vigneshio opened a new issue, #4657:
URL: https://github.com/apache/polaris/issues/4657

   When Polaris handles an external table notification (CREATE/UPDATE), 
creating or updating the table entity can fail under a concurrent modification, 
and there is currently no retry. 
   
   updateTableLike fails with a concurrent-modification conflict, and 
createTableLike fails when the entity was created concurrently, so the 
notification just errors out instead of retrying.
   
   A naive retry of the same write would not help, because the entity is read 
once from the request-scoped resolution view and its version becomes stale. A 
correct fix needs to re-read the current entity on each attempt, re-apply the 
new metadata location and timestamp, and retry with a bound.
   
   I wanted to discuss the approach before raising a PR: would a retry scoped 
to the update path be acceptable as a first step, with the create-then-update 
race handled as a follow-up? And is there an existing retry helper you would 
prefer to reuse?
   
   Location: 
runtime/service/src/main/java/org/apache/polaris/service/catalog/iceberg/IcebergCatalog.java
 (sendNotificationForTableLike).


-- 
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