You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hbase.apache.org by ap...@apache.org on 2017/07/17 17:59:16 UTC
[83/94] [abbrv] hbase git commit: HBASE-18175 Add hbase-spark
integration test into hbase-spark-it
HBASE-18175 Add hbase-spark integration test into hbase-spark-it
* adds module hbase-spark-it
* adds test IntegrationTestSparkBulkLoad
* adds resultant jar to bin assembly
Signed-off-by: Mike Drob <md...@apache.org>
Signed-off-by: Sean Busbey <bu...@apache.org>
Project: http://git-wip-us.apache.org/repos/asf/hbase/repo
Commit: http://git-wip-us.apache.org/repos/asf/hbase/commit/79a702d1
Tree: http://git-wip-us.apache.org/repos/asf/hbase/tree/79a702d1
Diff: http://git-wip-us.apache.org/repos/asf/hbase/diff/79a702d1
Branch: refs/heads/HBASE-14070.HLC
Commit: 79a702d1126288a0dd45ba9458265cff8e3619c6
Parents: 81ffd6a
Author: Yi Liang <ea...@gmail.com>
Authored: Wed Jul 12 17:12:52 2017 -0700
Committer: Sean Busbey <bu...@apache.org>
Committed: Fri Jul 14 10:53:40 2017 -0500
----------------------------------------------------------------------
hbase-assembly/pom.xml | 6 +
hbase-assembly/src/main/assembly/src.xml | 1 +
hbase-spark-it/pom.xml | 361 ++++++++++
.../spark/IntegrationTestSparkBulkLoad.java | 661 +++++++++++++++++++
pom.xml | 1 +
5 files changed, 1030 insertions(+)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/hbase/blob/79a702d1/hbase-assembly/pom.xml
----------------------------------------------------------------------
diff --git a/hbase-assembly/pom.xml b/hbase-assembly/pom.xml
index eda54cb..1699ea0 100644
--- a/hbase-assembly/pom.xml
+++ b/hbase-assembly/pom.xml
@@ -269,6 +269,12 @@
<groupId>org.apache.httpcomponents</groupId>
<artifactId>httpcore</artifactId>
</dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-spark-it</artifactId>
+ <version>${project.version}</version>
+ <type>test-jar</type>
+ </dependency>
</dependencies>
<profiles>
http://git-wip-us.apache.org/repos/asf/hbase/blob/79a702d1/hbase-assembly/src/main/assembly/src.xml
----------------------------------------------------------------------
diff --git a/hbase-assembly/src/main/assembly/src.xml b/hbase-assembly/src/main/assembly/src.xml
index befcce0..e5d3faf 100644
--- a/hbase-assembly/src/main/assembly/src.xml
+++ b/hbase-assembly/src/main/assembly/src.xml
@@ -56,6 +56,7 @@
<include>org.apache.hbase:hbase-shaded</include>
<include>org.apache.hbase:hbase-shell</include>
<include>org.apache.hbase:hbase-spark</include>
+ <include>org.apache.hbase:hbase-spark-it</include>
<include>org.apache.hbase:hbase-testing-util</include>
<include>org.apache.hbase:hbase-thrift</include>
</includes>
http://git-wip-us.apache.org/repos/asf/hbase/blob/79a702d1/hbase-spark-it/pom.xml
----------------------------------------------------------------------
diff --git a/hbase-spark-it/pom.xml b/hbase-spark-it/pom.xml
new file mode 100644
index 0000000..94ed27f
--- /dev/null
+++ b/hbase-spark-it/pom.xml
@@ -0,0 +1,361 @@
+<?xml version="1.0"?>
+<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
+<!--
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+-->
+ <modelVersion>4.0.0</modelVersion>
+ <parent>
+ <artifactId>hbase</artifactId>
+ <groupId>org.apache.hbase</groupId>
+ <version>3.0.0-SNAPSHOT</version>
+ <relativePath>..</relativePath>
+ </parent>
+
+ <artifactId>hbase-spark-it</artifactId>
+ <name>Apache HBase - Spark Integration Tests</name>
+ <description>Integration and System tests for HBase</description>
+
+
+ <properties>
+ <spark.version>1.6.0</spark.version>
+ <scala.version>2.10.4</scala.version>
+ <scala.binary.version>2.10</scala.binary.version>
+ <!-- Test inclusion patterns used by failsafe configuration -->
+ <unittest.include>**/Test*.java</unittest.include>
+ <integrationtest.include>**/IntegrationTest*.java</integrationtest.include>
+ <!-- To Run Tests with a particular Xmx Value use -Dfailsafe.Xmx=XXXg -->
+ <failsafe.Xmx>4g</failsafe.Xmx>
+ <!-- To run a single integration test, use -Dit.test=IntegrationTestXXX -->
+ </properties>
+ <build>
+ <pluginManagement>
+ <plugins>
+ <plugin>
+ <groupId>org.apache.maven.plugins</groupId>
+ <artifactId>maven-site-plugin</artifactId>
+ <configuration>
+ <skip>true</skip>
+ </configuration>
+ </plugin>
+ <!-- Make a jar and put the sources in the jar -->
+ <plugin>
+ <groupId>org.apache.maven.plugins</groupId>
+ <artifactId>maven-source-plugin</artifactId>
+ </plugin>
+ <plugin>
+ <!--Make it so assembly:single does nothing in here-->
+ <artifactId>maven-assembly-plugin</artifactId>
+ <configuration>
+ <skipAssembly>true</skipAssembly>
+ </configuration>
+ </plugin>
+ <plugin>
+ <groupId>org.apache.maven.plugins</groupId>
+ <artifactId>maven-failsafe-plugin</artifactId>
+ <version>${surefire.version}</version>
+ <dependencies>
+ <dependency>
+ <groupId>org.apache.maven.surefire</groupId>
+ <artifactId>surefire-junit4</artifactId>
+ <version>${surefire.version}</version>
+ </dependency>
+ </dependencies>
+ <configuration>
+ <includes>
