TEZ-4598: JDK-17: Migrate Tez to jdk17 (compile & runtime)

maheshrajus · maheshrajus · commit 1369989a40e6 · 2025-05-13T14:40:14.000+05:30
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
@@ -27,7 +27,7 @@ jobs:
   build:
     strategy:
       matrix:
-        java-version: [8, 11, 17]
+        java-version: [17]
         os: [ubuntu-latest, macos-latest]
     runs-on: ${{ matrix.os }}
     steps:
diff --git a/BUILDING.txt b/BUILDING.txt
@@ -8,7 +8,7 @@ Requirements:
 
 * JDK 1.8+
 * Maven 3.6.3 or later
-* spotbugs 4.2.2 or later (if running spotbugs)
+* spotbugs 4.9.3 or later (if running spotbugs)
 * ProtocolBuffer 3.21.1
 * Internet connection for first build (to fetch all dependencies)
 * Hadoop version should be 2.7.0 or higher.
diff --git a/Jenkinsfile b/Jenkinsfile
@@ -147,10 +147,8 @@ pipeline {
                         # help keep the ASF boxes clean
                         YETUS_ARGS+=("--sentinel")
 
-                        # test with Java 8 and 11
-                        YETUS_ARGS+=("--java-home=/usr/lib/jvm/java-8-openjdk-amd64")
-                        YETUS_ARGS+=("--multijdkdirs=/usr/lib/jvm/java-11-openjdk-amd64")
-                        YETUS_ARGS+=("--multijdktests=compile")
+                        # test with Java 17
+                        YETUS_ARGS+=("--java-home=/usr/lib/jvm/java-17-openjdk-amd64")
                         YETUS_ARGS+=("--debug")
 
                         # write Yetus report as GitHub comment (YETUS-1102)
diff --git a/build-tools/docker/Dockerfile b/build-tools/docker/Dockerfile
@@ -165,35 +165,24 @@ SHELL ["/bin/bash", "-o", "pipefail", "-c"]
 ####
 
 ####
-# OpenJDK 8
-####
+# OpenJDK 17
 # hadolint ignore=DL3008
-RUN apt-get -q update && apt-get -q install --no-install-recommends -y openjdk-8-jdk-headless \
-    && apt-get clean \
-    && rm -rf /var/lib/apt/lists/*
+RUN apt-get -q update \
+    && DEBIAN_FRONTEND=noninteractive apt-get install --no-install-recommends -y openjdk-17-jdk \
+    && apt-get clean && rm -rf /var/lib/apt/lists/*
 
-####
-# OpenJDK 11 (but keeps default to JDK8)
-# NOTE: This default only works when Apache Yetus is launched
-# _in_ the container and not outside of it!
-####
-# hadolint ignore=DL3008
-RUN apt-get -q update && apt-get -q install --no-install-recommends -y default-jre-headless openjdk-11-jdk-headless \
-    && apt-get clean \
-    && rm -rf /var/lib/apt/lists/* \
-    && update-java-alternatives -s java-1.8.0-openjdk-amd64 || :
-# since update alternatives might fail on executables that we don't really need (e.g. appletviewer)
-# and return with exit code <0 (actually: 2), we can simply do a sanity check if the version is
-# as expected for "java" executable after the update and go on
-RUN java -version 2>&1 | grep "1.8.0" && rm -f /usr/lib/jvm/default-java \
-    && ln -s java-8-openjdk-amd64 /usr/lib/jvm/default-java
-ENV JAVA_HOME /usr/lib/jvm/java-8-openjdk-amd64
+# Set JAVA_HOME and PATH environment variables
+ENV JAVA_HOME=/usr/lib/jvm/java-17-openjdk-amd64
+ENV PATH="${JAVA_HOME}/bin:${PATH}"
+
+# Set the default Java version using update-alternatives
+RUN update-alternatives --install /usr/bin/java java /usr/lib/jvm/java-17-openjdk-amd64/bin/java 1
 
 #######
-# Install SpotBugs 4.2.2
+# Install SpotBugs 4.9.3
 #######
 RUN mkdir -p /opt/spotbugs \
-    && curl -L -s -S https://github.com/spotbugs/spotbugs/releases/download/4.2.2/spotbugs-4.2.2.tgz \
+    && curl -L -s -S https://github.com/spotbugs/spotbugs/releases/download/4.9.3/spotbugs-4.9.3.tgz \
       -o /opt/spotbugs.tgz \
     && tar xzf /opt/spotbugs.tgz --strip-components 1 -C /opt/spotbugs \
     && chmod +x /opt/spotbugs/bin/*
diff --git a/pom.xml b/pom.xml
@@ -41,8 +41,8 @@
 
