From cd6bb9569e5f8a0a9c6b55473c13a0b453ee6c8f Mon Sep 17 00:00:00 2001
From: Moritz Mack <mmack@talend.com>
Date: Wed, 29 Jun 2022 14:51:50 +0200
Subject: [PATCH] Deprecate runner support for Spark 2.4 (closes #22094)

---
 CHANGES.md                                        |  1 +
 .../spark/translation/SparkContextFactory.java    |  8 +++++++-
 .../content/en/documentation/runners/spark.md     | 15 ++++++++-------
 3 files changed, 16 insertions(+), 8 deletions(-)

diff --git a/CHANGES.md b/CHANGES.md
index 9a5873ae940b..53af5dd28b0b 100644
--- a/CHANGES.md
+++ b/CHANGES.md
@@ -70,6 +70,7 @@
 
 ## Deprecations
 
+* Support for Spark 2.4.x is deprecated and will be dropped with the release of Beam 2.44.0 or soon after (Spark runner) ([#22094](https://github.com/apache/beam/issues/22094)).
 * X behavior is deprecated and will be removed in X versions ([#X](https://github.com/apache/beam/issues/X)).
 
 ## Bugfixes
diff --git a/runners/spark/src/main/java/org/apache/beam/runners/spark/translation/SparkContextFactory.java b/runners/spark/src/main/java/org/apache/beam/runners/spark/translation/SparkContextFactory.java
index 9f9465ccde8f..4b714b655818 100644
--- a/runners/spark/src/main/java/org/apache/beam/runners/spark/translation/SparkContextFactory.java
+++ b/runners/spark/src/main/java/org/apache/beam/runners/spark/translation/SparkContextFactory.java
@@ -143,6 +143,12 @@ private static JavaSparkContext createSparkContext(SparkPipelineOptions options)
     conf.setAppName(options.getAppName());
     // register immutable collections serializers because the SDK uses them.
     conf.set("spark.kryo.registrator", SparkRunnerKryoRegistrator.class.getName());
-    return new JavaSparkContext(conf);
+    JavaSparkContext jsc = new JavaSparkContext(conf);
+    if (jsc.sc().version().startsWith("2")) {
+      LOG.warn(
+          "Support for Spark 2 is deprecated, this runner will be removed in a few releases.\n"
+              + "Spark 2 is reaching its EOL, consider migrating to Spark 3.");
+    }
+    return jsc;
   }
 }
diff --git a/website/www/site/content/en/documentation/runners/spark.md b/website/www/site/content/en/documentation/runners/spark.md
index 91b72d542a75..abc1031840ba 100644
--- a/website/www/site/content/en/documentation/runners/spark.md
+++ b/website/www/site/content/en/documentation/runners/spark.md
@@ -67,7 +67,8 @@ the portable Runner. For more information on portability, please visit the
 
 ## Spark Runner prerequisites and setup
 
-The Spark runner currently supports Spark's 2.x branch, and more specifically any version greater than 2.4.0.
+The Spark runner currently supports Spark's 3.1.x branch.
+> **Note:** Support for Spark 2.4.x is deprecated and will be dropped with the release of Beam 2.44.0 (or soon after).
 
 {{< paragraph class="language-java" >}}
 You can add a dependency on the latest version of the Spark runner by adding to your pom.xml the following:
@@ -76,7 +77,7 @@ You can add a dependency on the latest version of the Spark runner by adding to
 {{< highlight java >}}
 <dependency>
   <groupId>org.apache.beam</groupId>
-  <artifactId>beam-runners-spark</artifactId>
+  <artifactId>beam-runners-spark-3</artifactId>
   <version>{{< param release_latest >}}</version>
 </dependency>
 {{< /highlight >}}
@@ -90,13 +91,13 @@ In some cases, such as running in local mode/Standalone, your (self-contained) a
 {{< highlight java >}}
 <dependency>
   <groupId>org.apache.spark</groupId>
-  <artifactId>spark-core_2.11</artifactId>
+  <artifactId>spark-core_2.12</artifactId>
   <version>${spark.version}</version>
 </dependency>
 
 <dependency>
   <groupId>org.apache.spark</groupId>
-  <artifactId>spark-streaming_2.11</artifactId>
+  <artifactId>spark-streaming_2.12</artifactId>
   <version>${spark.version}</version>
 </dependency>
 {{< /highlight >}}
@@ -193,7 +194,7 @@ download it on the [Downloads page](/get-started/downloads/).
 {{< paragraph class="language-py" >}}
 1. Start the JobService endpoint:
     * with Docker (preferred): `docker run --net=host apache/beam_spark_job_server:latest`
-    * or from Beam source code: `./gradlew :runners:spark:2:job-server:runShadow`
+    * or from Beam source code: `./gradlew :runners:spark:3:job-server:runShadow`
 {{< /paragraph >}}
 
 {{< paragraph class="language-py" >}}
@@ -228,7 +229,7 @@ For more details on the different deployment modes see: [Standalone](https://spa
 {{< paragraph class="language-py" >}}
 2. Start JobService that will connect with the Spark master:
     * with Docker (preferred): `docker run --net=host apache/beam_spark_job_server:latest --spark-master-url=spark://localhost:7077`
-    * or from Beam source code: `./gradlew :runners:spark:2:job-server:runShadow -PsparkMasterUrl=spark://localhost:7077`
+    * or from Beam source code: `./gradlew :runners:spark:3:job-server:runShadow -PsparkMasterUrl=spark://localhost:7077`
 {{< /paragraph >}}
 
 {{< paragraph class="language-py" >}}3. Submit the pipeline as above.
@@ -246,7 +247,7 @@ To run Beam jobs written in Python, Go, and other supported languages, you can u
 
 The following example runs a portable Beam job in Python from the Dataproc cluster's master node with Yarn backed.
 
-> Note: This example executes successfully with Dataproc 2.0, Spark 2.4.8 and 3.1.2 and Beam 2.37.0.
+> Note: This example executes successfully with Dataproc 2.0, Spark 3.1.2 and Beam 2.37.0.
 
 1. Create a Dataproc cluster with [Docker](https://cloud.google.com/dataproc/docs/concepts/components/docker) component enabled.