Partager via


Acheminer les messages Apache Kafka® vers Azure Cosmos DB pour Apache Cassandra, avec Apache Flink® sur HDInsight sur AKS

Important

Azure HDInsight sur AKS a été mis hors service le 31 janvier 2025. En savoir plus avec cette annonce.

Vous devez migrer vos charges de travail vers Microsoft Fabric ou un produit Azure équivalent pour éviter l’arrêt brusque de vos charges de travail.

Important

Cette fonctionnalité est actuellement en préversion. Les Conditions d’utilisation supplémentaires pour les préversions Microsoft Azure incluent des termes juridiques supplémentaires qui s’appliquent aux fonctionnalités Azure en version bêta, en préversion ou qui ne sont pas encore publiées en disponibilité générale. Pour plus d'informations sur cette préversion spécifique, consultez les informations sur Azure HDInsight sur AKS en préversion . Pour des questions ou des suggestions de fonctionnalités, envoyez une demande sur AskHDInsight avec les détails et suivez-nous pour plus de mises à jour sur Communauté Azure HDInsight.

Cet exemple utilise Apache Flink pour récepteur HDInsight pour les messages d’Apache Kafka dans Azure Cosmos DB pour Apache Cassandra.

Cet exemple est important lorsque les ingénieurs préfèrent les données agrégées en temps réel pour l’analyse. Avec l’accès aux données agrégées historiques, vous pouvez créer des modèles Machine Learning (ML) pour générer des insights ou des actions. Vous pouvez également ingérer des données IoT dans Apache Flink pour agréger des données en temps réel et les stocker dans Apache Cassandra.

Conditions préalables

Azure Cosmos DB pour Apache Cassandra

Azure Cosmos DB pour Apache Cassandra peut être utilisé comme magasin de données pour les applications écrites pour Apache Cassandra. Cette compatibilité signifie qu’en utilisant des pilotes Apache existants conformes à CQLv4, votre application Cassandra existante peut désormais communiquer avec l’API pour Cassandra.

Pour plus d’informations, consultez les liens suivants.

Capture d’écran montrant comment créer Azure Cosmos DB pour Apache Cassandra sur le portail Azure.

Obtenir des informations d'identification les utilise dans le code source de Stream :

Capture d’écran montrant comment obtenir des informations d’identification sur le code source de flux.

Implémentation

sur une machine virtuelle Ubuntu, préparons l’environnement de développement

Clonage du référentiel des exemples Azure

Reportez-vous au fichier README de GitHub pour télécharger Maven et clonez ce référentiel en utilisant Azure-Samples/azure-cosmos-db-cassandra-java-getting-started.git depuis les exemples Azure .

Mise à jour du projet maven pour Cassandra

Accédez au dossier de projet maven azure-cosmos-db-cassandra-java-getting-started-main et mettez à jour les modifications requises pour cet exemple.

maven pom.xml

<?xml version="1.0" encoding="UTF-8"?>
<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">
  <modelVersion>4.0.0</modelVersion>
 
