Bump SLF4J from 1.7.30 to 2.0.16.

.github/workflows/beam_PerformanceTests_SparkReceiver_IO.yml

@@ -96,7 +96,7 @@ jobs:
       - name: run integrationTest
         uses: ./.github/actions/gradle-command-self-hosted-action
         with:
-          gradle-command: :sdks:java:io:sparkreceiver:2:integrationTest
+          gradle-command: :sdks:java:io:sparkreceiver:3:integrationTest
           arguments: |
             --info \
             --tests org.apache.beam.sdk.io.sparkreceiver.SparkReceiverIOIT \

CHANGES.md

@@ -68,12 +68,13 @@
 ## New Features / Improvements
 
 * Support custom coders in Reshuffle ([#29908](https://github.com/apache/beam/issues/29908), [#33356](https://github.com/apache/beam/issues/33356)).
-
+* [Java] Upgrade SLF4J to 2.0.16. Update default Spark version to 3.5.0. ([#33574](https://github.com/apache/beam/pull/33574))
 * X feature added (Java/Python) ([#X](https://github.com/apache/beam/issues/X)).
 
 ## Breaking Changes
-* [Python] Reshuffle now correctly respects user-specified type hints, fixing a previous bug where it might use FastPrimitivesCoder wrongly. This change could break pipelines with incorrect type hints in Reshuffle. If you have issues after upgrading, temporarily set update_compatibility_version to a previous Beam version to use the old behavior. The recommended solution is to fix the type hints in your code. ([#33932](https://github.com/apache/beam/pull/33932))
 
+* [Python] Reshuffle now correctly respects user-specified type hints, fixing a previous bug where it might use FastPrimitivesCoder wrongly. This change could break pipelines with incorrect type hints in Reshuffle. If you have issues after upgrading, temporarily set update_compatibility_version to a previous Beam version to use the old behavior. The recommended solution is to fix the type hints in your code. ([#33932](https://github.com/apache/beam/pull/33932))
+* [Java] SparkReceiver 2 has been moved to SparkReceiver 3 that supports Spark 3.x. ([#33574](https://github.com/apache/beam/pull/33574))
 * X behavior was changed ([#X](https://github.com/apache/beam/issues/X)).
 
 ## Deprecations

build.gradle.kts

@@ -304,7 +304,7 @@ tasks.register("javaPreCommit") {
   dependsOn(":sdks:java:io:contextualtextio:build")
   dependsOn(":sdks:java:io:expansion-service:build")
   dependsOn(":sdks:java:io:file-based-io-tests:build")
-  dependsOn(":sdks:java:io:sparkreceiver:2:build")
+  dependsOn(":sdks:java:io:sparkreceiver:3:build")
   dependsOn(":sdks:java:io:synthetic:build")
   dependsOn(":sdks:java:io:xml:build")
   dependsOn(":sdks:java:javadoc:allJavadoc")

buildSrc/src/main/groovy/org/apache/beam/gradle/BeamModulePlugin.groovy

@@ -636,12 +636,12 @@ class BeamModulePlugin implements Plugin<Project> {
     def quickcheck_version = "1.0"
     def sbe_tool_version = "1.25.1"
     def singlestore_jdbc_version = "1.1.4"
-    def slf4j_version = "1.7.30"
+    def slf4j_version = "2.0.16"
     def snakeyaml_engine_version = "2.6"
     def snakeyaml_version = "2.2"
     def solace_version = "10.21.0"
     def spark2_version = "2.4.8"
-    def spark3_version = "3.2.2"
+    def spark3_version = "3.5.0"
     def spotbugs_version = "4.0.6"
     def testcontainers_version = "1.19.7"
     // [bomupgrader] determined by: org.apache.arrow:arrow-memory-core, consistent with: google_cloud_platform_libraries_bom

