Skip to content

[SPARK-33277][PYSPARK][SQL] Use ContextAwareIterator to stop consuming after the task ends. #30899

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
ueshin wants to merge 3 commits into from
Closed

[SPARK-33277][PYSPARK][SQL] Use ContextAwareIterator to stop consuming after the task ends. #30899

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
03c7aac
Use ContextAwareIterator to stop consuming after the task ends.
ueshin Dec 23, 2020
3606a4c
Move ContextAwareIterator to org.apache.spark.util.
ueshin Dec 23, 2020
36176f7
Move to org.apache.spark and make it follow the way of InterruptibleI…
ueshin Dec 23, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 17 additions & 1 deletion sql/core/src/main/scala/org/apache/spark/sql/execution/python/EvalPythonExec.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -89,6 +89,7 @@ trait EvalPythonExec extends UnaryExecNode {

inputRDD.mapPartitions { iter =>
val context = TaskContext.get()
val contextAwareIterator = new ContextAwareIterator(iter, context)

// The queue used to buffer input rows so we can drain it to
// combine input with output from Python.
Expand Down Expand Up @@ -120,7 +121,7 @@ trait EvalPythonExec extends UnaryExecNode {
}.toSeq)

// Add rows to queue to join later with the result.
val projectedRowIter = iter.map { inputRow =>
val projectedRowIter = contextAwareIterator.map { inputRow =>
queue.add(inputRow.asInstanceOf[UnsafeRow])
projection(inputRow)
}
Expand All @@ -137,3 +138,18 @@ trait EvalPythonExec extends UnaryExecNode {
}
}
}

/**
* A TaskContext aware iterator.
*
* As the Python evaluation consumes the parent iterator in a separate thread,
* it could consume more data from the parent even after the task ends and the parent is closed.
Copy link
Member

@viirya viirya Dec 23, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe add "If an off-heap column vector exists in the parent iterator, it could cause segmentation fault which crashes the executor." here too?

Current phrase is not clear why it is bad to read closed parent.

Copy link
Member Author

@ueshin ueshin Dec 23, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I added the sentence. Thanks!

* Thus, we should use ContextAwareIterator to stop consuming after the task ends.
*/
class ContextAwareIterator[IN](iter: Iterator[IN], context: TaskContext) extends Iterator[IN] {
Copy link
Member

@dongjoon-hyun dongjoon-hyun Dec 23, 2020 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This looks like a general class. Can we put this into more general package instead of org.apache.sql.execution.python package as a separate file?

Copy link
Member Author

@ueshin ueshin Dec 23, 2020 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I moved the class to org.apache.spark.util package. cc @gatorsmile


override def hasNext: Boolean =
!context.isCompleted() && !context.isInterrupted() && iter.hasNext

override def next(): IN = iter.next()
}
7 changes: 4 additions & 3 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/python/MapInPandasExec.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -61,16 +61,17 @@ case class MapInPandasExec(
val pythonRunnerConf = ArrowUtils.getPythonRunnerConfMap(conf)
val outputTypes = child.schema

val context = TaskContext.get()
val contextAwareIterator = new ContextAwareIterator(inputIter, context)

// Here we wrap it via another row so that Python sides understand it
// as a DataFrame.
val wrappedIter = inputIter.map(InternalRow(_))
val wrappedIter = contextAwareIterator.map(InternalRow(_))

// DO NOT use iter.grouped(). See BatchIterator.
val batchIter =
if (batchSize > 0) new BatchIterator(wrappedIter, batchSize) else Iterator(wrappedIter)

val context = TaskContext.get()

val columnarBatchIter = new ArrowPythonRunner(
chainedFunc,
PythonEvalType.SQL_MAP_PANDAS_ITER_UDF,
Expand Down