+ <include>${integrationtest.include}</include>
+ </includes>
+ <excludes>
+ <exclude>${unittest.include}</exclude>
+ <exclude>**/*$*</exclude>
+ </excludes>
+ <redirectTestOutputToFile>${test.output.tofile}</redirectTestOutputToFile>
+ <failIfNoTests>false</failIfNoTests>
+ <testFailureIgnore>false</testFailureIgnore>
+ </configuration>
+ <executions>
+ <execution>
+ <id>integration-test</id>
+ <phase>integration-test</phase>
+ <goals>
+ <goal>integration-test</goal>
+ </goals>
+ </execution>
+ <execution>
+ <id>verify</id>
+ <phase>verify</phase>
+ <goals>
+ <goal>verify</goal>
+ </goals>
+ </execution>
+ </executions>
+ </plugin>
+ </plugins>
+ </pluginManagement>
+
+ <plugins>
+ <!-- Run integration tests with mvn verify -->
+ <plugin>
+ <groupId>org.apache.maven.plugins</groupId>
+ <artifactId>maven-failsafe-plugin</artifactId>
+ <configuration>
+ <skip>false</skip>
+ <forkMode>always</forkMode>
+ <!-- TODO: failsafe does timeout, but verify does not fail the build because of the timeout.
+ I believe it is a failsafe bug, we may consider using surefire -->
+ <forkedProcessTimeoutInSeconds>1800</forkedProcessTimeoutInSeconds>
+ <argLine>-enableassertions -Xmx${failsafe.Xmx}
+ -Djava.security.egd=file:/dev/./urandom -XX:+CMSClassUnloadingEnabled
+ -verbose:gc -XX:+PrintCommandLineFlags -XX:+PrintFlagsFinal</argLine>
+ </configuration>
+ </plugin>
+ <plugin>
+ <groupId>org.apache.maven.plugins</groupId>
+ <artifactId>maven-enforcer-plugin</artifactId>
+ <executions>
+ <!-- purposefully have jsr 305 exclusion only warn in this module -->
+ <execution>
+ <id>banned-jsr305</id>
+ <goals>
+ <goal>enforce</goal>
+ </goals>
+ <configuration>
+ <fail>false</fail>
+ </configuration>
+ </execution>
+ <!-- hbase-spark is ok in this modules -->
+ <execution>
+ <id>banned-hbase-spark</id>
+ <goals>
+ <goal>enforce</goal>
+ </goals>
+ <configuration>
+ <skip>true</skip>
+ </configuration>
+ </execution>
+ <execution>
+ <id>banned-scala</id>
+ <goals>
+ <goal>enforce</goal>
+ </goals>
+ <configuration>
+ <skip>true</skip>
+ </configuration>
+ </execution>
+ </executions>
+ </plugin>
+ <plugin>
+ <artifactId>maven-dependency-plugin</artifactId>
+ <executions>
+ <execution>
+ <id>create-mrapp-generated-classpath</id>
+ <phase>generate-test-resources</phase>
+ <goals>
+ <goal>build-classpath</goal>
+ </goals>
+ <configuration>
+ <!-- needed to run the unit test for DS to generate
+ the required classpath that is required in the env
+ of the launch container in the mini cluster
+ -->
+ <outputFile>${project.build.directory}/test-classes/spark-generated-classpath</outputFile>
+ </configuration>
+ </execution>
+ </executions>
+ </plugin>
+
+ </plugins>
+ </build>
+
+ <dependencies>
+ <!-- Intra-project dependencies -->
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-annotations</artifactId>
+ <type>test-jar</type>
+ <scope>test</scope>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-common</artifactId>
+ <type>jar</type>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-protocol</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-client</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-rsgroup</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-rsgroup</artifactId>
+ <type>test-jar</type>
+ <scope>test</scope>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-server</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-spark</artifactId>
+ <version>${project.version}</version>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-it</artifactId>
+ <type>test-jar</type>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-hadoop-compat</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>${compat.module}</artifactId>
+ <version>${project.version}</version>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.hbase</groupId>
+ <artifactId>hbase-testing-util</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>com.google.guava</groupId>
+ <artifactId>guava</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>io.dropwizard.metrics</groupId>
+ <artifactId>metrics-core</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>commons-logging</groupId>
+ <artifactId>commons-logging</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>commons-cli</groupId>
+ <artifactId>commons-cli</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.commons</groupId>
+ <artifactId>commons-math</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>commons-lang</groupId>
+ <artifactId>commons-lang</artifactId>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.htrace</groupId>
+ <artifactId>htrace-core</artifactId>
+ </dependency>
+ <!-- Hadoop needs Netty 3.x at test scope for the minicluster -->
+ <dependency>
+ <groupId>io.netty</groupId>
+ <artifactId>netty</artifactId>
+ <version>${netty.hadoop.version}</version>
+ <scope>test</scope>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.spark</groupId>
+ <artifactId>spark-core_${scala.binary.version}</artifactId>
+ <version>${spark.version}</version>
+ <scope>provided</scope>
