From 27c910f7f29087d1ac216d4933d641d6515fd6ad Mon Sep 17 00:00:00 2001
From: Xiangrui Meng <meng@databricks.com>
Date: Mon, 25 Jan 2016 22:53:34 -0800
Subject: [PATCH] [SPARK-10086][MLLIB][STREAMING][PYSPARK] ignore
StreamingKMeans test in PySpark for now
I saw several failures from recent PR builds, e.g., https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/50015/consoleFull. This PR marks the test as ignored and we will fix the flakyness in SPARK-10086.
gliptak Do you know why the test failure didn't show up in the Jenkins "Test Result"?
cc: jkbradley
Author: Xiangrui Meng <meng@databricks.com>
Closes #10909 from mengxr/SPARK-10086.
---
python/pyspark/mllib/tests.py | 1 +
1 file changed, 1 insertion(+)
diff --git a/python/pyspark/mllib/tests.py b/python/pyspark/mllib/tests.py
index 79ce4959c9..25a7c29982 100644
--- a/python/pyspark/mllib/tests.py
+++ b/python/pyspark/mllib/tests.py
@@ -1189,6 +1189,7 @@ class StreamingKMeansTest(MLLibStreamingTestCase):
self._eventually(condition, catch_assertions=True)
+ @unittest.skip("SPARK-10086: Flaky StreamingKMeans test in PySpark")
def test_trainOn_predictOn(self):
"""Test that prediction happens on the updated model."""
stkm = StreamingKMeans(decayFactor=0.0, k=2)
--
GitLab