Skip to content

[SPARK-34726][SQL][2.4] Fix collectToPython timeouts #31818

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
peter-toth wants to merge 4 commits into from
Closed

[SPARK-34726][SQL][2.4] Fix collectToPython timeouts #31818

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
b4b425e
[SPARK-34726][SQL] Fix collectToPython timeouts
peter-toth Mar 12, 2021
6b18cc7
add UT
peter-toth Mar 12, 2021
aae1992
speed up UT
peter-toth Mar 12, 2021
8f6b811
address comments
peter-toth Mar 12, 2021
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
7 changes: 3 additions & 4 deletions sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3257,12 +3257,11 @@ class Dataset[T] private[sql](

private[sql] def collectToPython(): Array[Any] = {
EvaluatePython.registerPicklers()
withAction("collectToPython", queryExecution) { plan =>
val iter = withAction("collectToPython", queryExecution) { plan =>
val toJava: (Any) => Any = EvaluatePython.toJava(_, schema)
val iter: Iterator[Array[Byte]] = new SerDeUtil.AutoBatchedPickler(
plan.executeCollect().iterator.map(toJava))
PythonRDD.serveIterator(iter, "serve-DataFrame")
new SerDeUtil.AutoBatchedPickler(plan.executeCollect().iterator.map(toJava))
}
PythonRDD.serveIterator(iter, "serve-DataFrame")
}

private[sql] def getRowsToPython(
Expand Down
33 changes: 31 additions & 2 deletions sql/core/src/test/scala/org/apache/spark/sql/DatasetSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,19 +18,24 @@
package org.apache.spark.sql

import java.io.{Externalizable, ObjectInput, ObjectOutput}
import java.net.{InetAddress, Socket}
import java.sql.{Date, Timestamp}

import org.apache.spark.SparkException
import scala.io.Source

import org.apache.spark.{SparkConf, SparkException}
import org.apache.spark.security.SocketAuthHelper
import org.apache.spark.sql.catalyst.encoders.{OuterScopes, RowEncoder}
import org.apache.spark.sql.catalyst.plans.{LeftAnti, LeftSemi}
import org.apache.spark.sql.catalyst.util.sideBySide
import org.apache.spark.sql.execution.{LogicalRDD, RDDScanExec}
import org.apache.spark.sql.execution.{LogicalRDD, QueryExecution, RDDScanExec}
import org.apache.spark.sql.execution.exchange.{BroadcastExchangeExec, ShuffleExchangeExec}
import org.apache.spark.sql.execution.streaming.MemoryStream
import org.apache.spark.sql.functions._
import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.test.SharedSQLContext
import org.apache.spark.sql.types._
import org.apache.spark.sql.util.QueryExecutionListener

case class TestDataPoint(x: Int, y: Double, s: String, t: TestDataPoint2)
case class TestDataPoint2(x: Int, s: String)
Expand Down Expand Up @@ -1586,6 +1591,30 @@ class DatasetSuite extends QueryTest with SharedSQLContext {
}
}
}

test("SPARK-34726: Fix collectToPython timeouts") {
val listener = new QueryExecutionListener {
override def onFailure(funcName: String, qe: QueryExecution, exception: Exception): Unit = {}

override def onSuccess(funcName: String, qe: QueryExecution, duration: Long): Unit = {
// Longer than 15s in `PythonServer.setupOneConnectionServer`
Copy link
Contributor

@attilapiros attilapiros Mar 12, 2021

Choose a reason for hiding this comment

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

Let's extract the 15000 into a private[spark] var as member:

spark/core/src/main/scala/org/apache/spark/api/python/PythonRDD.scala

Line 899 in 6b18cc7

serverSocket.setSoTimeout(15000)

And add a comment above // visible for testing.
Then we can speed up this single test.

Copy link
Contributor Author

@peter-toth peter-toth Mar 12, 2021

Choose a reason for hiding this comment

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

Ok, added in 6b18cc7 and 8f6b811

Thread.sleep(20 * 1000)
Copy link
Contributor

@gaborgsomogyi gaborgsomogyi Mar 12, 2021

Choose a reason for hiding this comment

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

Just for my own understanding does this mean the test waits 15 seconds to pass?

Copy link
Contributor Author

@peter-toth peter-toth Mar 12, 2021

Choose a reason for hiding this comment

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

Yes, we need to wait a bit longer than the timeout. Without the fix in Dataset.collectToPython this UT fails.

}
}
spark.listenerManager.register(listener)

val Array(port: Int, secretToPython: String) = spark.range(5).toDF().collectToPython()

// Mimic Python side
val socket = new Socket(InetAddress.getByAddress(Array(127, 0, 0, 1)), port)
val authHelper = new SocketAuthHelper(new SparkConf()) {
override val secret: String = secretToPython
}
authHelper.authToServer(socket)
Source.fromInputStream(socket.getInputStream)

spark.listenerManager.unregister(listener)
Copy link
Contributor

@gaborgsomogyi gaborgsomogyi Mar 12, 2021

Choose a reason for hiding this comment

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

I would put this into finally to be on the safe side.

Copy link
Contributor Author

@peter-toth peter-toth Mar 12, 2021

Choose a reason for hiding this comment

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

All right, added in 8f6b811

}
}

case class TestDataUnion(x: Int, y: Int, z: Int)
Expand Down