If I understand your question correctly, the answer is yes. You can retrieve rows of the rdd which are distributed across the nodes.
> On Dec 22, 2015, at 7:34 PM, Gaurav Agarwal <gaurav130...@gmail.com> wrote: > > If I have 3 more cluster and spark is running there .if I load the records > from phoenix to spark rdd and fetch the records from the spark through data > frame. > > Now I want to know that spark is distributed? > So I fetch the records from any of the node, records will be retrieved > present on any node present in spark rdd .