+ <exclusions>
+ <exclusion>
+ <!-- make sure wrong scala version is not pulled in -->
+ <groupId>org.scala-lang</groupId>
+ <artifactId>scala-library</artifactId>
+ </exclusion>
+ <exclusion>
+ <!-- make sure wrong scala version is not pulled in -->
+ <groupId>org.scala-lang</groupId>
+ <artifactId>scalap</artifactId>
+ </exclusion>
+ <exclusion>
+ <groupId>com.google.code.findbugs</groupId>
+ <artifactId>jsr305</artifactId>
+ </exclusion>
+ </exclusions>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.spark</groupId>
+ <artifactId>spark-sql_${scala.binary.version}</artifactId>
+ <version>${spark.version}</version>
+ <scope>provided</scope>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.spark</groupId>
+ <artifactId>spark-streaming_${scala.binary.version}</artifactId>
+ <version>${spark.version}</version>
+ <scope>provided</scope>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.spark</groupId>
+ <artifactId>spark-streaming_${scala.binary.version}</artifactId>
+ <version>${spark.version}</version>
+ <type>test-jar</type>
+ <classifier>tests</classifier>
+ <scope>test</scope>
+ </dependency>
+ </dependencies>
+
+ <profiles>
+ <!-- Skip the tests in this module -->
+ <profile>
+ <id>skipIntegrationTests</id>
+ <activation>
+ <property>
+ <name>skipIntegrationTests</name>
+ </property>
+ </activation>
+ <properties>
+ <skipTests>true</skipTests>
+ </properties>
+ </profile>
+ </profiles>
+
+ <reporting>
+ <plugins>
+ <plugin>
+ <groupId>org.apache.maven.plugins</groupId>
+ <artifactId>maven-surefire-report-plugin</artifactId>
+ <version>2.7.2</version>
+ <reportSets>
+ <reportSet>
+ <id>spark-integration-tests</id>
+ <reports>
+ <report>report-only</report>
+ </reports>
+ <configuration>
+ <outputName>failsafe-report</outputName>
+ <reportsDirectories>
+ <reportsDirectory>${project.build.directory}/failsafe-reports</reportsDirectory>
+ </reportsDirectories>
+ </configuration>
+ </reportSet>
+ </reportSets>
+ </plugin>
+ </plugins>
+ </reporting>
+
+</project>
http://git-wip-us.apache.org/repos/asf/hbase/blob/79a702d1/hbase-spark-it/src/test/java/org/apache/hadoop/hbase/spark/IntegrationTestSparkBulkLoad.java
----------------------------------------------------------------------
diff --git a/hbase-spark-it/src/test/java/org/apache/hadoop/hbase/spark/IntegrationTestSparkBulkLoad.java b/hbase-spark-it/src/test/java/org/apache/hadoop/hbase/spark/IntegrationTestSparkBulkLoad.java
new file mode 100644
index 0000000..2d84914
--- /dev/null
+++ b/hbase-spark-it/src/test/java/org/apache/hadoop/hbase/spark/IntegrationTestSparkBulkLoad.java
@@ -0,0 +1,661 @@
+/**
+ *
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.hadoop.hbase.spark;
+
+import com.google.common.collect.Sets;
+import org.apache.commons.cli.CommandLine;
+import org.apache.commons.lang.RandomStringUtils;
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.Path;
+
+import org.apache.hadoop.hbase.Cell;
+import org.apache.hadoop.hbase.CellUtil;
+import org.apache.hadoop.hbase.HBaseConfiguration;
+import org.apache.hadoop.hbase.HBaseTestingUtility;
+import org.apache.hadoop.hbase.HConstants;
+import org.apache.hadoop.hbase.HTableDescriptor;
+import org.apache.hadoop.hbase.IntegrationTestBase;
+import org.apache.hadoop.hbase.IntegrationTestingUtility;
+import org.apache.hadoop.hbase.TableName;
+
+import org.apache.hadoop.hbase.client.Admin;
+import org.apache.hadoop.hbase.client.Connection;
+import org.apache.hadoop.hbase.client.ConnectionFactory;
+import org.apache.hadoop.hbase.client.Consistency;
+import org.apache.hadoop.hbase.client.RegionLocator;
+import org.apache.hadoop.hbase.client.Result;
+import org.apache.hadoop.hbase.client.Scan;
+import org.apache.hadoop.hbase.client.Table;
+
+import org.apache.hadoop.hbase.io.ImmutableBytesWritable;
+
+import org.apache.hadoop.hbase.mapreduce.IntegrationTestBulkLoad;
+import org.apache.hadoop.hbase.mapreduce.LoadIncrementalHFiles;
+import org.apache.hadoop.hbase.util.Bytes;
+import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;
+import org.apache.hadoop.hbase.util.Pair;
+import org.apache.hadoop.hbase.util.RegionSplitter;
+
+import org.apache.hadoop.util.StringUtils;
+import org.apache.hadoop.util.ToolRunner;
+import org.apache.spark.SerializableWritable;
+import org.apache.spark.SparkConf;
+import org.apache.spark.api.java.JavaRDD;
+import org.apache.spark.api.java.JavaSparkContext;
+
+import org.apache.spark.Partitioner;
+
+import org.apache.spark.api.java.function.Function;
+import org.apache.spark.api.java.function.Function2;
+import org.apache.spark.api.java.function.PairFlatMapFunction;
+import org.apache.spark.api.java.function.VoidFunction;
+import org.junit.Test;
+import scala.Tuple2;
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.HashMap;
+import java.util.Iterator;
+import java.util.LinkedList;
+import java.util.List;
+import java.util.Map;
+import java.util.Random;
+import java.util.Set;
+
+/**
+ * Test Bulk Load and Spark on a distributed cluster.
