When saving an RDD to disk, if the spark node doing the saving crashes, is the entire file rewritten to disk or can spark figure out where it left off upon RDD recomputation?
On Thu, Feb 20, 2014 at 10:41 AM, Adrian Mocanu <[hidden email]> wrote:
Hi
When saving an RDD to disk, if the spark node doing the saving crashes, is the entire file rewritten to disk or can spark figure out where it left off upon RDD recomputation?