[KYUUBI #6939] Bump Spark 3.5.5

### Why are the changes needed?

Test Spark 3.5.5 Release Notes

https://spark.apache.org/releases/spark-release-3-5-5.html

### How was this patch tested?

Pass GHA.

### Was this patch authored or co-authored using generative AI tooling?

No.

Closes #6939 from pan3793/spark-3.5.5.

Closes #6939

8c0288ae5 [Cheng Pan] ga
78b0e72db [Cheng Pan] nit
686a7b0a9 [Cheng Pan] fix
d40cc5bba [Cheng Pan] Bump Spark 3.5.5

Authored-by: Cheng Pan <chengpan@apache.org>
Signed-off-by: Cheng Pan <chengpan@apache.org>

.github/workflows/master.yml

@@ -444,8 +444,8 @@ jobs:
           # https://minikube.sigs.k8s.io/docs/handbook/pushing/#7-loading-directly-to-in-cluster-container-runtime
           minikube image load apache/kyuubi:latest
           # pre-install spark into minikube
-          docker pull apache/spark:3.5.4
+          docker pull apache/spark:3.5.5
-          minikube image load apache/spark:3.5.4
+          minikube image load apache/spark:3.5.5
       - name: kubectl pre-check
         run: |
           kubectl get nodes

bin/docker-image-tool.sh

@@ -181,8 +181,8 @@ Examples:
     $0 -r docker.io/myrepo -t v1.8.1 build
     $0 -r docker.io/myrepo -t v1.8.1 push
 
-  - Build and push with tag "v1.8.1" and Spark-3.5.4 as base image to docker.io/myrepo
+  - Build and push with tag "v1.8.1" and Spark-3.5.5 as base image to docker.io/myrepo
-    $0 -r docker.io/myrepo -t v1.8.1 -b BASE_IMAGE=repo/spark:3.5.4 build
+    $0 -r docker.io/myrepo -t v1.8.1 -b BASE_IMAGE=repo/spark:3.5.5 build
     $0 -r docker.io/myrepo -t v1.8.1 push
 
   - Build and push for multiple archs to docker.io/myrepo

docs/deployment/kyuubi_on_kubernetes.md

@@ -42,8 +42,8 @@ Examples:
     $0 -r docker.io/myrepo -t v1.8.1 build
     $0 -r docker.io/myrepo -t v1.8.1 push
 
-  - Build and push with tag "v1.8.1" and Spark-3.5.4 as base image to docker.io/myrepo
+  - Build and push with tag "v1.8.1" and Spark-3.5.5 as base image to docker.io/myrepo
-    $0 -r docker.io/myrepo -t v1.8.1 -b BASE_IMAGE=repo/spark:3.5.4 build
+    $0 -r docker.io/myrepo -t v1.8.1 -b BASE_IMAGE=repo/spark:3.5.5 build
     $0 -r docker.io/myrepo -t v1.8.1 push
 
   - Build and push for multiple archs to docker.io/myrepo

docs/extensions/engines/spark/lineage.md

@@ -117,7 +117,7 @@ Sometimes, it may be incompatible with other Spark distributions, then you may n
 For example,
 
 ```shell
-build/mvn clean package -pl :kyuubi-spark-lineage_2.12 -am -DskipTests -Dspark.version=3.5.4
+build/mvn clean package -pl :kyuubi-spark-lineage_2.12 -am -DskipTests -Dspark.version=3.5.5
 ```
 
 The available `spark.version`s are shown in the following table.

extensions/spark/kyuubi-spark-connector-hive/src/main/scala/org/apache/kyuubi/spark/connector/hive/HiveConnectorUtils.scala

