1 Answers
The simple answer is no. In the scenario Ryan described there is one replication rule, so any objects added to the destination bucket would not be replicated to the source bucket. S3 replication has a single source and destination per rule which means replication goes in one direction (source to destination). You would need to setup another replication rule on the destination bucket to sync it with the source bucket. With the identical replication rules setup on each of the buckets in the course scenario you would have a 2-way sync between both buckets. Then any objects uploaded to either bucket would get replicated to the other bucket.
Yes if you enable replication in your "destination" bucket. It will not result in an infinite loop.