HyukjinKwon commented on code in PR #37417:
URL: https://github.com/apache/spark/pull/37417#discussion_r939616121
##########
core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala:
##########
@@ -381,45 +381,52 @@ private[spark] class SparkSubmit extends Logging {
localPyFiles = Option(args.pyFiles).map {
downloadFileList(_, targetDir, sparkConf, hadoopConf)
}.orNull
-
if (isKubernetesClusterModeDriver) {
- // Replace with the downloaded local jar path to avoid propagating
hadoop compatible uris.
- // Executors will get the jars from the Spark file server.
- // Explicitly download the related files here
- args.jars = localJars
- val filesLocalFiles = Option(args.files).map {
- downloadFileList(_, targetDir, sparkConf, hadoopConf)
- }.orNull
- val archiveLocalFiles = Option(args.archives).map { uris =>
+ // SPARK-33748: this mimics the behaviour of Yarn cluster mode. If the
driver is running
+ // in cluster mode, the archives should be available in the driver's
current working
+ // directory too.
+ // SPARK-33782 : This downloads all the files , jars , archiveFiles
and pyfiles to current
+ // working directory
+ def downloadResourcesToCurrentDirectory(uris: String): String = {
val resolvedUris = Utils.stringToSeq(uris).map(Utils.resolveURI)
- val localArchives = downloadFileList(
+ val localResources = downloadFileList(
resolvedUris.map(
UriBuilder.fromUri(_).fragment(null).build().toString).mkString(","),
targetDir, sparkConf, hadoopConf)
-
- // SPARK-33748: this mimics the behaviour of Yarn cluster mode. If
the driver is running
- // in cluster mode, the archives should be available in the driver's
current working
- // directory too.
-
Utils.stringToSeq(localArchives).map(Utils.resolveURI).zip(resolvedUris).map {
- case (localArchive, resolvedUri) =>
- val source = new File(localArchive.getPath)
+
Utils.stringToSeq(localResources).map(Utils.resolveURI).zip(resolvedUris).map {
+ case (localResources, resolvedUri) =>
+ val source = new File(localResources.getPath)
val dest = new File(
".",
if (resolvedUri.getFragment != null) resolvedUri.getFragment
else source.getName)
logInfo(
- s"Unpacking an archive $resolvedUri " +
+ s"Files $resolvedUri " +
s"from ${source.getAbsolutePath} to ${dest.getAbsolutePath}")
Utils.deleteRecursively(dest)
Utils.unpack(source, dest)
Review Comment:
Hm, why do we try this unpack for jars and files too? I think we should just
call `downloadFileList` for them
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