+ * It starts an Spark job that creates linked chains.
+ * This test mimic {@link IntegrationTestBulkLoad} in mapreduce.
+ *
+ * Usage on cluster:
+ * spark-submit --class org.apache.hadoop.hbase.spark.IntegrationTestSparkBulkLoad
+ * HBASE_HOME/lib/hbase-it-XXX-tests.jar -m slowDeterministic -Dhbase.spark.bulkload.chainlength=300
+ */
+public class IntegrationTestSparkBulkLoad extends IntegrationTestBase {
+
+ private static final Log LOG = LogFactory.getLog(IntegrationTestSparkBulkLoad.class);
+
+ // The number of partitions for random generated data
+ private static String BULKLOAD_PARTITIONS_NUM = "hbase.spark.bulkload.partitionsnum";
+ private static int DEFAULT_BULKLOAD_PARTITIONS_NUM = 3;
+
+ private static String BULKLOAD_CHAIN_LENGTH = "hbase.spark.bulkload.chainlength";
+ private static int DEFAULT_BULKLOAD_CHAIN_LENGTH = 200000;
+
+ private static String BULKLOAD_IMPORT_ROUNDS = "hbase.spark.bulkload.importround";
+ private static int DEFAULT_BULKLOAD_IMPORT_ROUNDS = 1;
+
+ private static String CURRENT_ROUND_NUM = "hbase.spark.bulkload.current.roundnum";
+
+ private static String NUM_REPLICA_COUNT_KEY = "hbase.spark.bulkload.replica.countkey";
+ private static int DEFAULT_NUM_REPLICA_COUNT = 1;
+
+ private static String BULKLOAD_TABLE_NAME = "hbase.spark.bulkload.tableName";
+ private static String DEFAULT_BULKLOAD_TABLE_NAME = "IntegrationTestSparkBulkLoad";
+
+ private static String BULKLOAD_OUTPUT_PATH = "hbase.spark.bulkload.output.path";
+
+ private static final String OPT_LOAD = "load";
+ private static final String OPT_CHECK = "check";
+
+ private boolean load = false;
+ private boolean check = false;
+
+ private static final byte[] CHAIN_FAM = Bytes.toBytes("L");
+ private static final byte[] SORT_FAM = Bytes.toBytes("S");
+ private static final byte[] DATA_FAM = Bytes.toBytes("D");
+
+ /**
+ * Running spark job to load data into hbase table
+ */
+ public void runLoad() throws Exception {
+ setupTable();
+ int numImportRounds = getConf().getInt(BULKLOAD_IMPORT_ROUNDS, DEFAULT_BULKLOAD_IMPORT_ROUNDS);
+ LOG.info("Running load with numIterations:" + numImportRounds);
+ for (int i = 0; i < numImportRounds; i++) {
+ runLinkedListSparkJob(i);
+ }
+ }
+
+ /**
+ * Running spark job to create LinkedList for testing
+ * @param iteration iteration th of this job
+ * @throws Exception
+ */
+ public void runLinkedListSparkJob(int iteration) throws Exception {
+ String jobName = IntegrationTestSparkBulkLoad.class.getSimpleName() + " _load " +
+ EnvironmentEdgeManager.currentTime();
+
+ LOG.info("Running iteration " + iteration + "in Spark Job");
+
+ Path output = null;
+ if (conf.get(BULKLOAD_OUTPUT_PATH) == null) {
+ output = util.getDataTestDirOnTestFS(getTablename() + "-" + iteration);
+ } else {
+ output = new Path(conf.get(BULKLOAD_OUTPUT_PATH));
+ }
+
+ SparkConf sparkConf = new SparkConf().setAppName(jobName).setMaster("local");
+ Configuration hbaseConf = new Configuration(getConf());
+ hbaseConf.setInt(CURRENT_ROUND_NUM, iteration);
+ int partitionNum = hbaseConf.getInt(BULKLOAD_PARTITIONS_NUM, DEFAULT_BULKLOAD_PARTITIONS_NUM);
+
+
+ JavaSparkContext jsc = new JavaSparkContext(sparkConf);
+ JavaHBaseContext hbaseContext = new JavaHBaseContext(jsc, hbaseConf);
+
+
+ LOG.info("Partition RDD into " + partitionNum + " parts");
+ List<String> temp = new ArrayList<>();
+ JavaRDD<List<byte[]>> rdd = jsc.parallelize(temp, partitionNum).
