roryqi commented on issue #10713:
URL: https://github.com/apache/gravitino/issues/10713#issuecomment-4212514614

   > I do agree, we should have some mechanism to show this, but it doesn't 
look very straightforward, just on a glance. Might be better to handle it 
separately. If we already use 206 for partial operation success then we can use 
it here, but I think it might be slightly misleading. Also, looking at the 
code, I think that might require some more changes, as we just return the 
object created so, if we created schema but setOwner failed, we simply return 
schema object back, so no way for the rest endpoint to know to send 200 or 206 
back. If we want to propagate this back, we would need to make changes in all 
operations to handle this.
   > 
   > As well as the client today treats 206 as a failure(
   > 
   > 
[gravitino/clients/client-java/src/main/java/org/apache/gravitino/client/HTTPClient.java](https://github.com/apache/gravitino/blob/d07a94f6b4b8222b7a45118b3459ea97de7450d2/clients/client-java/src/main/java/org/apache/gravitino/client/HTTPClient.java#L170)
   > 
   > Line 170 in 
[d07a94f](/apache/gravitino/commit/d07a94f6b4b8222b7a45118b3459ea97de7450d2)
   > 
   >  return code == HttpStatus.SC_OK 
   > ), not success, might have to change that as well. Also it depends on 
other clients as well, if they treat this as success or failure.
   > I feel like we should simply give a success response as the actual 
operation was done, but internal gravitino set owner call failed. This can then 
be reset by the owner of metalake/catalog by themselves in a separate call.
   
   A simple success response will make users confused. It will occur that some 
operations can set the owner, some operations can't set the owner. It will be a 
random event from the user's view.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to