The master is what monitors and manages jobs. It can be a master in a Spark standalone scheduler (which itself might be running inside of YARN) or a Mesos master.
The driver is the program that the user interfaces with to talk to their Spark job. That communication is primarily dive through a SparkContext object. Have you read through http://spark.apache.org/docs/latest/index.html yet? If not, it and the other docs might be helpful. In the future, this level of question would probably be better suited for the user list. Best, Andy On Mar 31, 2014 7:48 AM, "Dan" <zsh912...@gmail.com> wrote: Hi, I've been recently reading spark code and confused about driver and master. What's the difference between them? When I run spark in standalone cluster, from the log it seems that the driver has not been launched. Thanks, Dan -- View this message in context: http://apache-spark-developers-list.1001551.n3.nabble.com/The-difference-between-driver-and-master-in-Spark-tp6158.html Sent from the Apache Spark Developers List mailing list archive at Nabble.com.
