PySpark tests are flaky

[dcrankshaw]

Our PySpark tests are really flaky due to a Spark bug. I think it's due to SPARK-18406 which was patched in 2.1.1 while our tests currently run on Spark 2.1.1.

They fail with the following stack trace (example:

17/09/01 12:42:30 ERROR Executor: Exception in task 0.0 in stage 18.0 (TID 30)
java.lang.AssertionError: assertion failed
	at scala.Predef$.assert(Predef.scala:156)
	at org.apache.spark.storage.BlockInfo.checkInvariants(BlockInfoManager.scala:84)
	at org.apache.spark.storage.BlockInfo.readerCount_$eq(BlockInfoManager.scala:66)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:362)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:361)
	at scala.Option.foreach(Option.scala:257)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:361)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:356)
	at scala.collection.Iterator$class.foreach(Iterator.scala:893)
	at scala.collection.AbstractIterator.foreach(Iterator.scala:1336)
	at org.apache.spark.storage.BlockInfoManager.releaseAllLocksForTask(BlockInfoManager.scala:356)
	at org.apache.spark.storage.BlockManager.releaseAllLocksForTask(BlockManager.scala:676)
	at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:329)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	at java.lang.Thread.run(Thread.java:748)
17/09/01 12:42:31 WARN TaskSetManager: Lost task 0.0 in stage 18.0 (TID 30, localhost, executor driver): java.lang.AssertionError: assertion failed
	at scala.Predef$.assert(Predef.scala:156)
	at org.apache.spark.storage.BlockInfo.checkInvariants(BlockInfoManager.scala:84)
	at org.apache.spark.storage.BlockInfo.readerCount_$eq(BlockInfoManager.scala:66)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:362)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:361)
	at scala.Option.foreach(Option.scala:257)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:361)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:356)
	at scala.collection.Iterator$class.foreach(Iterator.scala:893)
	at scala.collection.AbstractIterator.foreach(Iterator.scala:1336)
	at org.apache.spark.storage.BlockInfoManager.releaseAllLocksForTask(BlockInfoManager.scala:356)
	at org.apache.spark.storage.BlockManager.releaseAllLocksForTask(BlockManager.scala:676)
	at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:329)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	at java.lang.Thread.run(Thread.java:748)

17/09/01 12:42:31 ERROR TaskSetManager: Task 0 in stage 18.0 failed 1 times; aborting job
17-09-01:12:42:31 ERROR    [deploy_pyspark_models.py:205] Exception
Traceback (most recent call last):
  File "../integration-tests/deploy_pyspark_models.py", line 175, in <module>
    svm_model = train_svm(trainRDD)
  File "../integration-tests/deploy_pyspark_models.py", line 120, in train_svm
    return SVMWithSGD.train(trainRDD)
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/mllib/classification.py", line 553, in train
    return _regression_train_wrapper(train, SVMModel, data, initialWeights)
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/mllib/regression.py", line 208, in _regression_train_wrapper
    first = data.first()
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/rdd.py", line 1360, in first
    rs = self.take(1)
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/rdd.py", line 1342, in take
    res = self.context.runJob(self, takeUpToNumLeft, p)
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/context.py", line 968, in runJob
    port = self._jvm.PythonRDD.runJob(self._jsc.sc(), mappedRDD._jrdd, partitions)
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/lib/py4j-0.10.4-src.zip/py4j/java_gateway.py", line 1133, in __call__
    answer, self.gateway_client, self.target_id, self.name)
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/pyspark/sql/utils.py", line 63, in deco
    return f(*a, **kw)
  File "/spark/spark-2.1.1-bin-hadoop2.7/python/lib/py4j-0.10.4-src.zip/py4j/protocol.py", line 319, in get_return_value
    format(target_id, ".", name), value)
Py4JJavaError: An error occurred while calling z:org.apache.spark.api.python.PythonRDD.runJob.
: org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 18.0 failed 1 times, most recent failure: Lost task 0.0 in stage 18.0 (TID 30, localhost, executor driver): java.lang.AssertionError: assertion failed
	at scala.Predef$.assert(Predef.scala:156)
	at org.apache.spark.storage.BlockInfo.checkInvariants(BlockInfoManager.scala:84)
	at org.apache.spark.storage.BlockInfo.readerCount_$eq(BlockInfoManager.scala:66)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:362)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:361)
	at scala.Option.foreach(Option.scala:257)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:361)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:356)
	at scala.collection.Iterator$class.foreach(Iterator.scala:893)
	at scala.collection.AbstractIterator.foreach(Iterator.scala:1336)
	at org.apache.spark.storage.BlockInfoManager.releaseAllLocksForTask(BlockInfoManager.scala:356)
	at org.apache.spark.storage.BlockManager.releaseAllLocksForTask(BlockManager.scala:676)
	at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:329)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	at java.lang.Thread.run(Thread.java:748)

Driver stacktrace:
	at org.apache.spark.scheduler.DAGScheduler.org$apache$spark$scheduler$DAGScheduler$$failJobAndIndependentStages(DAGScheduler.scala:1435)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1423)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1422)
	at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59)
	at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:48)
	at org.apache.spark.scheduler.DAGScheduler.abortStage(DAGScheduler.scala:1422)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:802)
	at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:802)
	at scala.Option.foreach(Option.scala:257)
	at org.apache.spark.scheduler.DAGScheduler.handleTaskSetFailed(DAGScheduler.scala:802)
	at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.doOnReceive(DAGScheduler.scala:1650)
	at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1605)
	at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1594)
	at org.apache.spark.util.EventLoop$$anon$1.run(EventLoop.scala:48)
	at org.apache.spark.scheduler.DAGScheduler.runJob(DAGScheduler.scala:628)
	at org.apache.spark.SparkContext.runJob(SparkContext.scala:1925)
	at org.apache.spark.SparkContext.runJob(SparkContext.scala:1938)
	at org.apache.spark.SparkContext.runJob(SparkContext.scala:1951)
	at org.apache.spark.api.python.PythonRDD$.runJob(PythonRDD.scala:441)
	at org.apache.spark.api.python.PythonRDD.runJob(PythonRDD.scala)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.lang.reflect.Method.invoke(Method.java:498)
	at py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)
	at py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)
	at py4j.Gateway.invoke(Gateway.java:280)
	at py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
	at py4j.commands.CallCommand.execute(CallCommand.java:79)
	at py4j.GatewayConnection.run(GatewayConnection.java:214)
	at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.AssertionError: assertion failed
	at scala.Predef$.assert(Predef.scala:156)
	at org.apache.spark.storage.BlockInfo.checkInvariants(BlockInfoManager.scala:84)
	at org.apache.spark.storage.BlockInfo.readerCount_$eq(BlockInfoManager.scala:66)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:362)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2$$anonfun$apply$2.apply(BlockInfoManager.scala:361)
	at scala.Option.foreach(Option.scala:257)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:361)
	at org.apache.spark.storage.BlockInfoManager$$anonfun$releaseAllLocksForTask$2.apply(BlockInfoManager.scala:356)
	at scala.collection.Iterator$class.foreach(Iterator.scala:893)
	at scala.collection.AbstractIterator.foreach(Iterator.scala:1336)
	at org.apache.spark.storage.BlockInfoManager.releaseAllLocksForTask(BlockInfoManager.scala:356)
	at org.apache.spark.storage.BlockManager.releaseAllLocksForTask(BlockManager.scala:676)
	at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:329)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	... 1 more