From b1c6fa1584099b3a1e0615c100f10ea90b1ad2c9 Mon Sep 17 00:00:00 2001
From: Patrick Wendell <pwendell@gmail.com>
Date: Wed, 4 Dec 2013 18:39:34 -0800
Subject: Document missing configs and set shuffle consolidation to false.

---
 .../apache/spark/storage/ShuffleBlockManager.scala |  2 +-
 docs/configuration.md                              | 37 +++++++++++++++++++++-
 2 files changed, 37 insertions(+), 2 deletions(-)

diff --git a/core/src/main/scala/org/apache/spark/storage/ShuffleBlockManager.scala b/core/src/main/scala/org/apache/spark/storage/ShuffleBlockManager.scala
index 2f1b049ce4..e828e1d1c5 100644
--- a/core/src/main/scala/org/apache/spark/storage/ShuffleBlockManager.scala
+++ b/core/src/main/scala/org/apache/spark/storage/ShuffleBlockManager.scala
@@ -62,7 +62,7 @@ class ShuffleBlockManager(blockManager: BlockManager) {
   // Turning off shuffle file consolidation causes all shuffle Blocks to get their own file.
   // TODO: Remove this once the shuffle file consolidation feature is stable.
   val consolidateShuffleFiles =
-    System.getProperty("spark.shuffle.consolidateFiles", "true").toBoolean
+    System.getProperty("spark.shuffle.consolidateFiles", "false").toBoolean
 
   private val bufferSize = System.getProperty("spark.shuffle.file.buffer.kb", "100").toInt * 1024
 
diff --git a/docs/configuration.md b/docs/configuration.md
index 97183bafdb..1a3eef345c 100644
--- a/docs/configuration.md
+++ b/docs/configuration.md
@@ -327,7 +327,42 @@ Apart from these, the following properties are also available, and may be useful
     Too large a value decreases parallelism during broadcast (makes it slower); however, if it is too small, <code>BlockManager</code> might take a performance hit.
   </td>
 </tr>
-
+<tr>
+  <td>spark.shuffle.consolidateFiles</td>
+  <td>false</td>
+  <td>
+    If set to "true", consolidates intermediate files created during a shuffle.
+  </td>
+</tr>
+<tr>
+<tr>
+  <td>spark.speculation</td>
+  <td>false</td>
+  <td>
+    If set to "true", performs speculative execution of tasks. This means if one or more tasks are running slowly in a stage, they will be re-launched.
+  </td>
+</tr>
+<tr>
+  <td>spark.speculation.interval</td>
+  <td>100</td>
+  <td>
+    How often Spark will check for tasks to speculate, in seconds.
+  </td>
+</tr>
+<tr>
+  <td>spark.speculation.quantile</td>
+  <td>0.75</td>
+  <td>
+    Percentage of tasks which must be complete before speculation is enabled for a particular stage.
+  </td>
+</tr>
+<tr>
+  <td>spark.speculation.multiplier</td>
+  <td>1.5</td>
+  <td>
+    How many times slower a task is than the median to be considered for speculation.
+  </td>
+</tr>
 </table>
 
 # Environment Variables
-- 
cgit v1.2.3