the overwrite will add new files, keep the old ones and in a log keeps track of what is current data and what is old data.

If the overwrite fails, you will get an error message in the spark program, and the data to be overwritten will still be the current state.