@@ -19,15 +19,16 @@ package org.apache.kyuubi.spark.connector.hive
 
 import java.lang.{Boolean => JBoolean, Long => JLong}
 
+import scala.util.Try
+
 import org.apache.hadoop.fs.{FileStatus, Path}
 import org.apache.hadoop.hive.ql.plan.{FileSinkDesc, TableDesc}
-import org.apache.spark.SPARK_VERSION
 import org.apache.spark.internal.Logging
 import org.apache.spark.sql.SparkSession
 import org.apache.spark.sql.catalyst.InternalRow
 import org.apache.spark.sql.catalyst.catalog.{CatalogTable, CatalogTablePartition}
 import org.apache.spark.sql.connector.catalog.TableChange
-import org.apache.spark.sql.connector.catalog.TableChange.{AddColumn, After, ColumnPosition, DeleteColumn, First, RenameColumn, UpdateColumnComment, UpdateColumnNullability, UpdateColumnPosition, UpdateColumnType}
+import org.apache.spark.sql.connector.catalog.TableChange._
 import org.apache.spark.sql.execution.command.CommandUtils
 import org.apache.spark.sql.execution.command.CommandUtils.{calculateMultipleLocationSizes, calculateSingleLocationSize}
 import org.apache.spark.sql.execution.datasources.{PartitionDirectory, PartitionedFile}
@@ -35,20 +36,18 @@ import org.apache.spark.sql.hive.execution.HiveFileFormat
 import org.apache.spark.sql.internal.SQLConf
 import org.apache.spark.sql.types.{ArrayType, MapType, StructField, StructType}
 
-import org.apache.kyuubi.spark.connector.common.SparkUtils.SPARK_RUNTIME_VERSION
 import org.apache.kyuubi.util.reflect.{DynClasses, DynConstructors, DynMethods}
 import org.apache.kyuubi.util.reflect.ReflectUtils.invokeAs
 
 object HiveConnectorUtils extends Logging {
 
-  // SPARK-43186
+  def getHiveFileFormat(fileSinkConf: FileSinkDesc): HiveFileFormat =
-  def getHiveFileFormat(fileSinkConf: FileSinkDesc): HiveFileFormat = {
+    Try { // SPARK-43186: 3.5.0
-    if (SPARK_RUNTIME_VERSION >= "3.5") {
       DynConstructors.builder()
         .impl(classOf[HiveFileFormat], classOf[FileSinkDesc])
         .build[HiveFileFormat]()
         .newInstance(fileSinkConf)
-    } else if (SPARK_RUNTIME_VERSION >= "3.3") {
+    }.recover { case _: Exception =>
       val shimFileSinkDescClz = DynClasses.builder()
         .impl("org.apache.spark.sql.hive.HiveShim$ShimFileSinkDesc")
         .build()
@@ -67,34 +66,26 @@ object HiveConnectorUtils extends Logging {
         .impl(classOf[HiveFileFormat], shimFileSinkDescClz)
         .build[HiveFileFormat]()
         .newInstance(shimFileSinkDesc)
-    } else {
+    }.get
-      throw unsupportedSparkVersion()
-    }
-  }
 
-  // SPARK-41970
+  def partitionedFilePath(file: PartitionedFile): String =
-  def partitionedFilePath(file: PartitionedFile): String = {
+    Try { // SPARK-41970: 3.4.0
-    if (SPARK_RUNTIME_VERSION >= "3.4") {
       invokeAs[String](file, "urlEncodedPath")
-    } else if (SPARK_RUNTIME_VERSION >= "3.3") {
+    }.recover { case _: Exception =>
       invokeAs[String](file, "filePath")
-    } else {
+    }.get
-      throw unsupportedSparkVersion()
-    }
-  }
 
