Skip to content
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

Jump to bottom

[SPARK-45814][CONNECT][SQL]Make ArrowConverters.createEmptyArrowBatch call close() to avoid memory leak #43691

Closed
wants to merge 4 commits into from
Closed

[SPARK-45814][CONNECT][SQL]Make ArrowConverters.createEmptyArrowBatch call close() to avoid memory leak #43691

Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
e16e996
call hasNext to avoid memory leak
Nov 7, 2023
8652a40
better close()
Nov 8, 2023
860c8a1
better close()
Nov 8, 2023
2574223
follow review
Nov 8, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
21 changes: 14 additions & 7 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/arrow/ArrowConverters.scala
View file
Open in desktop
Expand Up @@ -80,7 +80,7 @@ private[sql] object ArrowConverters extends Logging {
maxRecordsPerBatch: Long,
timeZoneId: String,
errorOnDuplicatedFieldNames: Boolean,
context: TaskContext) extends Iterator[Array[Byte]] {
context: TaskContext) extends Iterator[Array[Byte]] with AutoCloseable {

protected val arrowSchema =
ArrowUtils.toArrowSchema(schema, timeZoneId, errorOnDuplicatedFieldNames)
Expand All @@ -93,13 +93,11 @@ private[sql] object ArrowConverters extends Logging {
protected val arrowWriter = ArrowWriter.create(root)

Option(context).foreach {_.addTaskCompletionListener[Unit] { _ =>
root.close()
allocator.close()
close()
}}

override def hasNext: Boolean = rowIter.hasNext || {
root.close()
allocator.close()
close()
false
}

Expand All @@ -124,6 +122,11 @@ private[sql] object ArrowConverters extends Logging {

out.toByteArray
}

override def close(): Unit = {
root.close()
allocator.close()
}
}

private[sql] class ArrowBatchWithSchemaIterator(
Expand Down Expand Up @@ -226,11 +229,15 @@ private[sql] object ArrowConverters extends Logging {
schema: StructType,
timeZoneId: String,
errorOnDuplicatedFieldNames: Boolean): Array[Byte] = {
new ArrowBatchWithSchemaIterator(
val batches = new ArrowBatchWithSchemaIterator(
Iterator.empty, schema, 0L, 0L,
timeZoneId, errorOnDuplicatedFieldNames, TaskContext.get()) {
override def hasNext: Boolean = true
}.next()
}
val emptyBatch = batches.next()
Copy link
Contributor

Choose a reason for hiding this comment

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

Utils.tryWithSafeFinally {
  batches.next()
} {
  if (TaskContext.get() == null) { // If taskContext is not null, it can be successfully closed, right?
    batches.close()
  }
}

Copy link
Author

Choose a reason for hiding this comment

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

It's right

// SPARK-45814: We need to call `close()` to avoid memory leak.
batches.close()
emptyBatch
}

/**
Expand Down