[SPARK-4761][SQL] Enables Kryo by default in Spark SQL Thrift server

Enables Kryo and disables reference tracking by default in Spark SQL Thrift server. Configurations explicitly defined by users in `spark-defaults.conf` are respected (the Thrift server is started by `spark-submit`, which handles configuration properties properly).

<!-- Reviewable:start -->
[<img src="https://reviewable.io/review_button.png" height=40 alt="Review on Reviewable"/>](https://reviewable.io/reviews/apache/spark/3621)
<!-- Reviewable:end -->

Author: Cheng Lian <lian@databricks.com>

Closes #3621 from liancheng/kryo-by-default and squashes the following commits:

70c2775 [Cheng Lian] Enables Kryo by default in Spark SQL Thrift server

(cherry picked from commit 6f61e1f)
Signed-off-by: Patrick Wendell <pwendell@gmail.com>

sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLEnv.scala

@@ -32,11 +32,21 @@ private[hive] object SparkSQLEnv extends Logging {
 
   def init() {
     if (hiveContext == null) {
-      val sparkConf = new SparkConf()
+      val sparkConf = new SparkConf(loadDefaults = true)
+      val maybeSerializer = sparkConf.getOption("spark.serializer")
+      val maybeKryoReferenceTracking = sparkConf.getOption("spark.kryo.referenceTracking")
+
+      sparkConf
         .setAppName(s"SparkSQL::${java.net.InetAddress.getLocalHost.getHostName}")
         .set("spark.sql.hive.version", HiveShim.version)
-      sparkContext = new SparkContext(sparkConf)
+        .set(
+          "spark.serializer",
+          maybeSerializer.getOrElse("org.apache.spark.serializer.KryoSerializer"))
+        .set(
+          "spark.kryo.referenceTracking",
+          maybeKryoReferenceTracking.getOrElse("false"))
 
+      sparkContext = new SparkContext(sparkConf)
       sparkContext.addSparkListener(new StatsReportListener())
       hiveContext = new HiveContext(sparkContext)