  <groupId>com.azure.cosmosdb.cassandra</groupId>
  <artifactId>cosmosdb-cassandra-examples</artifactId>
   <version>1.0-SNAPSHOT</version>
  <dependencies>
            <dependency>
          <groupId>org.apache.flink</groupId>
          <artifactId>flink-java</artifactId>
           <version>1.17.0</version>
       </dependency>
       <!-- https://mvnrepository.com/artifact/org.apache.flink/flink-streaming-java -->
       <dependency>
          <groupId>org.apache.flink</groupId>
          <artifactId>flink-streaming-java</artifactId>
           <version>1.17.0</version>
       </dependency>
       <!-- https://mvnrepository.com/artifact/org.apache.flink/flink-clients -->
       <dependency>
          <groupId>org.apache.flink</groupId>
          <artifactId>flink-clients</artifactId>
           <version>1.17.0</version>
       </dependency>
       <dependency>
          <groupId>org.apache.flink</groupId>
          <artifactId>flink-connector-files</artifactId>
           <version>1.17.0</version>
       </dependency>
       <dependency>
          <groupId>org.apache.flink</groupId>
          <artifactId>flink-connector-kafka</artifactId>
           <version>1.17.0</version>
       </dependency>
       <dependency>
          <groupId>com.datastax.cassandra</groupId>
          <artifactId>cassandra-driver-core</artifactId>
          <version>3.3.0</version>
       </dependency>
       <dependency>
          <groupId>com.datastax.cassandra</groupId>
          <artifactId>cassandra-driver-mapping</artifactId>
          <version>3.1.4</version>
       </dependency>
       <dependency>
          <groupId>com.datastax.cassandra</groupId>
          <artifactId>cassandra-driver-extras</artifactId>
          <version>3.1.4</version>
       </dependency>
       <dependency>
          <groupId>org.slf4j</groupId>
          <artifactId>slf4j-api</artifactId>
          <version>1.7.5</version>
       </dependency>
       <dependency>
          <groupId>org.slf4j</groupId>
          <artifactId>slf4j-log4j12</artifactId>
          <version>1.7.5</version>
       </dependency>
   </dependencies>
 
   <build>
       <plugins>
           <plugin>
              <artifactId>maven-assembly-plugin</artifactId>
               <configuration>
                   <descriptorRefs>
                      <descriptorRef>jar-with-dependencies</descriptorRef>
                   </descriptorRefs>
                  <finalName>cosmosdb-cassandra-examples</finalName>
                  <appendAssemblyId>false</appendAssemblyId>
               </configuration>
               <executions>
                   <execution>
                      <id>make-assembly</id>
                      <phase>package</phase>
                       <goals>
                          <goal>single</goal>
                       </goals>
                   </execution>
               </executions>
           </plugin>
           <plugin>
              <groupId>org.apache.maven.plugins</groupId>
              <artifactId>maven-compiler-plugin</artifactId>
               <configuration>
                  <source>1.8</source>
                  <target>1.8</target>
               </configuration>
           </plugin>
       </plugins>
   </build>
 
</project>

Cosmos DB pour la configuration de connexion d’Apache Cassandra

Vous devez mettre à jour votre nom d’hôte et votre nom d’utilisateur et vos clés dans l’extrait de code ci-dessous.

root@flinkvm:/home/flinkvm/azure-cosmos-db-cassandra-java-getting-started-main/src/main/resources# cat config.properties
###Cassandra endpoint details on cosmosdb
cassandra_host=<update-host-name>.cassandra.cosmos.azure.com
cassandra_port = 10350
cassandra_username=<update-user-name>
cassandra_password=mxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
#ssl_keystore_file_path=<SSL key store file location>
#ssl_keystore_password=<SSL key store password>

structure source

root@flinkvm:/home/flinkvm/azure-cosmos-db-cassandra-java-getting-started-main/src/main/java/com/azure/cosmosdb/cassandra# ll
total 24
drwxr-xr-x 5 root root 4096 May 12 12:46 ./
drwxr-xr-x 3 root root 4096 Apr  9  2020 ../
-rw-r--r-- 1 root root 1105 Apr  9  2020 User.java
drwxr-xr-x 2 root root 4096 May 15 03:53 examples/
drwxr-xr-x 2 root root 4096 Apr  9  2020 repository/
drwxr-xr-x 2 root root 4096 May 15 02:43 util/

dossier utilitaire
CassandraUtils.java

Note

La modification ssl_keystore_file_path dépend de l’emplacement du certificat Java. Cluster Apache Flink sur HDInsight sur AKS, le chemin d’accès est /usr/lib/jvm/msopenjdk-11-jre/lib/security

package com.azure.cosmosdb.cassandra.util;

import com.datastax.driver.core.*;

import javax.net.ssl.*;
import java.io.File;
import java.io.FileInputStream;
import java.io.InputStream;
import java.security.*;