   <properties>
     <!-- Build Properties -->
-    <maven.compiler.source>1.8</maven.compiler.source>
-    <maven.compiler.target>1.8</maven.compiler.target>
+    <maven.compiler.source>17</maven.compiler.source>
+    <maven.compiler.target>17</maven.compiler.target>
     <maven.test.redirectTestOutputToFile>true</maven.test.redirectTestOutputToFile>
     <surefire.version>3.0.0-M4</surefire.version>
     <failIfNoTests>false</failIfNoTests>
@@ -70,8 +70,8 @@
     <commons-lang.version>2.6</commons-lang.version>
     <clover.license>${user.home}/clover.license</clover.license>
     <dependency-check-maven.version>3.2.0</dependency-check-maven.version>
-    <spotbugs.version>4.2.2</spotbugs.version>
-    <spotbugs-maven-plugin.version>4.2.0</spotbugs-maven-plugin.version>
+    <spotbugs.version>4.9.3</spotbugs.version>
+    <spotbugs-maven-plugin.version>4.9.3.0</spotbugs-maven-plugin.version>
     <frontend-maven-plugin.version>1.8.0</frontend-maven-plugin.version>
     <guava.version>32.0.1-jre</guava.version>
     <hadoop.version>3.4.1</hadoop.version>
@@ -102,6 +102,11 @@
     <snappy-java.version>1.1.10.4</snappy-java.version>
     <test.build.data>${project.build.directory}/tmp</test.build.data>
     <wro4j-maven-plugin.version>1.7.9</wro4j-maven-plugin.version>
+    <test.jvm.args>
+      --add-opens=java.base/java.lang=ALL-UNNAMED
+      --add-opens=java.base/java.util=ALL-UNNAMED
+      --add-opens java.base/java.io=ALL-UNNAMED
+    </test.jvm.args>
 
     <maven.javadoc.skip>true</maven.javadoc.skip> <!-- enabled only in relevant modules separately -->
   </properties>
@@ -922,7 +927,9 @@
             <forkCount>1</forkCount>
             <reuseForks>false</reuseForks>
             <forkedProcessTimeoutInSeconds>900</forkedProcessTimeoutInSeconds>
+            <testFailureIgnore>true</testFailureIgnore>
             <argLine>-Xmx1024m -XX:+HeapDumpOnOutOfMemoryError</argLine>
+            <argLine>${test.jvm.args}</argLine>
             <environmentVariables>
               <JAVA_HOME>${java.home}</JAVA_HOME>
               <MALLOC_ARENA_MAX>4</MALLOC_ARENA_MAX>
diff --git a/tez-api/src/main/java/org/apache/tez/client/TezClientUtils.java b/tez-api/src/main/java/org/apache/tez/client/TezClientUtils.java
@@ -1063,6 +1063,7 @@ static String constructAMLaunchOpts(TezConfiguration tezConf, Resource capabilit
     amOpts = amOpts + tezConf.get(TezConfiguration.TEZ_AM_LAUNCH_CMD_OPTS,
         TezConfiguration.TEZ_AM_LAUNCH_CMD_OPTS_DEFAULT);
 
+    amOpts = amOpts + TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_JDK17_CMD_OPTS_DEFAULT;
     amOpts = maybeAddDefaultMemoryJavaOpts(amOpts, capability,
         tezConf.getDouble(TezConfiguration.TEZ_CONTAINER_MAX_JAVA_HEAP_FRACTION,
             TezConfiguration.TEZ_CONTAINER_MAX_JAVA_HEAP_FRACTION_DEFAULT));
diff --git a/tez-api/src/main/java/org/apache/tez/dag/api/TezConfiguration.java b/tez-api/src/main/java/org/apache/tez/dag/api/TezConfiguration.java
@@ -399,6 +399,8 @@ public TezConfiguration(boolean loadDefaults) {
   public static final String TEZ_AM_LAUNCH_CLUSTER_DEFAULT_CMD_OPTS_DEFAULT =
       "-server -Djava.net.preferIPv4Stack=true -Dhadoop.metrics.log.level=WARN";
 