runners/google-cloud-dataflow-java/worker/build.gradle

@@ -87,7 +87,7 @@ applyJavaNature(
                 // TODO(https://github.com/apache/beam/issues/19114): Move DataflowRunnerHarness class under org.apache.beam.runners.dataflow.worker namespace
                 "com/google/cloud/dataflow/worker/DataflowRunnerHarness.class",
                 // Allow slf4j implementation worker for logging during pipeline execution
-                "org/slf4j/impl/**"
+                "org/slf4j/jul/**"
         ],
         generatedClassPatterns: [
                 /^org\.apache\.beam\.runners\.dataflow\.worker\.windmill.*/,
@@ -240,7 +240,7 @@ project.task('validateShadedJarContainsSlf4jJdk14', dependsOn: 'shadowJar') {
     doLast {
         project.configurations.shadow.artifacts.files.each {
             FileTree slf4jImpl = project.zipTree(it).matching {
-                include "org/slf4j/impl/JDK14LoggerAdapter.class"
+                include "org/slf4j/jul/JDK14LoggerAdapter.class"
             }
             outFile.text = slf4jImpl.files
             if (slf4jImpl.files.isEmpty()) {

runners/spark/3/build.gradle

@@ -56,6 +56,26 @@ sparkVersions.each { kv ->
   }
 
   dependencies {
+    // Spark versions prior to 3.4.0 are compiled against SLF4J 1.x.  The
+    // `org.apache.spark.internal.Logging.isLog4j12()` function references an
+    // SLF4J 1.x binding class (org.slf4j.impl.StaticLoggerBinder) which is
+    // no longer available in SLF4J 2.x.  This results in a
+    // `java.lang.NoClassDefFoundError`.
+    //
+    // The workaround is to provide an SLF4J 1.x binding module out of group
+    // `org.slf4j` to resolve the issue.
+    // Module `org.apache.logging.log4j:log4j-slf4j-impl` is an example that
+    // provides a compatible SLF4J 1.x binding regardless SLF4J upgrade.
+    // Binding/provider modules under group `org.slf4j` (e.g.,
+    // slf4j-simple, slf4j-reload4j) get upgraded as a new SLF4J version is in
+    // use, and therefore do not contain the 1.x binding classes.
+    //
+    // Notice that Spark 3.1.x uses `ch.qos.logback:logback-classic` and is
+    // unaffected by the SLF4J upgrade. Spark 3.3.x already uses
+    // `log4j-slf4j-impl` so it is also unaffected.
+    if ("$kv.key" >= "320" && "$kv.key" <= "324") {
+      "sparkVersion$kv.key" library.java.log4j2_slf4j_impl
+    }
     spark.components.each { component -> "sparkVersion$kv.key" "$component:$kv.value" }
   }
 

runners/spark/spark_runner.gradle

@@ -176,6 +176,10 @@ dependencies {
   spark.components.each { component ->
     provided "$component:$spark_version"
   }
+  if ("$spark_version" >= "3.5.0") {
+    implementation "org.apache.spark:spark-common-utils_$spark_scala_version:$spark_version"
+    implementation "org.apache.spark:spark-sql-api_$spark_scala_version:$spark_version"
+  }
   permitUnusedDeclared "org.apache.spark:spark-network-common_$spark_scala_version:$spark_version"
   implementation "io.dropwizard.metrics:metrics-core:4.1.1" // version used by Spark 3.1
   compileOnly "org.scala-lang:scala-library:2.12.15"
@@ -202,6 +206,10 @@ dependencies {
   testImplementation library.java.mockito_core
   testImplementation "org.assertj:assertj-core:3.11.1"
   testImplementation "org.apache.zookeeper:zookeeper:3.4.11"
+  if ("$spark_version" >= "3.5.0") {
+    testImplementation "org.apache.spark:spark-common-utils_$spark_scala_version:$spark_version"
+    testImplementation "org.apache.spark:spark-sql-api_$spark_scala_version:$spark_version"
+  }
   validatesRunner project(path: ":sdks:java:core", configuration: "shadowTest")
   validatesRunner project(path: ":runners:core-java", configuration: "testRuntimeMigration")
   validatesRunner project(":sdks:java:io:hadoop-format")

