steveloughran opened a new pull request #2310: URL: https://github.com/apache/hadoop/pull/2310
Successor to #2280, which as well as handling the rename() failure problem does the same for delete _and_ for multipart failure handling. that failure handling is the most complex as the objects have been deleted; need to go through the list of objects submitted, remove those, determine which are directories, and for each one, look up on S3Guard to see if they have children, and add a tombstone on that record iff they were (a) deleted and (b) empty. This makes the cost of handling delete failures of directory markers slightly more expensive., but it should generally be considered a failure condition. Testing: s3a london; new tests added, -Dmarkers=keep. A new test is going to be needed to only create a dir tree and then delete that, with the role set up to fail to delete some of the path underneath ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