/**
 * Cassandra utility class to handle the Cassandra Sessions
 */
public class CassandraUtils {

    private Cluster cluster;
    private Configurations config = new Configurations();
    private String cassandraHost = "<cassandra-host-ip>";
    private int cassandraPort = 10350;
    private String cassandraUsername = "localhost";
    private String cassandraPassword = "<cassandra-password>";
    private File sslKeyStoreFile = null;
    private String sslKeyStorePassword = "<keystore-password>";


    /**
     * This method creates a Cassandra Session based on the end-point details given in config.properties.
     * This method validates the SSL certificate based on ssl_keystore_file_path & ssl_keystore_password properties.
     * If ssl_keystore_file_path & ssl_keystore_password are not given then it uses 'cacerts' from JDK.
     * @return Session Cassandra Session
     */
    public Session getSession() {

        try {
            //Load cassandra endpoint details from config.properties
            loadCassandraConnectionDetails();

            final KeyStore keyStore = KeyStore.getInstance("JKS");
            try (final InputStream is = new FileInputStream(sslKeyStoreFile)) {
                keyStore.load(is, sslKeyStorePassword.toCharArray());
            }

            final KeyManagerFactory kmf = KeyManagerFactory.getInstance(KeyManagerFactory
                    .getDefaultAlgorithm());
            kmf.init(keyStore, sslKeyStorePassword.toCharArray());
            final TrustManagerFactory tmf = TrustManagerFactory.getInstance(TrustManagerFactory
                    .getDefaultAlgorithm());
            tmf.init(keyStore);

            // Creates a socket factory for HttpsURLConnection using JKS contents.
            final SSLContext sc = SSLContext.getInstance("TLSv1.2");
            sc.init(kmf.getKeyManagers(), tmf.getTrustManagers(), new java.security.SecureRandom());

            JdkSSLOptions sslOptions = RemoteEndpointAwareJdkSSLOptions.builder()
                    .withSSLContext(sc)
                    .build();
            cluster = Cluster.builder()
                    .addContactPoint(cassandraHost)
                    .withPort(cassandraPort)
                    .withCredentials(cassandraUsername, cassandraPassword)
                    .withSSL(sslOptions)
                    .build();

            return cluster.connect();
        } catch (Exception ex) {
            ex.printStackTrace();
        }
        return null;
    }

    public Cluster getCluster() {
        return cluster;
    }

    /**
     * Closes the cluster and Cassandra session
     */
    public void close() {
        cluster.close();
    }

    /**
     * Loads Cassandra end-point details from config.properties.
     * @throws Exception
     */
    private void loadCassandraConnectionDetails() throws Exception {
        cassandraHost = config.getProperty("cassandra_host");
        cassandraPort = Integer.parseInt(config.getProperty("cassandra_port"));
        cassandraUsername = config.getProperty("cassandra_username");
        cassandraPassword = config.getProperty("cassandra_password");
        String ssl_keystore_file_path = config.getProperty("ssl_keystore_file_path");
        String ssl_keystore_password = config.getProperty("ssl_keystore_password");

        // If ssl_keystore_file_path, build the path using JAVA_HOME directory.
        if (ssl_keystore_file_path == null || ssl_keystore_file_path.isEmpty()) {
            String javaHomeDirectory = System.getenv("JAVA_HOME");
            if (javaHomeDirectory == null || javaHomeDirectory.isEmpty()) {
                throw new Exception("JAVA_HOME not set");
            }
            ssl_keystore_file_path = new StringBuilder(javaHomeDirectory).append("/lib/security/cacerts").toString();
        }

        sslKeyStorePassword = (ssl_keystore_password != null && !ssl_keystore_password.isEmpty()) ?
                ssl_keystore_password : sslKeyStorePassword;

        sslKeyStoreFile = new File(ssl_keystore_file_path);

        if (!sslKeyStoreFile.exists() || !sslKeyStoreFile.canRead()) {
            throw new Exception(String.format("Unable to access the SSL Key Store file from %s", ssl_keystore_file_path));
        }
    }
}

Configurations.java

package com.azure.cosmosdb.cassandra.util;

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.IOException;
import java.io.InputStream;
import java.util.Properties;

