This post has NOT been accepted by the mailing list yet.
I have a small program in SparkSQL that in my local machine with Mac OS X uses 100% of 4 cpus (when using --master local[*] option). But when I execute the same code with the same option on a virtual machine based on Debian Jessie, it only uses 1 cpu. Looks like it is unable to assign more than 1 cpu, no matter what --master option you send.
I'm using spark-2.1.0-bin-hadoop2.7 with java 8 on both cases.