0

I have a 3 node cluster managed by Cloudera Manager. I have edited BlockMatrix.scala file in the Spark MLLib source code and packaged it using:

mvn -DskipTests package

command. It has created new jar file. Now I want to run this newly created MLLib jar on my cluster. What is the way to do that?

pheeleeppoo
  • 1,491
  • 6
  • 25
  • 29
Chandan
  • 764
  • 2
  • 8
  • 21
  • 1
    Why change in the source code? It would be better to simply extend `BlockMatrix` and add an implicit conversion if you want to add functionality. – Shaido Nov 07 '17 at 06:57
  • Actually I wanted to check how much time it requires to complete each of the transformations (to check the computation and communication cost) involves in BlockMatrix multiply method. – Chandan Nov 07 '17 at 08:11

0 Answers0