On a cluster where the nodes and the master all have access to a shared filesystem/files - does spark read a file (like one resulting from sc.textFile()) in parallel/different sections on each node? Or is the file read on master in sequence and chunks processed on the nodes afterwards?
Thanks! Ognen