+  public static final String TEZ_AM_LAUNCH_CLUSTER_JDK17_CMD_OPTS_DEFAULT =
+      " --add-opens java.base/java.lang=ALL-UNNAMED";
   /**
    * String value. Command line options provided during the launch of the Tez
    * AppMaster process. Its recommended to not set any Xmx or Xms in these launch opts so that
diff --git a/tez-api/src/test/java/org/apache/tez/client/TestTezClientUtils.java b/tez-api/src/test/java/org/apache/tez/client/TestTezClientUtils.java
@@ -534,7 +534,8 @@ public void testAMCommandOpts() {
         TezClientUtils.constructAMLaunchOpts(tezConf, Resource.newInstance(1024, 1));
     assertEquals(tmpOpts + " "
         + TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_DEFAULT_CMD_OPTS_DEFAULT + " "
-        + amCommandOpts,
+        + amCommandOpts
+        + TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_JDK17_CMD_OPTS_DEFAULT,
         amOptsConstructed);
 
     // Test2: Setup cluster-default command opts explicitly
@@ -543,7 +544,8 @@ public void testAMCommandOpts() {
     tezConf.set(TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_DEFAULT_CMD_OPTS, clusterDefaultCommandOpts);
     amOptsConstructed =
         TezClientUtils.constructAMLaunchOpts(tezConf, Resource.newInstance(1024, 1));
-    assertEquals(tmpOpts + " " + clusterDefaultCommandOpts + " " + amCommandOpts, amOptsConstructed);
+    assertEquals(tmpOpts + " " + clusterDefaultCommandOpts + " " + amCommandOpts
+        + TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_JDK17_CMD_OPTS_DEFAULT, amOptsConstructed);
 
 
     // Test3: Don't setup Xmx explicitly
@@ -555,7 +557,7 @@ public void testAMCommandOpts() {
     // It's OK for the Xmx value to show up before cluster default options, since Xmx will not be replaced if it already exists.
     assertEquals(
         " -Xmx" + ((int) (1024 * factor)) + "m" + " " + tmpOpts + " " + clusterDefaultCommandOpts + " " +
-            amCommandOpts,
+            amCommandOpts + TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_JDK17_CMD_OPTS_DEFAULT,
         amOptsConstructed);
 
     // Test4: Ensure admin options with Xmx does not cause them to be overridden. This should almost never be done though.
@@ -564,7 +566,8 @@ public void testAMCommandOpts() {
     tezConf.set(TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_DEFAULT_CMD_OPTS, clusterDefaultCommandOpts);
     amOptsConstructed =
         TezClientUtils.constructAMLaunchOpts(tezConf, Resource.newInstance(1024, 1));
-    assertEquals(tmpOpts + " " + clusterDefaultCommandOpts + " " + amCommandOpts, amOptsConstructed);
+    assertEquals(tmpOpts + " " + clusterDefaultCommandOpts + " " + amCommandOpts
+        + TezConfiguration.TEZ_AM_LAUNCH_CLUSTER_JDK17_CMD_OPTS_DEFAULT, amOptsConstructed);
   }
 
   @Test(timeout = 5000)
diff --git a/tez-dag/pom.xml b/tez-dag/pom.xml
@@ -200,6 +200,7 @@
         <groupId>org.apache.maven.plugins</groupId>
         <artifactId>maven-surefire-plugin</artifactId>
         <configuration>
+          <argLine>${test.jvm.args}</argLine>
           <environmentVariables>
             <LOG_DIRS>${test.log.dir}</LOG_DIRS>
           </environmentVariables>
diff --git a/tez-ext-service-tests/pom.xml b/tez-ext-service-tests/pom.xml
@@ -161,6 +161,7 @@
         <groupId>org.apache.maven.plugins</groupId>
         <artifactId>maven-surefire-plugin</artifactId>
         <configuration>
+          <argLine>${test.jvm.args}</argLine>
           <environmentVariables>
             <LOG_DIRS>${test.log.dir}</LOG_DIRS>
           </environmentVariables>
diff --git a/tez-ext-service-tests/src/test/java/org/apache/tez/dag/app/rm/TezTestServiceTaskSchedulerService.java b/tez-ext-service-tests/src/test/java/org/apache/tez/dag/app/rm/TezTestServiceTaskSchedulerService.java
@@ -60,7 +60,7 @@ public class TezTestServiceTaskSchedulerService extends TaskScheduler {
   // AppIdIdentifier to avoid conflicts with other containers in the system.
 