/**
 * Configuration utility to read the configurations from properties file
 */
public class Configurations {
    private static final Logger LOGGER = LoggerFactory.getLogger(Configurations.class);
    private static String PROPERTY_FILE = "config.properties";
    private static Properties prop = null;

    private void loadProperties() throws IOException {
        InputStream input = getClass().getClassLoader().getResourceAsStream(PROPERTY_FILE);
        if (input == null) {
            LOGGER.error("Sorry, unable to find {}", PROPERTY_FILE);
            return;
        }
        prop = new Properties();
        prop.load(input);
    }

    public String getProperty(String propertyName) throws IOException {
        if (prop == null) {
            loadProperties();
        }
        return prop.getProperty(propertyName);

    }
}

dossier Exemples

CassandraSink.java

package com.azure.cosmosdb.cassandra.examples;

import com.datastax.driver.core.PreparedStatement;
import com.datastax.driver.core.Session;
import org.apache.flink.api.java.tuple.Tuple3;
import org.apache.flink.streaming.api.functions.sink.SinkFunction;
import com.azure.cosmosdb.cassandra.repository.UserRepository;
import com.azure.cosmosdb.cassandra.util.CassandraUtils;

public class CassandraSink implements SinkFunction<Tuple3<Integer, String, String>> {

    @Override
    public void invoke(Tuple3<Integer, String, String> value, Context context) throws Exception {

            CassandraUtils utils = new CassandraUtils();
            Session cassandraSession = utils.getSession();
            try {
                UserRepository repository = new UserRepository(cassandraSession);

                //Insert rows into user table
                PreparedStatement preparedStatement = repository.prepareInsertStatement();
                repository.insertUser(preparedStatement, value.f0, value.f1, value.f2);

                } finally {
                        if (null != utils) utils.close();
                        if (null != cassandraSession) cassandraSession.close();
                        }
            }
}

classe principale : CassandraDemo.java

Remarque

  • Remplacer les adresses IP du répartiteur Kafka par vos adresses IP de répartiteur de cluster Kafka
  • Préparer le sujet
    • utilisateur /usr/hdp/current/kafka-broker/bin/kafka-topics.sh --create --replication-factor 2 --partitions 3 --topic user --bootstrap-server wn0-flinkd:9092
package com.azure.cosmosdb.cassandra.examples;

import org.apache.flink.api.common.eventtime.WatermarkStrategy;
import org.apache.flink.api.common.serialization.SimpleStringSchema;
import org.apache.flink.api.common.typeinfo.Types;
import org.apache.flink.api.java.tuple.Tuple3;
import org.apache.flink.connector.kafka.source.KafkaSource;
import org.apache.flink.connector.kafka.source.enumerator.initializer.OffsetsInitializer;
import org.apache.flink.streaming.api.datastream.DataStream;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;

public class CassandraDemo {
    public static void main(String[] args) throws Exception {
        StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment().setParallelism(1);

        // 1. read kafka message as stream input, update the broker IPs from your Kafka setup
        String brokers = "<update-broker-ips>:9092,<update-broker-ips>:9092,<update-broker-ips>:9092";

        KafkaSource<String> source = KafkaSource.<String>builder()
                .setBootstrapServers(brokers)
                .setTopics("user")
                .setGroupId("my-group")
                .setStartingOffsets(OffsetsInitializer.earliest())
                .setValueOnlyDeserializer(new SimpleStringSchema())
                .build();

        DataStream<String> kafka = env.fromSource(source, WatermarkStrategy.noWatermarks(), "Kafka Source");
        kafka.print();

        DataStream<Tuple3<Integer,String,String>> dataStream = kafka.map(line-> {
            String[] fields = line.split(",");
            int v1 = Integer.parseInt(fields[0]);
            Tuple3<Integer,String,String> tuple3 = Tuple3.of(v1,fields[1],fields[2]);
            return tuple3;
        }).returns(Types.TUPLE(Types.INT,Types.STRING,Types.STRING));


        dataStream.addSink(new CassandraSink());

