I have a very simple question for Spark configuration, Does in local mode, the number of executors is 1, as there are just one working node?
Another question I have is the difference between deploy modes as client or cluster mode? The documentation mention in the client model, the driver initiated outside the cluster, I can not get this statement..
If I remember correctly, you get one executor in local mode, if you do not specify otherwise.
The difference between client and cluster mode is primarily where your driver runs. In client mode it runs on the machine where you deploy the job. In cluster mode, the driver will run on some node in the cluster. In client mode you thus have easy acces to whatever the driver outputs to stdout and stderr. Go with client mode if you are in doubt.