+ mapPartitionsWithIndex(new LinkedListCreationMapper(new SerializableWritable<>(hbaseConf)), false);
+
+ hbaseContext.bulkLoad(rdd, getTablename(), new ListToKeyValueFunc(), output.toUri().getPath(),
+ new HashMap<>(), false, HConstants.DEFAULT_MAX_FILE_SIZE);
+
+ try (Connection conn = ConnectionFactory.createConnection(conf);
+ Admin admin = conn.getAdmin();
+ Table table = conn.getTable(getTablename());
+ RegionLocator regionLocator = conn.getRegionLocator(getTablename())) {
+ // Create a new loader.
+ LoadIncrementalHFiles loader = new LoadIncrementalHFiles(conf);
+
+ // Load the HFiles into table.
+ loader.doBulkLoad(output, admin, table, regionLocator);
+ }
+
+
+ // Delete the files.
+ util.getTestFileSystem().delete(output, true);
+ jsc.close();
+ }
+
+ // See mapreduce.IntegrationTestBulkLoad#LinkedListCreationMapper
+ // Used to generate test data
+ public static class LinkedListCreationMapper implements
+ Function2<Integer, Iterator<String>, Iterator<List<byte[]>>> {
+
+ SerializableWritable swConfig = null;
+ private Random rand = new Random();
+
+ public LinkedListCreationMapper(SerializableWritable conf) {
+ this.swConfig = conf;
+ }
+
+ @Override
+ public Iterator<List<byte[]>> call(Integer v1, Iterator v2) throws Exception {
+ Configuration config = (Configuration) swConfig.value();
+ int partitionId = v1.intValue();
+ LOG.info("Starting create List in Partition " + partitionId);
+
+ int partitionNum = config.getInt(BULKLOAD_PARTITIONS_NUM, DEFAULT_BULKLOAD_PARTITIONS_NUM);
+ int chainLength = config.getInt(BULKLOAD_CHAIN_LENGTH, DEFAULT_BULKLOAD_CHAIN_LENGTH);
+ int iterationsNum = config.getInt(BULKLOAD_IMPORT_ROUNDS, DEFAULT_BULKLOAD_IMPORT_ROUNDS);
+ int iterationsCur = config.getInt(CURRENT_ROUND_NUM, 0);
+ List<List<byte[]>> res = new LinkedList<>();
+
+
+ long tempId = partitionId + iterationsCur * partitionNum;
+ long totalPartitionNum = partitionNum * iterationsNum;
+ long chainId = Math.abs(rand.nextLong());
+ chainId = chainId - (chainId % totalPartitionNum) + tempId;
+
+ byte[] chainIdArray = Bytes.toBytes(chainId);
+ long currentRow = 0;
+ long nextRow = getNextRow(0, chainLength);
+ for(long i = 0; i < chainLength; i++) {
+ byte[] rk = Bytes.toBytes(currentRow);
+ // Insert record into a list
+ List<byte[]> tmp1 = Arrays.asList(rk, CHAIN_FAM, chainIdArray, Bytes.toBytes(nextRow));
+ List<byte[]> tmp2 = Arrays.asList(rk, SORT_FAM, chainIdArray, Bytes.toBytes(i));
+ List<byte[]> tmp3 = Arrays.asList(rk, DATA_FAM, chainIdArray, Bytes.toBytes(
+ RandomStringUtils.randomAlphabetic(50)));
+ res.add(tmp1);
+ res.add(tmp2);
+ res.add(tmp3);
+
+ currentRow = nextRow;
+ nextRow = getNextRow(i+1, chainLength);
+ }
+ return res.iterator();
+ }
+
+ /** Returns a unique row id within this chain for this index */
+ private long getNextRow(long index, long chainLength) {
+ long nextRow = Math.abs(new Random().nextLong());
+ // use significant bits from the random number, but pad with index to ensure it is unique
+ // this also ensures that we do not reuse row = 0
+ // row collisions from multiple mappers are fine, since we guarantee unique chainIds
+ nextRow = nextRow - (nextRow % chainLength) + index;
+ return nextRow;
+ }
+ }
+
+
+
+ public static class ListToKeyValueFunc implements
+ Function<List<byte[]>, Pair<KeyFamilyQualifier, byte[]>> {
+ @Override
+ public Pair<KeyFamilyQualifier, byte[]> call(List<byte[]> v1) throws Exception {
+ if (v1 == null || v1.size() != 4) {
+ return null;
+ }
+ KeyFamilyQualifier kfq = new KeyFamilyQualifier(v1.get(0), v1.get(1), v1.get(2));
+
+ return new Pair<>(kfq, v1.get(3));
+ }
+ }
+
+ /**
+ * After adding data to the table start a mr job to
+ * @throws IOException
+ * @throws ClassNotFoundException
+ * @throws InterruptedException
+ */
+ public void runCheck() throws Exception {
+ LOG.info("Running check");
