diff --git a/python/pyspark/tests.py b/python/pyspark/tests.py
index b687d695b01c4..8ad01bf3fb1d0 100644
--- a/python/pyspark/tests.py
+++ b/python/pyspark/tests.py
@@ -955,8 +955,6 @@ def test_newhadoop(self):
             conf=input_conf).collect())
         self.assertEqual(new_dataset, data)
 
-    @unittest.skipIf(sys.version_info[:2] <= (2, 6) or python_implementation() == "PyPy",
-                     "Skipped on 2.6 and PyPy until SPARK-2951 is fixed")
     def test_newhadoop_with_array(self):
         basepath = self.tempdir.name
         # use custom ArrayWritable types and converters to handle arrays