   // Per instance
-  private final int memoryPerInstance;
+  private final long memoryPerInstance;
   private final int coresPerInstance;
   private final int executorsPerInstance;
 
@@ -84,7 +84,7 @@ public TezTestServiceTaskSchedulerService(TaskSchedulerContext taskSchedulerCont
       throw new TezUncheckedException(e);
     }
     this.memoryPerInstance = conf
-        .getInt(TezTestServiceConfConstants.TEZ_TEST_SERVICE_MEMORY_PER_INSTANCE_MB, -1);
+        .getLong(TezTestServiceConfConstants.TEZ_TEST_SERVICE_MEMORY_PER_INSTANCE_MB, -1);
     Preconditions.checkArgument(memoryPerInstance > 0,
         TezTestServiceConfConstants.TEZ_TEST_SERVICE_MEMORY_PER_INSTANCE_MB +
             " must be configured");
@@ -145,7 +145,7 @@ public void dagComplete() {
   @Override
   public Resource getTotalResources() {
     return Resource
-        .newInstance(Ints.checkedCast(serviceHosts.size() * memoryPerInstance),
+        .newInstance(serviceHosts.size() * memoryPerInstance,
             serviceHosts.size() * coresPerInstance);
   }
 
diff --git a/tez-runtime-library/src/main/java/org/apache/tez/runtime/library/common/shuffle/impl/SimpleFetchedInputAllocator.java b/tez-runtime-library/src/main/java/org/apache/tez/runtime/library/common/shuffle/impl/SimpleFetchedInputAllocator.java
@@ -20,6 +20,7 @@
 
 import java.io.IOException;
 
+import com.google.common.annotations.VisibleForTesting;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 import org.apache.hadoop.classification.InterfaceAudience.Private;
@@ -56,7 +57,8 @@ public class SimpleFetchedInputAllocator implements FetchedInputAllocator,
 
   // Configuration parameters
   private final long memoryLimit;
-  private final long maxSingleShuffleLimit;
+  @VisibleForTesting
+  final long maxSingleShuffleLimit;
 
   private final long maxAvailableTaskMemory;
   private final long initialMemoryAvailable;
diff --git a/tez-runtime-library/src/test/java/org/apache/tez/runtime/library/common/shuffle/impl/TestSimpleFetchedInputAllocator.java b/tez-runtime-library/src/test/java/org/apache/tez/runtime/library/common/shuffle/impl/TestSimpleFetchedInputAllocator.java
@@ -19,6 +19,7 @@
 package org.apache.tez.runtime.library.common.shuffle.impl;
 
 import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertTrue;
 
 import java.io.File;
 import java.io.IOException;
@@ -36,55 +37,88 @@
 public class TestSimpleFetchedInputAllocator {
 
   private static final Logger LOG = LoggerFactory.getLogger(TestSimpleFetchedInputAllocator.class);
-  
+
   @Test(timeout = 5000)
   public void testInMemAllocation() throws IOException {
     File localDirs = new File(System.getProperty("test.build.data", "/tmp"), this.getClass().getName());
     Configuration conf = new Configuration();
-    
-    long jvmMax = Runtime.getRuntime().maxMemory();
+
+    long jvmMax = 954728448L;
     LOG.info("jvmMax: " + jvmMax);
-    
+
     float bufferPercent = 0.1f;
     conf.setFloat(TezRuntimeConfiguration.TEZ_RUNTIME_SHUFFLE_FETCH_BUFFER_PERCENT, bufferPercent);
     conf.setFloat(TezRuntimeConfiguration.TEZ_RUNTIME_SHUFFLE_MEMORY_LIMIT_PERCENT, 1.0f);
     conf.setStrings(TezRuntimeFrameworkConfigs.LOCAL_DIRS, localDirs.getAbsolutePath());
-    
+
     long inMemThreshold = (long) (bufferPercent * jvmMax);
     LOG.info("InMemThreshold: " + inMemThreshold);
 
