diff --git a/core/src/main/scala/org/apache/spark/SparkContext.scala b/core/src/main/scala/org/apache/spark/SparkContext.scala
index dd3157990ef2d..3d72f1d943915 100644
--- a/core/src/main/scala/org/apache/spark/SparkContext.scala
+++ b/core/src/main/scala/org/apache/spark/SparkContext.scala
@@ -239,6 +239,8 @@ class SparkContext(config: SparkConf) extends Logging {
       None
     }
 
+  ui.foreach(_.bind())
+
   /** A default Hadoop Configuration for the Hadoop code (e.g. file systems) that we reuse. */
   val hadoopConfiguration = SparkHadoopUtil.get.newConfiguration(conf)
 
@@ -341,10 +343,6 @@ class SparkContext(config: SparkConf) extends Logging {
   postEnvironmentUpdate()
   postApplicationStart()
 
-  // Bind the SparkUI after starting the task scheduler
-  // because certain pages and listeners depend on it
-  ui.foreach(_.bind())
-
   private[spark] var checkpointDir: Option[String] = None
 
   // Thread Local variable that can be used by users to pass information down the stack
diff --git a/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressPage.scala b/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressPage.scala
index a82f71ed08475..1e02f1225d344 100644
--- a/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressPage.scala
+++ b/core/src/main/scala/org/apache/spark/ui/jobs/JobProgressPage.scala
@@ -29,7 +29,7 @@ private[ui] class JobProgressPage(parent: JobProgressTab) extends WebUIPage("")
   private val live = parent.live
   private val sc = parent.sc
   private val listener = parent.listener
-  private lazy val isFairScheduler = parent.isFairScheduler
+  private def isFairScheduler = parent.isFairScheduler
 
   def render(request: HttpServletRequest): Seq[Node] = {
     listener.synchronized {