+ String jobName = IntegrationTestSparkBulkLoad.class.getSimpleName() + "_check" + EnvironmentEdgeManager.currentTime();
+
+ SparkConf sparkConf = new SparkConf().setAppName(jobName).setMaster("local");
+ Configuration hbaseConf = new Configuration(getConf());
+ JavaSparkContext jsc = new JavaSparkContext(sparkConf);
+ JavaHBaseContext hbaseContext = new JavaHBaseContext(jsc, hbaseConf);
+
+ Scan scan = new Scan();
+ scan.addFamily(CHAIN_FAM);
+ scan.addFamily(SORT_FAM);
+ scan.setMaxVersions(1);
+ scan.setCacheBlocks(false);
+ scan.setBatch(1000);
+ int replicaCount = conf.getInt(NUM_REPLICA_COUNT_KEY, DEFAULT_NUM_REPLICA_COUNT);
+ if (replicaCount != DEFAULT_NUM_REPLICA_COUNT) {
+ scan.setConsistency(Consistency.TIMELINE);
+ }
+
+ // 1. Using TableInputFormat to get data from HBase table
+ // 2. Mimic LinkedListCheckingMapper in mapreduce.IntegrationTestBulkLoad
+ // 3. Sort LinkKey by its order ID
+ // 4. Group LinkKey if they have same chainId, and repartition RDD by NaturalKeyPartitioner
+ // 5. Check LinkList in each Partition using LinkedListCheckingFlatMapFunc
+ hbaseContext.hbaseRDD(getTablename(), scan).flatMapToPair(new LinkedListCheckingFlatMapFunc())
+ .sortByKey()
+ .combineByKey(new createCombinerFunc(), new mergeValueFunc(), new mergeCombinersFunc(),
+ new NaturalKeyPartitioner(new SerializableWritable<>(hbaseConf)))
+ .foreach(new LinkedListCheckingForeachFunc(new SerializableWritable<>(hbaseConf)));
+ jsc.close();
+ }
+
+ private void runCheckWithRetry() throws Exception {
+ try {
+ runCheck();
+ } catch (Throwable t) {
+ LOG.warn("Received " + StringUtils.stringifyException(t));
+ LOG.warn("Running the check MR Job again to see whether an ephemeral problem or not");
+ runCheck();
+ throw t; // we should still fail the test even if second retry succeeds
+ }
+ // everything green
+ }
+
+ /**
+ * PairFlatMapFunction used to transfer <Row, Result> to Tuple <SparkLinkKey, SparkLinkChain>
+ */
+ public static class LinkedListCheckingFlatMapFunc implements
+ PairFlatMapFunction<Tuple2<ImmutableBytesWritable, Result>, SparkLinkKey, SparkLinkChain> {
+
+ @Override
+ public Iterable<Tuple2<SparkLinkKey, SparkLinkChain>> call(Tuple2<ImmutableBytesWritable, Result> v)
+ throws Exception {
+ Result value = v._2();
+ long longRk = Bytes.toLong(value.getRow());
+ List<Tuple2<SparkLinkKey, SparkLinkChain>> list = new LinkedList<>();
+
+ for (Map.Entry<byte[], byte[]> entry : value.getFamilyMap(CHAIN_FAM).entrySet()) {
+ long chainId = Bytes.toLong(entry.getKey());
+ long next = Bytes.toLong(entry.getValue());
+ Cell c = value.getColumnCells(SORT_FAM, entry.getKey()).get(0);
+ long order = Bytes.toLong(CellUtil.cloneValue(c));
+ Tuple2<SparkLinkKey, SparkLinkChain> tuple2 =
+ new Tuple2<>(new SparkLinkKey(chainId, order), new SparkLinkChain(longRk, next));
+ list.add(tuple2);
+ }
+ return list;
+ }
+ }
+
+ public static class createCombinerFunc implements
+ Function<SparkLinkChain, List<SparkLinkChain>> {
+ @Override
+ public List<SparkLinkChain> call(SparkLinkChain v1) throws Exception {
+ List<SparkLinkChain> list = new LinkedList<>();
+ list.add(v1);
+ return list;
+ }
+ }
+
+ public static class mergeValueFunc implements
+ Function2<List<SparkLinkChain>, SparkLinkChain, List<SparkLinkChain>> {
+ @Override
+ public List<SparkLinkChain> call(List<SparkLinkChain> v1, SparkLinkChain v2) throws Exception {
+ if (v1 == null)
+ v1 = new LinkedList<>();
+ v1.add(v2);
+ return v1;
+ }
+ }
+
+ public static class mergeCombinersFunc implements
+ Function2<List<SparkLinkChain>, List<SparkLinkChain>, List<SparkLinkChain>> {
+ @Override
+ public List<SparkLinkChain> call(List<SparkLinkChain> v1, List<SparkLinkChain> v2) throws Exception {
+ v1.addAll(v2);
+ return v1;
+ }
+ }
+
+ /**
+ * Class to figure out what partition to send a link in the chain to. This is based upon
+ * the linkKey's ChainId.
