diff --git a/python-package/xgboost/spark/core.py b/python-package/xgboost/spark/core.py
index ffeeae8a7..03d431dd4 100644
--- a/python-package/xgboost/spark/core.py
+++ b/python-package/xgboost/spark/core.py
@@ -88,6 +88,7 @@ _pyspark_specific_params = [
     "features_cols",
     "enable_sparse_data_optim",
     "qid_col",
+    "repartition_random_shuffle",
 ]
 
 _non_booster_params = ["missing", "n_estimators", "feature_types", "feature_weights"]
@@ -477,7 +478,7 @@ class _SparkXGBEstimator(Estimator, _SparkXGBParams, MLReadable, MLWritable):
             num_workers=1,
             use_gpu=False,
             force_repartition=False,
-            repartition_random_shuffle=True,
+            repartition_random_shuffle=False,
             feature_names=None,
             feature_types=None,
             arbitrary_params_dict={},
diff --git a/python-package/xgboost/spark/data.py b/python-package/xgboost/spark/data.py
index a5b3b1e81..11dc02340 100644
--- a/python-package/xgboost/spark/data.py
+++ b/python-package/xgboost/spark/data.py
@@ -9,6 +9,8 @@ from xgboost.compat import concat
 
 from xgboost import DataIter, DeviceQuantileDMatrix, DMatrix
 
+from .utils import get_logger  # type: ignore
+
 
 def stack_series(series: pd.Series) -> np.ndarray:
     """Stack a series of arrays."""
@@ -246,6 +248,11 @@ def create_dmatrix_from_partitions(  # pylint: disable=too-many-arguments
         else:
             append_fn = append_m
         cache_partitions(iterator, append_fn)
+        if len(train_data) == 0:
+            get_logger("XGBoostPySpark").warning(
+                "Detected an empty partition in the training data. "
+                "Consider to enable repartition_random_shuffle"
+            )
         dtrain = make(train_data, kwargs)
     else:
         cache_partitions(iterator, append_dqm)