Skip to content

[SPARK-40635][YARN][TESTS] Fix yarn module daily test failed with hadoop2 #38079

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
LuciferYang wants to merge 8 commits into from
Closed

[SPARK-40635][YARN][TESTS] Fix yarn module daily test failed with hadoop2 #38079

Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
e8e6b7f
try filter some test deps from classpath
LuciferYang Oct 3, 2022
5ad8024
test hadoop 2
LuciferYang Oct 2, 2022
93bdf77
try dedup
LuciferYang Oct 3, 2022
64f0e0b
use File.pathSeparator
LuciferYang Oct 3, 2022
b17b27f
test hadoop3
LuciferYang Oct 3, 2022
bd89c95
do de duplication only when Utils.isTesting is true
LuciferYang Oct 5, 2022
4e9a0d8
re test haddop2
LuciferYang Oct 5, 2022
2712ae0
add comments and re-test hadoop-3
LuciferYang Oct 5, 2022
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
13 changes: 12 additions & 1 deletion resource-managers/yarn/src/main/scala/org/apache/spark/deploy/yarn/Client.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1470,6 +1470,11 @@ private[spark] object Client extends Logging {
addClasspathEntry(getClusterPath(sparkConf, cp), env)
}

val cpSet = extraClassPath match {
case Some(classPath) if Utils.isTesting => classPath.split(File.pathSeparator).toSet
case _ => Set.empty[String]
}

addClasspathEntry(Environment.PWD.$$(), env)

addClasspathEntry(Environment.PWD.$$() + Path.SEPARATOR + LOCALIZED_CONF_DIR, env)
Expand Down Expand Up @@ -1513,7 +1518,13 @@ private[spark] object Client extends Logging {
}

sys.env.get(ENV_DIST_CLASSPATH).foreach { cp =>
addClasspathEntry(getClusterPath(sparkConf, cp), env)
// SPARK-40635: during the test, add a jar de-duplication process to avoid
// that the startup command can't be executed due to the too long classpath.
val newCp = if (Utils.isTesting) {
Copy link
Member

@HyukjinKwon HyukjinKwon Oct 5, 2022

Choose a reason for hiding this comment

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

Maybe let's add a comment with mentioning the JIRA number.

Copy link
Contributor Author

@LuciferYang LuciferYang Oct 5, 2022

Choose a reason for hiding this comment

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

ok, will add comments after the hadoop2 check pass

cp.split(File.pathSeparator)
.filterNot(cpSet.contains).mkString(File.pathSeparator)
} else cp
addClasspathEntry(getClusterPath(sparkConf, newCp), env)
Copy link
Contributor Author

@LuciferYang LuciferYang Oct 3, 2022

Choose a reason for hiding this comment

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

Try to deduplicate jars that have been added to CLASSPATH through extraClassPath

Copy link
Contributor Author

@LuciferYang LuciferYang Oct 3, 2022 •
edited
Loading

Choose a reason for hiding this comment

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

Is it necessary to change it to

      val newCp = if (Utils.isTesting) {
        cp.split(File.pathSeparator)
          .filterNot(cpSet.contains).mkString(File.pathSeparator)
      } else cp

but I think de duplication may also be useful for the production environment

}

// Add the localized Hadoop config at the end of the classpath, in case it contains other
Expand Down