Kafka自定义分区机制

---

1.如何自定义分区机制

若需要使用自定义分区机制，需要完成两件事：
1)在 producer 程序中创建一个类，实现 org.apache.kafka.clients.producer.Partitioner 接口主要分区逻辑在 Partitioner.partition中实现。
2)在用于构造KafkaProducer的Properties对象中设置 partitioner.class 参数。

2.示例

假设我们的消息中有一些消息是用于审计功能的，这类消息的 key 会被固定地分配一个字符串“audit”。我们想要让这类消息发送到 topic 的最后一个分区上，便于后续统一处理，而对于相同 topic 下的其他消息则采用随机发送的策略发送到其他分区上。那么现在就可以这样来实现自定义的分区策略，如下列代码所示:

import org.apache.kafka.clients.producer.Partitioner;
import org.apache.kafka.common.Cluster;
import org.apache.kafka.common.PartitionInfo;
import java.util.List;
import java.util.Map;
import java.util.Random;
public class AuditPartitioner implements Partitioner {
    private Random random;
    @Override
    public void configure(Map<String, ?> map) {
        //该方法实现必要资源的初始化工作
        random= new Random();
    }
    @Override
    public int partition(String topic, Object keyObj, byte[] keyBytes, Object valueObj, byte[] valueBytes, Cluster cluster) {
        String key=(String)keyObj;
         //从集群元数据中把属于该topic的所有分区信息都读取出供分区策略使用
        List<PartitionInfo> partitionInfoList = cluster.availablePartitionsForTopic(topic);
        int partitionCount =partitionInfoList.size();
        int auditPartition=partitionCount-1;
        return key == null|| key.isEmpty()|| !key.contains ("audit")?random.nextInt(partitionCount-1):auditPartition;
    }
    @Override
    public void close() {
    //该方法实现必要资源的清理工作
    }
}

创建好自定义分区策略类后，在构建KafkaProducer 之前为Properties增加该属性；代码如下：

import org.apache.kafka.clients.producer.KafkaProducer;
import org.apache.kafka.clients.producer.Producer;
import org.apache.kafka.clients.producer.ProducerRecord;
import java.util.Properties;
import java.util.concurrent.ExecutionException;
public class ProducerTest {
    public static void main(String[] args) throws ExecutionException, InterruptedException {
        Properties props = new Properties();
        props.put("bootstrap.servers", "localhost:9092");//必须指定
        props.put("key.serializer", "org.apache.kafka.common.serialization.StringSerializer");//必须指定
        props.put("value.serializer", "org.apache.kafka.common.serialization.StringSerializer");//必须指定
        props.put("acks", "-1");
        props.put("retries", 3);
        props.put("batch.size", 323840);
        props.put("linger.ms", 10);
        props.put("buffer.memory", 33554432);
        props.put("max.block.ms", 3000);
		props.put("partitioner.class","com.exm.collectcodenew.kafka.producer.custompartitioner.AuditPartitioner");
        Producer<String, String> producer = new KafkaProducer<>(props);
        ProducerRecord nonKeyRecord = new ProducerRecord("topic-test","non-key record");
        ProducerRecord auditRecord = new ProducerRecord("topic-test", "audit","audit record");
        ProducerRecord nonAuditRecord =new ProducerRecord("topic-test","other","non-sudit record");
        producer.send(nonKeyRecord).get();
        producer.send(nonAuditRecord).get();
        producer.send(auditRecord).get();
        producer.send(nonKeyRecord).get();
        producer.send(nonAuditRecord).get();
        producer.close();
    }
}