefix access under a bucket link will be resolved to the source bucket
• Prefix ACL read: PrefixManagerImpl#getAcl will resolve the bucket to get the ACL
• Prefix ACL write: OMPrefixAclRequest#validateAndUpdateCache will resolve the bucket before writing the ACL
  • This requires passing the resolved prefix OzoneObj to the request hooks (onComplete, apply) so that the request implementation can use the resolved object and not the link object
• Fixed PrefixManagerImpl#setAcl so that default ACL is inherited only when the prefix does not exist
  • This was found in the integration test
• Some comments and cosmetics improvements

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-10412

How was this patch tested?

Acceptance and integration tests.

Clean CI run: https://github.com/ivandika3/ozone/actions/runs/8029229022

[ivandika3]

@whbing @ChenSammi Could you help take a look when you have time?

[adoroszlai]

Thanks @ivandika3 for the patch, LGTM.

[whbing]

String prefix or String prefixPath maybe only need one?

[ivandika3]

Thank you for the review. I removed all the unnecessary variables (volume, bucket, and prefix)

[whbing]

Is null condition necessary here ? I see a call to resolvedPrefixObj.getpath () on line 81, 82.

[ivandika3]

If IOException is thrown during the getResolvedPrefixObj (i.e. during bucket resolving process), the resolvedPrefixObj should be null and we should fall back to the initial ozone object from getOzoneObj.

[whbing]

@ivandika3 Thanks ! LGTM.

[adoroszlai]

@ChenSammi would you like to take a look?

[adoroszlai]

Thanks @ivandika3 for the patch, @whbing for the review.

[ivandika3]

Thank you @adoroszlai @whbing for the reviews.