   def splitFiles(
       sparkSession: SparkSession,
       file: AnyRef,
       filePath: Path,
-      isSplitable: Boolean,
+      isSplitable: JBoolean,
-      maxSplitBytes: Long,
+      maxSplitBytes: JLong,
-      partitionValues: InternalRow): Seq[PartitionedFile] = {
+      partitionValues: InternalRow): Seq[PartitionedFile] =
-
+    Try { // SPARK-42821: 4.0.0-preview2
-    if (SPARK_RUNTIME_VERSION >= "4.0") { // SPARK-42821
       val fileStatusWithMetadataClz = DynClasses.builder()
         .impl("org.apache.spark.sql.execution.datasources.FileStatusWithMetadata")
-        .build()
+        .buildChecked()
       DynMethods
         .builder("splitFiles")
         .impl(
@@ -103,17 +94,40 @@ object HiveConnectorUtils extends Logging {
           classOf[Boolean],
           classOf[Long],
           classOf[InternalRow])
-        .build()
+        .buildChecked()
-        .invoke[Seq[PartitionedFile]](
+        .invokeChecked[Seq[PartitionedFile]](
           null,
           file,
-          isSplitable.asInstanceOf[JBoolean],
+          isSplitable,
-          maxSplitBytes.asInstanceOf[JLong],
+          maxSplitBytes,
           partitionValues)
-    } else if (SPARK_RUNTIME_VERSION >= "3.5") { // SPARK-43039
+    }.recover { case _: Exception => // SPARK-51185: Spark 3.5.5
       val fileStatusWithMetadataClz = DynClasses.builder()
         .impl("org.apache.spark.sql.execution.datasources.FileStatusWithMetadata")
-        .build()
+        .buildChecked()
+      DynMethods
+        .builder("splitFiles")
+        .impl(
+          "org.apache.spark.sql.execution.PartitionedFileUtil",
+          classOf[SparkSession],
+          fileStatusWithMetadataClz,
+          classOf[Path],
+          classOf[Boolean],
+          classOf[Long],
+          classOf[InternalRow])
+        .buildChecked()
+        .invokeChecked[Seq[PartitionedFile]](
+          null,
+          sparkSession,
+          file,
+          filePath,
+          isSplitable,
+          maxSplitBytes,
+          partitionValues)
+    }.recover { case _: Exception => // SPARK-43039: 3.5.0
+      val fileStatusWithMetadataClz = DynClasses.builder()
+        .impl("org.apache.spark.sql.execution.datasources.FileStatusWithMetadata")
+        .buildChecked()
       DynMethods
         .builder("splitFiles")
         .impl(
@@ -123,15 +137,15 @@ object HiveConnectorUtils extends Logging {
           classOf[Boolean],
           classOf[Long],
           classOf[InternalRow])
-        .build()
+        .buildChecked()
-        .invoke[Seq[PartitionedFile]](
+        .invokeChecked[Seq[PartitionedFile]](
           null,
           sparkSession,
           file,
-          isSplitable.asInstanceOf[JBoolean],
+          isSplitable,
-          maxSplitBytes.asInstanceOf[JLong],
+          maxSplitBytes,
           partitionValues)
-    } else if (SPARK_RUNTIME_VERSION >= "3.3") {
+    }.recover { case _: Exception =>
       DynMethods
         .builder("splitFiles")
         .impl(
@@ -142,55 +156,41 @@ object HiveConnectorUtils extends Logging {
           classOf[Boolean],
           classOf[Long],
           classOf[InternalRow])
-        .build()
+        .buildChecked()
-        .invoke[Seq[PartitionedFile]](
+        .invokeChecked[Seq[PartitionedFile]](
           null,
           sparkSession,
           file,
           filePath,
-          isSplitable.asInstanceOf[JBoolean],
+          isSplitable,
-          maxSplitBytes.asInstanceOf[JLong],
+          maxSplitBytes,
           partitionValues)
-    } else {
+    }.get
-      throw unsupportedSparkVersion()
-    }
-  }
 
-  def createPartitionDirectory(values: InternalRow, files: Seq[FileStatus]): PartitionDirectory = {
+  def createPartitionDirectory(values: InternalRow, files: Seq[FileStatus]): PartitionDirectory =
-    if (SPARK_RUNTIME_VERSION >= "3.5") {
+    Try { // SPARK-43039: 3.5.0
       new DynMethods.Builder("apply")
         .impl(classOf[PartitionDirectory], classOf[InternalRow], classOf[Array[FileStatus]])
         .buildChecked()
         .asStatic()
         .invoke[PartitionDirectory](values, files.toArray)
-    } else if (SPARK_RUNTIME_VERSION >= "3.3") {
+    }.recover { case _: Exception =>
       new DynMethods.Builder("apply")
         .impl(classOf[PartitionDirectory], classOf[InternalRow], classOf[Seq[FileStatus]])
         .buildChecked()
         .asStatic()
         .invoke[PartitionDirectory](values, files)
-    } else {
+    }.get
-      throw unsupportedSparkVersion()
-    }
-  }
 