sdks/java/extensions/arrow/build.gradle

@@ -24,19 +24,10 @@ description = "Apache Beam :: SDKs :: Java :: Extensions :: Arrow"
 dependencies {
     implementation library.java.vendored_guava_32_1_2_jre
     implementation project(path: ":sdks:java:core", configuration: "shadow")
-    implementation(library.java.arrow_vector) {
-        // Arrow 15 has compile dependency of slf4j 2.x where Beam does not support
-        exclude group: 'org.slf4j', module: 'slf4j-api'
-    }
-    implementation(library.java.arrow_memory_core) {
-        // Arrow 15 has compile dependency of slf4j 2.x where Beam does not support
-        exclude group: 'org.slf4j', module: 'slf4j-api'
-    }
+    implementation(library.java.arrow_vector)
+    implementation(library.java.arrow_memory_core)
     implementation library.java.joda_time
-    testImplementation(library.java.arrow_memory_netty) {
-        // Arrow 15 has compile dependency of slf4j 2.x where Beam does not support
-        exclude group: 'org.slf4j', module: 'slf4j-api'
-    }
+    testImplementation(library.java.arrow_memory_netty)
     testImplementation library.java.junit
     testImplementation library.java.hamcrest
     testRuntimeOnly library.java.slf4j_simple

sdks/java/io/cdap/build.gradle

@@ -45,7 +45,11 @@ dependencies {
     implementation library.java.cdap_etl_api
     implementation library.java.cdap_etl_api_spark
     implementation library.java.cdap_hydrator_common
-    implementation library.java.cdap_plugin_hubspot
+    implementation (library.java.cdap_plugin_hubspot) {
+        // Excluding the module for scala 2.11, because Spark 3.x uses scala
+        // 2.12 instead.
+        exclude group: "com.fasterxml.jackson.module", module: "jackson-module-scala_2.11"
+    }
     implementation library.java.cdap_plugin_salesforce
     implementation library.java.cdap_plugin_service_now
     implementation library.java.cdap_plugin_zendesk
@@ -56,11 +60,17 @@ dependencies {
     implementation library.java.jackson_core
     implementation library.java.jackson_databind
     implementation library.java.slf4j_api
-    implementation library.java.spark_streaming
+    implementation (library.java.spark3_streaming) {
+        // Excluding `org.slf4j:jul-to-slf4j` which was introduced as a
+        // transitive dependency in Spark 3.5.0 (particularly from
+        // spark-common-utils_2.12) and would cause stack overflow together with
+        // `org.slf4j:slf4j-jdk14`.
+        exclude group: "org.slf4j", module: "jul-to-slf4j"
+    }
     implementation library.java.tephra
     implementation library.java.vendored_guava_32_1_2_jre
     implementation project(path: ":sdks:java:core", configuration: "shadow")
-    implementation project(":sdks:java:io:sparkreceiver:2")
+    implementation project(":sdks:java:io:sparkreceiver:3")
     implementation project(":sdks:java:io:hadoop-format")
     testImplementation library.java.cdap_plugin_service_now
     testImplementation library.java.cdap_etl_api

sdks/java/io/google-cloud-platform/build.gradle

@@ -138,22 +138,13 @@ dependencies {
   implementation library.java.slf4j_api
   implementation library.java.vendored_grpc_1_69_0
   implementation library.java.vendored_guava_32_1_2_jre
-  implementation(library.java.arrow_memory_core) {
-    // Arrow 15 has compile dependency of slf4j 2.x where Beam does not support
-    exclude group: 'org.slf4j', module: 'slf4j-api'
-  }
-  implementation(library.java.arrow_vector) {
-    // Arrow 15 has compile dependency of slf4j 2.x where Beam does not support
-    exclude group: 'org.slf4j', module: 'slf4j-api'
-  }
+  implementation library.java.arrow_memory_core
+  implementation library.java.arrow_vector
 
   implementation 'com.google.http-client:google-http-client-gson:1.41.2'
   implementation "org.threeten:threetenbp:1.4.4"
 
