Skip to content

[SPARK-20920][SQL] ForkJoinPool pools are leaked when writing hive tables with many partitions #18216

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
srowen wants to merge 2 commits into from
Closed

[SPARK-20920][SQL] ForkJoinPool pools are leaked when writing hive tables with many partitions #18216

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
27a76c1
Don't leave thread pool running from AlterTableRecoverPartitionsComma…
srowen Jun 6, 2017
ca6d3cd
Remove trailing whitespace
srowen Jun 6, 2017
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
21 changes: 13 additions & 8 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/command/ddl.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,6 @@ import java.util.Locale

import scala.collection.{GenMap, GenSeq}
import scala.collection.parallel.ForkJoinTaskSupport
import scala.concurrent.forkjoin.ForkJoinPool
import scala.util.control.NonFatal

import org.apache.hadoop.conf.Configuration
Expand All @@ -36,7 +35,7 @@ import org.apache.spark.sql.catalyst.catalog.CatalogTypes.TablePartitionSpec
import org.apache.spark.sql.catalyst.expressions.{Attribute, AttributeReference}
import org.apache.spark.sql.execution.datasources.PartitioningUtils
import org.apache.spark.sql.types._
import org.apache.spark.util.SerializableConfiguration
import org.apache.spark.util.{SerializableConfiguration, ThreadUtils}

// Note: The definition of these commands are based on the ones described in
// https://cwiki.apache.org/confluence/display/Hive/LanguageManual+DDL
Expand Down Expand Up @@ -588,8 +587,15 @@ case class AlterTableRecoverPartitionsCommand(
val threshold = spark.conf.get("spark.rdd.parallelListingThreshold", "10").toInt
val hadoopConf = spark.sparkContext.hadoopConfiguration
val pathFilter = getPathFilter(hadoopConf)
val partitionSpecsAndLocs = scanPartitions(spark, fs, pathFilter, root, Map(),
table.partitionColumnNames, threshold, spark.sessionState.conf.resolver)

val evalPool = ThreadUtils.newForkJoinPool("AlterTableRecoverPartitionsCommand", 8)
val partitionSpecsAndLocs: Seq[(TablePartitionSpec, Path)] =
try {
scanPartitions(spark, fs, pathFilter, root, Map(), table.partitionColumnNames, threshold,
spark.sessionState.conf.resolver, new ForkJoinTaskSupport(evalPool)).seq
} finally {
evalPool.shutdown()
}
val total = partitionSpecsAndLocs.length
logInfo(s"Found $total partitions in $root")

Expand All @@ -610,8 +616,6 @@ case class AlterTableRecoverPartitionsCommand(
Seq.empty[Row]
}

@transient private lazy val evalTaskSupport = new ForkJoinTaskSupport(new ForkJoinPool(8))

private def scanPartitions(
spark: SparkSession,
fs: FileSystem,
Expand All @@ -620,7 +624,8 @@ case class AlterTableRecoverPartitionsCommand(
spec: TablePartitionSpec,
partitionNames: Seq[String],
threshold: Int,
resolver: Resolver): GenSeq[(TablePartitionSpec, Path)] = {
resolver: Resolver,
evalTaskSupport: ForkJoinTaskSupport): GenSeq[(TablePartitionSpec, Path)] = {
if (partitionNames.isEmpty) {
return Seq(spec -> path)
}
Expand All @@ -644,7 +649,7 @@ case class AlterTableRecoverPartitionsCommand(
val value = ExternalCatalogUtils.unescapePathName(ps(1))
if (resolver(columnName, partitionNames.head)) {
scanPartitions(spark, fs, filter, st.getPath, spec ++ Map(partitionNames.head -> value),
partitionNames.drop(1), threshold, resolver)
partitionNames.drop(1), threshold, resolver, evalTaskSupport)
} else {
logWarning(
s"expected partition column ${partitionNames.head}, but got ${ps(0)}, ignoring it")
Expand Down