[SPARK-4672][Core]Checkpoint() should clear f to shorten the serialization chain

The related JIRA is https://issues.apache.org/jira/browse/SPARK-4672

The f closure of `PartitionsRDD(ZippedPartitionsRDD2)` contains a `$outer` that references EdgeRDD/VertexRDD, which causes task's serialization chain become very long in iterative GraphX applications. As a result, StackOverflow error will occur. If we set "f = null" in `clearDependencies()`, checkpoint() can cut off the long serialization chain. More details and explanation can be found in the JIRA.

Author: JerryLead <[email protected]>
Author: Lijie Xu <[email protected]>

Closes apache#3545 from JerryLead/my_core and squashes the following commits:

f7faea5 [JerryLead] checkpoint() should clear the f to avoid StackOverflow error
c0169da [JerryLead] Merge branch 'master' of https://github.com/apache/spark
52799e3 [Lijie Xu] Merge pull request #1 from apache/master

Author: JerryLead

core/src/main/scala/org/apache/spark/rdd/ZippedPartitionsRDD.scala

@@ -77,7 +77,7 @@ private[spark] abstract class ZippedPartitionsBaseRDD[V: ClassTag](
 
 private[spark] class ZippedPartitionsRDD2[A: ClassTag, B: ClassTag, V: ClassTag](
     sc: SparkContext,
-    f: (Iterator[A], Iterator[B]) => Iterator[V],
+    var f: (Iterator[A], Iterator[B]) => Iterator[V],
     var rdd1: RDD[A],
     var rdd2: RDD[B],
     preservesPartitioning: Boolean = false)
@@ -92,13 +92,14 @@ private[spark] class ZippedPartitionsRDD2[A: ClassTag, B: ClassTag, V: ClassTag]
     super.clearDependencies()
     rdd1 = null
     rdd2 = null
+    f = null
   }
 }
 
 private[spark] class ZippedPartitionsRDD3
   [A: ClassTag, B: ClassTag, C: ClassTag, V: ClassTag](
     sc: SparkContext,
-    f: (Iterator[A], Iterator[B], Iterator[C]) => Iterator[V],
+    var f: (Iterator[A], Iterator[B], Iterator[C]) => Iterator[V],
     var rdd1: RDD[A],
     var rdd2: RDD[B],
     var rdd3: RDD[C],
@@ -117,13 +118,14 @@ private[spark] class ZippedPartitionsRDD3
     rdd1 = null
     rdd2 = null
     rdd3 = null
+    f = null
   }
 }
 
 private[spark] class ZippedPartitionsRDD4
   [A: ClassTag, B: ClassTag, C: ClassTag, D:ClassTag, V: ClassTag](
     sc: SparkContext,
-    f: (Iterator[A], Iterator[B], Iterator[C], Iterator[D]) => Iterator[V],
+    var f: (Iterator[A], Iterator[B], Iterator[C], Iterator[D]) => Iterator[V],
     var rdd1: RDD[A],
     var rdd2: RDD[B],
     var rdd3: RDD[C],
@@ -145,5 +147,6 @@ private[spark] class ZippedPartitionsRDD4
     rdd2 = null
     rdd3 = null
     rdd4 = null
+    f = null
   }
 }