-  testImplementation(library.java.arrow_memory_netty) {
-    // Arrow 15 has compile dependency of slf4j 2.x where Beam does not support
-    exclude group: 'org.slf4j', module: 'slf4j-api'
-  }
+  testImplementation library.java.arrow_memory_netty
   testImplementation project(path: ":sdks:java:core", configuration: "shadowTest")
   testImplementation project(path: ":sdks:java:extensions:avro", configuration: "testRuntimeMigration")
   testImplementation project(path: ":sdks:java:extensions:google-cloud-platform-core", configuration: "testRuntimeMigration")

sdks/java/io/sparkreceiver/2/README.md → sdks/java/io/sparkreceiver/3/README.md

@@ -18,11 +18,11 @@
 -->
 # SparkReceiverIO
 
-SparkReceiverIO provides I/O transforms to read messages from an [Apache Spark Receiver](https://spark.apache.org/docs/2.4.0/streaming-custom-receivers.html) `org.apache.spark.streaming.receiver.Receiver` as an unbounded source.
+SparkReceiverIO provides I/O transforms to read messages from an [Apache Spark Receiver](https://spark.apache.org/docs/3.5.0/streaming-custom-receivers.html) `org.apache.spark.streaming.receiver.Receiver` as an unbounded source.
 
 ## Prerequistes
 
-SparkReceiverIO supports [Spark Receivers](https://spark.apache.org/docs/2.4.0/streaming-custom-receivers.html) (Spark version 2.4).
+SparkReceiverIO supports [Spark Receivers](https://spark.apache.org/docs/3.5.0/streaming-custom-receivers.html) (Spark version 3.x, tested on Spark version 3.5.0).
 1. Corresponding Spark Receiver should implement [HasOffset](https://github.com/apache/beam/blob/master/sdks/java/io/sparkreceiver/src/main/java/org/apache/beam/sdk/io/sparkreceiver/HasOffset.java) interface.
 2. Records should have the numeric field that represents record offset. *Example:* `RecordId` field for Salesforce and `vid` field for Hubspot Receivers.
    For more details please see [GetOffsetUtils](https://github.com/apache/beam/tree/master/examples/java/cdap/src/main/java/org/apache/beam/examples/complete/cdap/utils/GetOffsetUtils.java) class from CDAP plugins examples.
@@ -53,7 +53,7 @@ To learn more, please check out CDAP Streaming plugins [complete examples](https
 
 ## Dependencies
 
-To use SparkReceiverIO, add a dependency on `beam-sdks-java-io-sparkreceiver`.
+To use SparkReceiverIO, add a dependency on `beam-sdks-java-io-sparkreceiver-3`.
 
 ```maven
 <dependency>

sdks/java/io/sparkreceiver/2/build.gradle → sdks/java/io/sparkreceiver/3/build.gradle

@@ -43,8 +43,8 @@ dependencies {
     implementation library.java.commons_lang3
     implementation library.java.joda_time
     implementation library.java.slf4j_api
-    implementation library.java.spark_streaming
-    implementation library.java.spark_core
+    implementation library.java.spark3_streaming
+    implementation library.java.spark3_core
     implementation library.java.vendored_guava_32_1_2_jre
     implementation project(path: ":sdks:java:core", configuration: "shadow")
     compileOnly "org.scala-lang:scala-library:2.11.12"

settings.gradle.kts

@@ -252,7 +252,7 @@ include(":sdks:java:io:rabbitmq")
 include(":sdks:java:io:redis")
 include(":sdks:java:io:rrio")
 include(":sdks:java:io:solr")
-include(":sdks:java:io:sparkreceiver:2")
+include(":sdks:java:io:sparkreceiver:3")
 include(":sdks:java:io:snowflake")
 include(":sdks:java:io:snowflake:expansion-service")
 include(":sdks:java:io:splunk")