That makes it a lot easier. As you will need to reindex all the data there will be additional load and there is no way to avoid that. I would probebly try creating the new index and then redirecting new writes there. Once the old index is no longer written to you can start a reindexing job of this data to the new index. You will have duplicates for a period of time if both indices are included in queries, but this will go away once reindexing is done and the old index is deleted.