+ */
+ public static class NaturalKeyPartitioner extends Partitioner {
+
+ private int numPartions = 0;
+ public NaturalKeyPartitioner(SerializableWritable swConf) {
+ Configuration hbaseConf = (Configuration) swConf.value();
+ numPartions = hbaseConf.getInt(BULKLOAD_PARTITIONS_NUM, DEFAULT_BULKLOAD_PARTITIONS_NUM);
+
+ }
+
+ @Override
+ public int numPartitions() {
+ return numPartions;
+ }
+
+ @Override
+ public int getPartition(Object key) {
+ if (!(key instanceof SparkLinkKey))
+ return -1;
+ int hash = ((SparkLinkKey) key).getChainId().hashCode();
+ return Math.abs(hash % numPartions);
+
+ }
+ }
+
+ /**
+ * Sort all LinkChain for one LinkKey, and test List<LinkChain>
+ */
+ public static class LinkedListCheckingForeachFunc
+ implements VoidFunction<Tuple2<SparkLinkKey, List<SparkLinkChain>>> {
+
+ private SerializableWritable swConf = null;
+
+ public LinkedListCheckingForeachFunc(SerializableWritable conf) {
+ swConf = conf;
+ }
+
+ @Override
+ public void call(Tuple2<SparkLinkKey, List<SparkLinkChain>> v1) throws Exception {
+ long next = -1L;
+ long prev = -1L;
+ long count = 0L;
+
+ SparkLinkKey key = v1._1();
+ List<SparkLinkChain> values = v1._2();
+
+ for (SparkLinkChain lc : values) {
+
+ if (next == -1) {
+ if (lc.getRk() != 0L) {
+ String msg = "Chains should all start at rk 0, but read rk " + lc.getRk()
+ + ". Chain:" + key.getChainId() + ", order:" + key.getOrder();
+ throw new RuntimeException(msg);
+ }
+ next = lc.getNext();
+ } else {
+ if (next != lc.getRk()) {
+ String msg = "Missing a link in the chain. Prev rk " + prev + " was, expecting "
+ + next + " but got " + lc.getRk() + ". Chain:" + key.getChainId()
+ + ", order:" + key.getOrder();
+ throw new RuntimeException(msg);
+ }
+ prev = lc.getRk();
+ next = lc.getNext();
+ }
+ count++;
+ }
+ Configuration hbaseConf = (Configuration) swConf.value();
+ int expectedChainLen = hbaseConf.getInt(BULKLOAD_CHAIN_LENGTH, DEFAULT_BULKLOAD_CHAIN_LENGTH);
+ if (count != expectedChainLen) {
+ String msg = "Chain wasn't the correct length. Expected " + expectedChainLen + " got "
+ + count + ". Chain:" + key.getChainId() + ", order:" + key.getOrder();
+ throw new RuntimeException(msg);
+ }
+ }
+ }
+
+ /**
+ * Writable class used as the key to group links in the linked list.
+ *
+ * Used as the key emited from a pass over the table.
+ */
+ public static class SparkLinkKey implements java.io.Serializable, Comparable<SparkLinkKey> {
+
+ private Long chainId;
+ private Long order;
+
+ public Long getOrder() {
+ return order;
+ }
+
+ public Long getChainId() {
+ return chainId;
+ }
+
+ public SparkLinkKey(long chainId, long order) {
+ this.chainId = chainId;
+ this.order = order;
+ }
+
+ @Override
+ public int hashCode() {
+ return this.getChainId().hashCode();
+ }
+
+ @Override
+ public boolean equals(Object other) {
+ if (!(other instanceof SparkLinkKey))
+ return false;
+ SparkLinkKey otherKey = (SparkLinkKey) other;
+ return this.getChainId().equals(otherKey.getChainId());
+ }
+
+ @Override
+ public int compareTo(SparkLinkKey other) {
+ int res = getChainId().compareTo(other.getChainId());
+ if (res == 0)
+ res= getOrder().compareTo(other.getOrder());
+ return res;
+ }
+ }
+
+ /**
+ * Writable used as the value emitted from a pass over the hbase table.
