Skip to content

[SPARK-22950][SQL]Handle ChildFirstURLClassLoader's parent #20145

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
wants to merge 4 commits into from
Closed

[SPARK-22950][SQL]Handle ChildFirstURLClassLoader's parent #20145

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
7cb0c44
child first classpath first has no regular parent
yaooqinn Jan 4, 2018
bb27e17
doc
yaooqinn Jan 4, 2018
e1a9f2e
surround with try finally
yaooqinn Jan 4, 2018
fea2490
rm one unit test
yaooqinn Jan 4, 2018
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
4 changes: 3 additions & 1 deletion sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveUtils.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,7 +47,7 @@ import org.apache.spark.sql.internal.SQLConf
import org.apache.spark.sql.internal.SQLConf._
import org.apache.spark.sql.internal.StaticSQLConf.{CATALOG_IMPLEMENTATION, WAREHOUSE_PATH}
import org.apache.spark.sql.types._
import org.apache.spark.util.Utils
import org.apache.spark.util.{ChildFirstURLClassLoader, Utils}


private[spark] object HiveUtils extends Logging {
Expand Down Expand Up @@ -312,6 +312,8 @@ private[spark] object HiveUtils extends Logging {
// starting from the given classLoader.
def allJars(classLoader: ClassLoader): Array[URL] = classLoader match {
case null => Array.empty[URL]
case childFirst: ChildFirstURLClassLoader =>
childFirst.getURLs() ++ allJars(Utils.getSparkClassLoader)
case urlClassLoader: URLClassLoader =>
urlClassLoader.getURLs ++ allJars(urlClassLoader.getParent)
case other => allJars(other.getParent)
Expand Down
30 changes: 30 additions & 0 deletions sql/hive/src/test/scala/org/apache/spark/sql/hive/HiveUtilsSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,11 +17,16 @@

package org.apache.spark.sql.hive

import java.net.URL

import org.apache.hadoop.hive.conf.HiveConf.ConfVars

import org.apache.spark.SparkConf
import org.apache.spark.deploy.SparkHadoopUtil
import org.apache.spark.sql.QueryTest
import org.apache.spark.sql.hive.test.TestHiveSingleton
import org.apache.spark.sql.test.SQLTestUtils
import org.apache.spark.util.{ChildFirstURLClassLoader, MutableURLClassLoader}

class HiveUtilsSuite extends QueryTest with SQLTestUtils with TestHiveSingleton {

Expand All @@ -42,4 +47,29 @@ class HiveUtilsSuite extends QueryTest with SQLTestUtils with TestHiveSingleton
assert(hiveConf("foo") === "bar")
}
}

test("ChildFirstURLClassLoader's parent is null") {
Copy link
Contributor

@cloud-fan cloud-fan Jan 4, 2018

Choose a reason for hiding this comment

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

what is this test trying to test?

Copy link
Member Author

@yaooqinn yaooqinn Jan 4, 2018

Choose a reason for hiding this comment

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

just indicates that if we don't handle child first classloader's parent, we will get caught in exception. if it bothers, i can del it

Copy link
Contributor

@cloud-fan cloud-fan Jan 4, 2018

Choose a reason for hiding this comment

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

I think we can remove it. The below tests case would fail without your PR, that proves your patch already.

val conf = new SparkConf
val contextClassLoader = Thread.currentThread().getContextClassLoader
val loader = new FakeChildFirstURLClassLoader(Array(), contextClassLoader)
Thread.currentThread().setContextClassLoader(loader)
intercept[IllegalArgumentException](
HiveUtils.newClientForMetadata(conf, SparkHadoopUtil.newConfiguration(conf)))
Thread.currentThread().setContextClassLoader(contextClassLoader)
Copy link
Contributor

@cloud-fan cloud-fan Jan 4, 2018

Choose a reason for hiding this comment

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

we should use try .. finally to make sure we always set back the context class loader

Copy link
Member Author

@yaooqinn yaooqinn Jan 4, 2018

Choose a reason for hiding this comment

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

ok

}

test("ChildFirstURLClassLoader's parent is null, get spark classloader instead") {
val conf = new SparkConf
val contextClassLoader = Thread.currentThread().getContextClassLoader
val loader = new ChildFirstURLClassLoader(Array(), contextClassLoader)
Thread.currentThread().setContextClassLoader(loader)
HiveUtils.newClientForMetadata(conf, SparkHadoopUtil.newConfiguration(conf))
Thread.currentThread().setContextClassLoader(contextClassLoader)
Copy link
Contributor

@cloud-fan cloud-fan Jan 4, 2018

Choose a reason for hiding this comment

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

ditto

}
}

/**
* A Fake [[ChildFirstURLClassLoader]] used for test
*/
private[spark] class FakeChildFirstURLClassLoader(urls: Array[URL], parent: ClassLoader)
Copy link
Contributor

@cloud-fan cloud-fan Jan 4, 2018

Choose a reason for hiding this comment

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

This is not a ChildFirstURLClassLoader, shall we make it extend ChildFirstURLClassLoader?

Copy link
Member Author

@yaooqinn yaooqinn Jan 4, 2018

Choose a reason for hiding this comment

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

inheritation maketh case match

extends MutableURLClassLoader(urls, null)