        // 4. run stream
        env.execute("sink Kafka to Cosmos DB for Apache Cassandra");
    }
}

Construction du projet

Exécutez mvn clean install à partir du dossier azure-cosmos-db-cassandra-java-getting-started-main pour générer le projet. Cette commande génère cosmosdb-cassandra-examples.jar sous dossier cible.

root@flinkvm:/home/flinkvm/azure-cosmos-db-cassandra-java-getting-started-main/target# ll
total 91156
drwxr-xr-x 7 root root     4096 May 15 03:54 ./
drwxr-xr-x 7 root root     4096 May 15 03:54 ../
drwxr-xr-x 2 root root     4096 May 15 03:54 archive-tmp/
drwxr-xr-x 3 root root     4096 May 15 03:54 classes/
-rw-r--r-- 1 root root    15542 May 15 03:54 cosmosdb-cassandra-examples-1.0-SNAPSHOT.jar
-rw-r--r-- 1 root root 93290819 May 15 03:54 cosmosdb-cassandra-examples.jar
drwxr-xr-x 3 root root     4096 May 15 03:54 generated-sources/
drwxr-xr-x 2 root root     4096 May 15 03:54 maven-archiver/
drwxr-xr-x 3 root root     4096 May 15 03:54 maven-status/

Charger un fichier jar dans le stockage Azure et utiliser wget dans webssh

msdata@pod-0 [ ~ ]$ ls -l cosmosdb-cassandra-examples.jar
-rw-r----- 1 msdata msdata 93290819 May 15 04:02 cosmosdb-cassandra-examples.jar

Préparation de la KeyStore et de la table Cosmos DB

Exécutez la classe UserProfile dans /azure-cosmos-db-cassandra-java-getting-started-main/src/main/java/com/azure/cosmosdb/cassandra/examples pour créer le magasin de clés et la table d’Azure Cosmos DB.

bin/flink run -c com.azure.cosmosdb.cassandra.examples.UserProfile -j cosmosdb-cassandra-examples.jar

Intégrer les sujets Kafka dans Cosmos DB pour Apache Cassandra

Exécutez la classe CassandraDemo pour transférer le sujet Kafka vers Cosmos DB pour Apache Cassandra.

bin/flink run -c com.azure.cosmosdb.cassandra.examples.CassandraDemo -j cosmosdb-cassandra-examples.jar

Capture d’écran montrant comment exécuter la démonstration Cassandra.

Contrôlez le travail sur l'interface utilisateur de Flink dans HDInsight sur le cluster AKS.

Capture d’écran montrant comment vérifier la tâche sur HDInsight sur l'UI AKS Flink.

Production de messages dans Kafka

Produire un message dans la rubrique Kafka.

sshuser@hn0-flinkd:~$ cat user.py
import time
from datetime import datetime
import random

user_set = [
        'John',
        'Mike',
        'Lucy',
        'Tom',
        'Machael',
        'Lily',
        'Zark',
        'Tim',
        'Andrew',
        'Pick',
        'Sean',
        'Luke',
        'Chunck'
]

city_set = [
        'Atmore',
        'Auburn',
        'Bessemer',
        'Birmingham',
        'Chickasaw',
        'Clanton',
        'Decatur',
        'Florence',
        'Greenville',
        'Jasper',
        'Huntsville',
        'Homer',
        'Homer'
]

def main():
        while True:
                unique_id = str(int(time.time()))
                if random.randrange(10) < 4:
                        city = random.choice(city_set[:3])
                else:
                        city = random.choice(city_set)
                user = random.choice(user_set)
                print(unique_id + "," + user + "," + city )
                time.sleep(1)

if __name__ == "__main__":
    main()
sshuser@hn0-flinkd:~$ python user.py | /usr/hdp/current/kafka-broker/bin/kafka-console-producer.sh --bootstrap-server wn0-flinkd:9092 --topic user &
[2] 11516

Vérifier la table sur Cosmos DB pour Apache Cassandra sur le portail Azure

capture d’écran montrant Cosmos DB pour Apache Cassandra sur le portail Azure.

Préférences