+ */
+ public static class SparkLinkChain implements java.io.Serializable, Comparable<SparkLinkChain>{
+
+ public Long getNext() {
+ return next;
+ }
+
+ public Long getRk() {
+ return rk;
+ }
+
+
+ public SparkLinkChain(Long rk, Long next) {
+ this.rk = rk;
+ this.next = next;
+ }
+
+ private Long rk;
+ private Long next;
+
+ @Override
+ public int compareTo(SparkLinkChain linkChain) {
+ int res = getRk().compareTo(linkChain.getRk());
+ if (res == 0) {
+ res = getNext().compareTo(linkChain.getNext());
+ }
+ return res;
+ }
+
+ @Override
+ public int hashCode() {
+ return getRk().hashCode() ^ getNext().hashCode();
+ }
+
+ @Override
+ public boolean equals(Object other) {
+ if (!(other instanceof SparkLinkChain))
+ return false;
+ SparkLinkChain otherKey = (SparkLinkChain) other;
+ return this.getRk().equals(otherKey.getRk()) && this.getNext().equals(otherKey.getNext());
+ }
+ }
+
+
+ /**
+ * Allow the scan to go to replica, this would not affect the runCheck()
+ * Since data are BulkLoaded from HFile into table
+ * @throws IOException
+ * @throws InterruptedException
+ */
+ private void installSlowingCoproc() throws IOException, InterruptedException {
+ int replicaCount = conf.getInt(NUM_REPLICA_COUNT_KEY, DEFAULT_NUM_REPLICA_COUNT);
+ if (replicaCount == DEFAULT_NUM_REPLICA_COUNT) return;
+
+ TableName t = getTablename();
+ Admin admin = util.getAdmin();
+ HTableDescriptor desc = admin.getTableDescriptor(t);
+ desc.addCoprocessor(IntegrationTestBulkLoad.SlowMeCoproScanOperations.class.getName());
+ HBaseTestingUtility.modifyTableSync(admin, desc);
+ }
+
+ @Test
+ public void testBulkLoad() throws Exception {
+ runLoad();
+ installSlowingCoproc();
+ runCheckWithRetry();
+ }
+
+
+ private byte[][] getSplits(int numRegions) {
+ RegionSplitter.UniformSplit split = new RegionSplitter.UniformSplit();
+ split.setFirstRow(Bytes.toBytes(0L));
+ split.setLastRow(Bytes.toBytes(Long.MAX_VALUE));
+ return split.split(numRegions);
+ }
+
+ private void setupTable() throws IOException, InterruptedException {
+ if (util.getAdmin().tableExists(getTablename())) {
+ util.deleteTable(getTablename());
+ }
+
+ util.createTable(
+ getTablename(),
+ new byte[][]{CHAIN_FAM, SORT_FAM, DATA_FAM},
+ getSplits(16)
+ );
+
+ int replicaCount = conf.getInt(NUM_REPLICA_COUNT_KEY, DEFAULT_NUM_REPLICA_COUNT);
+ if (replicaCount == DEFAULT_NUM_REPLICA_COUNT) return;
+
+ TableName t = getTablename();
+ HBaseTestingUtility.setReplicas(util.getAdmin(), t, replicaCount);
+ }
+
+ @Override
+ public void setUpCluster() throws Exception {
+ util = getTestingUtil(getConf());
+ util.initializeCluster(1);
+ int replicaCount = getConf().getInt(NUM_REPLICA_COUNT_KEY, DEFAULT_NUM_REPLICA_COUNT);
+ if (LOG.isDebugEnabled() && replicaCount != DEFAULT_NUM_REPLICA_COUNT) {
+ LOG.debug("Region Replicas enabled: " + replicaCount);
+ }
+
+ // Scale this up on a real cluster
+ if (util.isDistributedCluster()) {
+ util.getConfiguration().setIfUnset(BULKLOAD_PARTITIONS_NUM, String.valueOf(DEFAULT_BULKLOAD_PARTITIONS_NUM));
+ util.getConfiguration().setIfUnset(BULKLOAD_IMPORT_ROUNDS, "1");
+ } else {
+ util.startMiniMapReduceCluster();
+ }
+ }
+
+ @Override
+ protected void addOptions() {
+ super.addOptions();
+ super.addOptNoArg(OPT_CHECK, "Run check only");
+ super.addOptNoArg(OPT_LOAD, "Run load only");
+ }
+
+ @Override
+ protected void processOptions(CommandLine cmd) {
+ super.processOptions(cmd);
+ check = cmd.hasOption(OPT_CHECK);
+ load = cmd.hasOption(OPT_LOAD);
+ }
+
+ @Override
+ public int runTestFromCommandLine() throws Exception {
+ if (load) {
+ runLoad();
+ } else if (check) {
+ installSlowingCoproc();
+ runCheckWithRetry();
+ } else {
+ testBulkLoad();
+ }
+ return 0;
+ }
+
+ @Override
+ public TableName getTablename() {
+ return getTableName(getConf());
+ }
+
+ public static TableName getTableName(Configuration conf) {
+ return TableName.valueOf(conf.get(BULKLOAD_TABLE_NAME, DEFAULT_BULKLOAD_TABLE_NAME));
+ }
+
+ @Override
+ protected Set<String> getColumnFamilies() {
+ return Sets.newHashSet(Bytes.toString(CHAIN_FAM) , Bytes.toString(DATA_FAM),
+ Bytes.toString(SORT_FAM));
+ }
+
+ public static void main(String[] args) throws Exception {
+ Configuration conf = HBaseConfiguration.create();
+ IntegrationTestingUtility.setUseDistributedCluster(conf);
+ int status = ToolRunner.run(conf, new IntegrationTestSparkBulkLoad(), args);
+ System.exit(status);
+ }
+}
http://git-wip-us.apache.org/repos/asf/hbase/blob/79a702d1/pom.xml
----------------------------------------------------------------------
diff --git a/pom.xml b/pom.xml
index 5e9e175..329c468 100644
--- a/pom.xml
+++ b/pom.xml
@@ -86,6 +86,7 @@
<module>hbase-archetypes</module>
<module>hbase-metrics-api</module>
<module>hbase-metrics</module>
+ <module>hbase-spark-it</module>
</modules>
<!--Add apache snapshots in case we want to use unreleased versions of plugins:
e.g. surefire 2.18-SNAPSHOT-->