-  def getPartitionFilePath(file: AnyRef): Path = {
+  def getPartitionFilePath(file: AnyRef): Path =
-    if (SPARK_RUNTIME_VERSION >= "3.5") {
+    Try { // SPARK-43039: 3.5.0
       new DynMethods.Builder("getPath")
         .impl("org.apache.spark.sql.execution.datasources.FileStatusWithMetadata")
         .build()
         .invoke[Path](file)
-    } else if (SPARK_RUNTIME_VERSION >= "3.3") {
+    }.recover { case _: Exception =>
       file.asInstanceOf[FileStatus].getPath
-    } else {
+    }.get
-      throw unsupportedSparkVersion()
-    }
-  }
-
-  private def unsupportedSparkVersion(): KyuubiHiveConnectorException = {
-    KyuubiHiveConnectorException(s"Spark version $SPARK_VERSION " +
-      "is not supported by Kyuubi spark hive connector.")
-  }
 
   def calculateTotalSize(
       spark: SparkSession,

integration-tests/kyuubi-kubernetes-it/src/test/scala/org/apache/kyuubi/kubernetes/test/deployment/KyuubiOnKubernetesTestsSuite.scala

@@ -56,7 +56,7 @@ class KyuubiOnKubernetesWithSparkTestsBase extends WithKyuubiServerOnKubernetes
       Map(
         "spark.master" -> s"k8s://$miniKubeApiMaster",
         // We should update spark docker image in ./github/workflows/master.yml at the same time
-        "spark.kubernetes.container.image" -> "apache/spark:3.5.4",
+        "spark.kubernetes.container.image" -> "apache/spark:3.5.5",
         "spark.kubernetes.container.image.pullPolicy" -> "IfNotPresent",
         "spark.executor.memory" -> "512M",
         "spark.driver.memory" -> "1024M",

integration-tests/kyuubi-kubernetes-it/src/test/scala/org/apache/kyuubi/kubernetes/test/spark/SparkOnKubernetesTestsSuite.scala

@@ -51,7 +51,7 @@ abstract class SparkOnKubernetesSuiteBase
     // TODO Support more Spark version
     // Spark official docker image: https://hub.docker.com/r/apache/spark/tags
     KyuubiConf().set("spark.master", s"k8s://$apiServerAddress")
-      .set("spark.kubernetes.container.image", "apache/spark:3.5.4")
+      .set("spark.kubernetes.container.image", "apache/spark:3.5.5")
       .set("spark.kubernetes.container.image.pullPolicy", "IfNotPresent")
       .set("spark.executor.instances", "1")
       .set("spark.executor.memory", "512M")

pom.xml

@@ -200,7 +200,7 @@
           DO NOT forget to change the following properties when change the minor version of Spark:
           `delta.version`, `delta.artifact`, `maven.plugin.scalatest.exclude.tags`
           -->
-        <spark.version>3.5.4</spark.version>
+        <spark.version>3.5.5</spark.version>
         <spark.binary.version>3.5</spark.binary.version>
         <spark.archive.scala.suffix></spark.archive.scala.suffix>
         <spark.archive.name>spark-${spark.version}-bin-hadoop3${spark.archive.scala.suffix}.tgz</spark.archive.name>
@@ -1932,7 +1932,7 @@
                 <java.version>8</java.version>
                 <!--
                   Iceberg drops support for Java 8 since 1.7.0.
-                  And it may have compatible issue with Spark 3.5.4, see Iceberg #11731
+                  And it may have compatible issue with Spark 3.5.4+, see Iceberg #11731
                   -->
                 <iceberg.version>1.6.1</iceberg.version>
             </properties>
@@ -2044,7 +2044,7 @@
                 <module>extensions/spark/kyuubi-spark-connector-hive</module>
             </modules>
             <properties>
-                <spark.version>3.5.4</spark.version>
+                <spark.version>3.5.5</spark.version>
                 <spark.binary.version>3.5</spark.binary.version>
                 <delta.version>3.3.0</delta.version>
                 <delta.artifact>delta-spark_${scala.binary.version}</delta.artifact>