Update checkpointing API docs in Python/Java.

2 files changed, 12 insertions, 16 deletions

diff --git a/python/pyspark/context.py b/python/pyspark/context.py
index 8beb8e2ae9..dcbed37270 100644
--- a/python/pyspark/context.py
+++ b/python/pyspark/context.py
@@ -202,9 +202,12 @@ class SparkContext(object):
 
     def setCheckpointDir(self, dirName, useExisting=False):
         """
-        Set the directory under which RDDs are going to be checkpointed. This
-        method will create this directory and will throw an exception of the
-        path already exists (to avoid overwriting existing files may be
-        overwritten). The directory will be deleted on exit if indicated.
+        Set the directory under which RDDs are going to be checkpointed. The
+        directory must be a HDFS path if running on a cluster.
+
+        If the directory does not exist, it will be created. If the directory
+        exists and C{useExisting} is set to true, then the exisiting directory
+        will be used.  Otherwise an exception will be thrown to prevent
+        accidental overriding of checkpoint files in the existing directory.
         """
         self._jsc.sc().setCheckpointDir(dirName, useExisting)
diff --git a/python/pyspark/rdd.py b/python/pyspark/rdd.py
index 2a2ff9b271..7b6ab956ee 100644
--- a/python/pyspark/rdd.py
+++ b/python/pyspark/rdd.py
@@ -52,18 +52,11 @@ class RDD(object):
 
     def checkpoint(self):
         """
-        Mark this RDD for checkpointing. The RDD will be saved to a file inside
-        `checkpointDir` (set using setCheckpointDir()) and all references to
-        its parent RDDs will be removed.  This is used to truncate very long
-        lineages.  In the current implementation, Spark will save this RDD to
-        a file (using saveAsObjectFile()) after the first job using this RDD is
-        done.  Hence, it is strongly recommended to use checkpoint() on RDDs
-        when
-
-        (i) checkpoint() is called before the any job has been executed on this
-        RDD.
-
-        (ii) This RDD has been made to persist in memory. Otherwise saving it
+        Mark this RDD for checkpointing. It will be saved to a file inside the
+        checkpoint directory set with L{SparkContext.setCheckpointDir()} and
+        all references to its parent RDDs will be removed. This function must
+        be called before any job has been executed on this RDD. It is strongly
+        recommended that this RDD is persisted in memory, otherwise saving it
         on a file will require recomputation.
         """
         self.is_checkpointed = True