     SimpleFetchedInputAllocator inputManager = new SimpleFetchedInputAllocator(
-        "srcName", UUID.randomUUID().toString(), 123, conf,
-        Runtime.getRuntime().maxMemory(), inMemThreshold);
+            "srcName", UUID.randomUUID().toString(), 123, conf,
+            jvmMax, inMemThreshold);
 
     long requestSize = (long) (0.4f * inMemThreshold);
-    long compressedSize = 1l;
+    long compressedSize = 1L;
     LOG.info("RequestSize: " + requestSize);
-    
+
     FetchedInput fi1 = inputManager.allocate(requestSize, compressedSize, new InputAttemptIdentifier(1, 1));
     assertEquals(FetchedInput.Type.MEMORY, fi1.getType());
-    
-    
+
     FetchedInput fi2 = inputManager.allocate(requestSize, compressedSize, new InputAttemptIdentifier(2, 1));
     assertEquals(FetchedInput.Type.MEMORY, fi2.getType());
-    
-    
+
     // Over limit by this point. Next reserve should give back a DISK allocation
     FetchedInput fi3 = inputManager.allocate(requestSize, compressedSize, new InputAttemptIdentifier(3, 1));
     assertEquals(FetchedInput.Type.DISK, fi3.getType());
-    
-    
+
     // Freed one memory allocation. Next should be mem again.
     fi1.abort();
     fi1.free();
     FetchedInput fi4 = inputManager.allocate(requestSize, compressedSize, new InputAttemptIdentifier(4, 1));
     assertEquals(FetchedInput.Type.MEMORY, fi4.getType());
-    
-    // Freed one disk allocation. Next sould be disk again (no mem freed)
+
+    // Freed one disk allocation. Next should be disk again (no mem freed)
     fi3.abort();
     fi3.free();
     FetchedInput fi5 = inputManager.allocate(requestSize, compressedSize, new InputAttemptIdentifier(4, 1));
     assertEquals(FetchedInput.Type.DISK, fi5.getType());
   }
 
+  /**
+   * This method tests the allocation behavior of SimpleFetchedInputAllocator when
+   * a high `maxMemory` is reported by the Runtime.The allocation results in a
+   * DISK input because the `requestSize` exceeds the `maxSingleShuffleLimit`.
+   */
+  @Test(timeout = 5000)
+  public void testInMemAllocationWithJvmMaxMemory() throws IOException {
+    File localDirs = new File(System.getProperty("test.build.data", "/tmp"), this.getClass().getName());
+    Configuration conf = new Configuration();
+
+    long jvmMax = Runtime.getRuntime().maxMemory();
+    LOG.info("jvmMax: " + jvmMax);
+
+    float bufferPercent = 0.1f;
+    conf.setFloat(TezRuntimeConfiguration.TEZ_RUNTIME_SHUFFLE_FETCH_BUFFER_PERCENT, bufferPercent);
+    conf.setFloat(TezRuntimeConfiguration.TEZ_RUNTIME_SHUFFLE_MEMORY_LIMIT_PERCENT, 1.0f);
+    conf.setStrings(TezRuntimeFrameworkConfigs.LOCAL_DIRS, localDirs.getAbsolutePath());
+
+    long inMemThreshold = (long) (bufferPercent * jvmMax);
+    LOG.info("InMemThreshold: " + inMemThreshold);
+
+    SimpleFetchedInputAllocator inputManager = new SimpleFetchedInputAllocator(
+            "srcName", UUID.randomUUID().toString(), 123, conf,
+            jvmMax, inMemThreshold);
+
+    long requestSize = (long) (0.4f * inMemThreshold) + 100L;
+    long compressedSize = 1L;
+    LOG.info("RequestSize: " + requestSize);
+
+    // check if requestSize is greater than maxSingleShuffleLimit
+    assertTrue(requestSize > inputManager.maxSingleShuffleLimit);
+
+    // requestSize is greater than the maxSingleShuffleLimit, so allocation is from DISK
+    FetchedInput fi1 = inputManager.allocate(requestSize, compressedSize, new InputAttemptIdentifier(1, 1));
+    assertEquals(FetchedInput.Type.DISK, fi1.getType());
+  }
 }
diff --git a/tez-tests/pom.xml b/tez-tests/pom.xml
@@ -160,6 +160,7 @@
         <groupId>org.apache.maven.plugins</groupId>
         <artifactId>maven-surefire-plugin</artifactId>
         <configuration>
+          <argLine>${test.jvm.args}</argLine>
           <environmentVariables>
             <LOG_DIRS>${test.log.dir}</LOG_DIRS>
           </environmentVariables>
