[SPARK-8282] [SPARKR] Make number of threads used in RBackend configurable

Read number of threads for RBackend from configuration.

[SPARK-8282] #comment Linking with JIRA

Author: Hossein <hossein@databricks.com>

Closes #6730 from falaki/SPARK-8282 and squashes the following commits:

33b3d98 [Hossein] Documented new config parameter
70f2a9c [Hossein] Fixing import
ec44225 [Hossein] Read number of threads for RBackend from configuration

core/src/main/scala/org/apache/spark/api/r/RBackend.scala

@@ -29,7 +29,7 @@ import io.netty.channel.socket.nio.NioServerSocketChannel
 import io.netty.handler.codec.LengthFieldBasedFrameDecoder
 import io.netty.handler.codec.bytes.{ByteArrayDecoder, ByteArrayEncoder}
 
-import org.apache.spark.Logging
+import org.apache.spark.{Logging, SparkConf}
 
 /**
  * Netty-based backend server that is used to communicate between R and Java.
@@ -41,7 +41,8 @@ private[spark] class RBackend {
   private[this] var bossGroup: EventLoopGroup = null
 
   def init(): Int = {
-    bossGroup = new NioEventLoopGroup(2)
+    val conf = new SparkConf()
+    bossGroup = new NioEventLoopGroup(conf.getInt("spark.r.numRBackendThreads", 2))
     val workerGroup = bossGroup
     val handler = new RBackendHandler(this)
 

docs/configuration.md

@@ -1495,6 +1495,18 @@ Apart from these, the following properties are also available, and may be useful
 </tr>
 </table>
 
+#### SparkR
+<table class="table">
+<tr><th>Property Name</th><th>Default</th><th>Meaning</th></tr>
+<tr>
+  <td><code>spark.r.numRBackendThreads</code></td>
+  <td>2</td>
+  <td>
+    Number of threads used by RBackend to handle RPC calls from SparkR package.
+  </td>
+</tr>
+</table>
+
 #### Cluster Managers
 Each cluster manager in Spark has additional configuration options. Configurations
 can be found on the pages for each mode: