实践代码采用kafka-clients V0.10.0.0 编写

一、编写producer

第一步：使用./kafka-topics.sh 命令创建topic及partitions 分区数

./kafka-topics.sh --create--zookepper "172.16.49.173:2181" --topic "producer_test" --partitions 10 replication-factor 3

第二步：实现org.apache.kafka.clients.producer.Partitioner 分区接口，以实现自定义的消息分区

import java.util.List;
import java.util.Map;
import org.apache.kafka.clients.producer.Partitioner;
import org.apache.kafka.common.Cluster;
import org.apache.kafka.common.PartitionInfo;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

public class MyPartition implements Partitioner {
    private static Logger LOG = LoggerFactory.getLogger(MyPartition.class);
    public MyPartition() {
        // TODO Auto-generated constructor stub
    }

    @Override
    public void configure(Map<String, ?> configs) {
        // TODO Auto-generated method stub

    }

    @Override
    public int partition(String topic, Object key, byte[] keyBytes, Object value, byte[] valueBytes, Cluster cluster) {
        // TODO Auto-generated method stub
        List<PartitionInfo> partitions = cluster.partitionsForTopic(topic);
        int numPartitions = partitions.size();
        int partitionNum = 0;
        try {
            partitionNum = Integer.parseInt((String) key);
        } catch (Exception e) {
            partitionNum = key.hashCode() ;
        }
        LOG.info("the message sendTo topic:"+ topic+" and the partitionNum:"+ partitionNum);
        return Math.abs(partitionNum  % numPartitions);
    }

    @Override
    public void close() {
        // TODO Auto-generated method stub

    }

}

第三步：编写 producer

import java.util.Properties;
import org.apache.kafka.clients.producer.Callback;
import org.apache.kafka.clients.producer.KafkaProducer;
import org.apache.kafka.clients.producer.ProducerRecord;
import org.apache.kafka.clients.producer.RecordMetadata;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

public class PartitionTest {
    private static Logger LOG = LoggerFactory.getLogger(PartitionTest.class);

    public static void main(String[] args) {
        // TODO Auto-generated method stub
        Properties props = new Properties();
        props.put("bootstrap.servers", "172.16.49.173:9092;172.16.49.173:9093");

        props.put("retries", 0);
        // props.put("batch.size", 16384);
        props.put("linger.ms", 1);
        // props.put("buffer.memory", 33554432);
        props.put("key.serializer", "org.apache.kafka.common.serialization.StringSerializer");
        props.put("value.serializer", "org.apache.kafka.common.serialization.StringSerializer");
        props.put("partitioner.class", "com.goodix.kafka.MyPartition");
        KafkaProducer<String, String> producer = new KafkaProducer<String, String>(props);
        ProducerRecord<String, String> record = new ProducerRecord<String, String>("producer_test", "2223132132",
                "test23_60");
        producer.send(record, new Callback() {
            @Override
            public void onCompletion(RecordMetadata metadata, Exception e) {
                // TODO Auto-generated method stub
                if (e != null)
                    LOG.error("the producer has a error:" + e.getMessage());
                else {
                    LOG.info("The offset of the record we just sent is: " + metadata.offset());
                    LOG.info("The partition of the record we just sent is: " + metadata.partition());
                }

            }
        });
        try {
            Thread.sleep(1000);
            producer.close();
        } catch (InterruptedException e1) {
            // TODO Auto-generated catch block
            e1.printStackTrace();
        }

    }

}

备注： 要先用命令创建topic及partitions 分区数;否则在自定义的分区中如果有大于1的情况下，发送数据消息到kafka时会报expired due to timeout while requesting metadata from brokers错误

二、使用Old Consumer High Level API编写consumer

第一步：编写具体处理消息的类

import java.io.UnsupportedEncodingException;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import kafka.consumer.ConsumerIterator;
import kafka.consumer.KafkaStream;
import kafka.message.MessageAndMetadata;


public class Consumerwork implements Runnable {
    private static Logger LOG = LoggerFactory.getLogger(Consumerwork.class);
     @SuppressWarnings("rawtypes")
    private KafkaStream m_stream;
     private int m_threadNumber;
     @SuppressWarnings("rawtypes")
    public Consumerwork(KafkaStream a_stream,int a_threadNumber) {
        // TODO Auto-generated constructor stub
         m_threadNumber = a_threadNumber;
         m_stream = a_stream;
    }

    @SuppressWarnings("unchecked")
    @Override
    public void run() {
        // TODO Auto-generated method stub
        ConsumerIterator<byte[], byte[]> it = m_stream.iterator();
           while (it.hasNext())
                try {
                    MessageAndMetadata<byte[], byte[]> thisMetadata=it.next();
                    String jsonStr = new String(thisMetadata.message(),"utf-8") ;
                    LOG.info("Thread " + m_threadNumber + ": " +jsonStr);
                    LOG.info("partion"+thisMetadata.partition()+",offset:"+thisMetadata.offset());
                    try {
                        Thread.sleep(1000);
                    } catch (InterruptedException e) {
                        // TODO Auto-generated catch block
                        e.printStackTrace();
                    }
                } catch (UnsupportedEncodingException e) {
                    // TODO Auto-generated catch block
                    e.printStackTrace();
                }
    }
}

第二步：编写启动Consumer主类

mport java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Properties;
import java.util.Scanner;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import kafka.consumer.ConsumerConfig;
import kafka.consumer.KafkaStream;
import kafka.javaapi.consumer.ConsumerConnector;
public class ConsumerGroup {
    private final ConsumerConnector consumer;
    private final String topic;
    private ExecutorService executor;
    private static Logger LOG = LoggerFactory.getLogger(ConsumerGroup.class);
    public ConsumerGroup(String a_zookeeper, String a_groupId, String a_topic) {
        consumer = kafka.consumer.Consumer.createJavaConsumerConnector(createConsumerConfig(a_zookeeper, a_groupId));
        this.topic = a_topic;
    }
    public static void main(String[] args) {
        Scanner sc = new Scanner(System.in);
        System.out.println("请输入zookeeper集群地址(如zk1:2181,zk2:2181,zk3:2181)：");
        String zooKeeper = sc.nextLine(); 
        System.out.println("请输入指定的消费group名称：");
        String groupId = sc.nextLine(); 
        System.out.println("请输入指定的消费topic名称：");
        String topic = sc.nextLine(); 
        System.out.println("请输入指定的消费处理线程数：");
        int threads = sc.nextInt();
        LOG.info("Starting consumer kafka messages with zk:" + zooKeeper + " and the topic is " + topic);
        ConsumerGroup example = new ConsumerGroup(zooKeeper, groupId, topic);
        example.run(threads);

        try {
            Thread.sleep(1000);
        } catch (InterruptedException ie) {

        }
        // example.shutdown();
    }

    private void shutdown() {
        // TODO Auto-generated method stub
        if (consumer != null)
            consumer.shutdown();
        if (executor != null)
            executor.shutdown();
        try {
            if (!executor.awaitTermination(5000, TimeUnit.MILLISECONDS)) {
                LOG.info("Timed out waiting for consumer threads to shut down, exiting uncleanly");
            }
        } catch (InterruptedException e) {
            LOG.info("Interrupted during shutdown, exiting uncleanly");
        }
    }

    private void run(int a_numThreads) {
        // TODO Auto-generated method stub
        Map<String, Integer> topicCountMap = new HashMap<String, Integer>();
        topicCountMap.put(topic, new Integer(a_numThreads));
        Map<String, List<KafkaStream<byte[], byte[]>>> consumerMap = consumer.createMessageStreams(topicCountMap);
        List<KafkaStream<byte[], byte[]>> streams = consumerMap.get(topic);

        // now launch all the threads
        //
        executor = Executors.newFixedThreadPool(a_numThreads);

        // now create an object to consume the messages
        //
        int threadNumber = 0;
        LOG.info("the streams size is "+streams.size());
        for (final KafkaStream stream : streams) {
            executor.submit(new com.goodix.kafka.oldconsumer.Consumerwork(stream, threadNumber));
    // consumer.commitOffsets();
            threadNumber++;
        }

    }

    private ConsumerConfig createConsumerConfig(String a_zookeeper, String a_groupId) {
        // TODO Auto-generated method stub
        Properties props = new Properties();
        props.put("zookeeper.connect", a_zookeeper);
        props.put("group.id", a_groupId);
        props.put("zookeeper.session.timeout.ms", "60000");
        props.put("zookeeper.sync.time.ms", "200");
        props.put("auto.commit.interval.ms", "1000");
        props.put("auto.offset.reset", "smallest");
// props.put("rebalance.max.retries", "5");
// props.put("rebalance.backoff.ms", "15000");
        return new ConsumerConfig(props);
    }

}

1. topicCountMap.put(topic, new Integer(a_numThreads)) 是告诉Kafka我有多少个线程来处理消息。

(1). 这个线程数必须是小等于topic的partition分区数；可以通过./kafka-topics.sh --describe --zookeeper "172.16.49.173:2181" --topic "producer_test"命令来查看分区的情况
(2). kafka会根据partition.assignment.strategy指定的分配策略来指定线程消费那些分区的消息；这里没有单独配置该项即是采用的默认值range策略（按照阶段平均分配）。比如分区有10个、线程数有3个，则线程 1消费0,1,2,3，线程2消费4,5,6,线程3消费7,8,9。另外一种是roundrobin(循环分配策略)，官方文档中写有使用该策略有两个前提条件的，所以一般不要去设定。
(3). 经过测试：consumerMap.get(topic).size()，应该是获得的目前该topic有数据的分区数
(4). stream即指的是来自一个或多个服务器上的一个或者多个partition的消息。每一个stream都对应一个单线程处理。因此，client能够设置满足自己需求的stream数目。总之，一个stream也许代表了多个服务器partion的消息的聚合，但是每一个 partition都只能到一个stream

2. Executors.newFixedThreadPool(a_numThreads)是创建一个创建固定容量大小的缓冲池：每次提交一个任务就创建一个线程，直到线程达到线程池的最大大小。线程池的大小一旦达到最大值就会保持不变，如果某个线程因为执行异常而结束，那么线程池会补充一个新线程。

3. props.put(“auto.offset.reset”, “smallest”) 是指定从最小没有被消费offset开始；如果没有指定该项则是默认的为largest，这样的话该consumer就得不到生产者先产生的消息。

4. 要使用old consumer API需要引用kafka_2.11以及kafka-clients。

<dependency>
<groupId>org.apache.kafka</groupId>
<artifactId>kafka_2.11</artifactId>
<version>0.10.0.0</version>
</dependency>
<dependency>
<groupId>org.apache.kafka</groupId>
<artifactId>kafka-clients</artifactId>
<version>0.10.0.0</version>
</dependency>

三、使用Old SimpleConsumerAPI编写consumer

这是一个更加底层和复杂的API

使用的场景

由于使用该API需要自己控制的项比较多，也比较复杂，官方给出了一些合适的适用场景，也可以理解成为这些场景是High Level Consumer API 不能够做到的

1. 针对一个消息读取多次
2. 在一个process中，仅仅处理一个topic中的一个partitions
3. 使用事务，确保每个消息只被处理一次

需要处理的事情

1. 必须在程序中跟踪offset值
2. 必须找出指定Topic Partition中的lead broker
3. 必须处理broker的变动

使用SimpleConsumer的步骤

首先，你必须知道读哪个topic的哪个partition
然后，找到负责该partition的broker leader，从而找到存有该partition副本的那个broker
再者，自己去写request并fetch数据
最终，还要注意需要识别和处理broker leader的改变

示例

package com.goodix.kafka.oldconsumer;
import kafka.api.FetchRequest;
import kafka.api.FetchRequestBuilder;
import kafka.api.PartitionOffsetRequestInfo;
import kafka.common.ErrorMapping;
import kafka.common.TopicAndPartition;
import kafka.javaapi.*;
import kafka.javaapi.consumer.SimpleConsumer;
import kafka.message.MessageAndOffset;

import java.nio.ByteBuffer;
import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Scanner;

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

public class SimpleExample {
    private static Logger LOG = LoggerFactory.getLogger(SimpleExample.class);
    public static void main(String args[]) {
        SimpleExample example = new SimpleExample();
        Scanner sc = new Scanner(System.in);
        System.out.println("请输入broker节点的ip地址(如172.16.49.173)");
        String brokerIp = sc.nextLine(); 
        List<String> seeds = new ArrayList<String>();
        seeds.add(brokerIp);
        System.out.println("请输入broker节点端口号(如9092)");
        int port = Integer.parseInt( sc.nextLine());
        System.out.println("请输入要订阅的topic名称(如test)");
        String topic = sc.nextLine();
        System.out.println("请输入要订阅要查找的分区(如0)");
        int partition = Integer.parseInt( sc.nextLine());
        System.out.println("请输入最大读取消息数量(如10000)");
        long maxReads = Long.parseLong( sc.nextLine());

        try {
            example.run(maxReads, topic, partition, seeds, port);
        } catch (Exception e) {
            LOG.error("Oops:" + e);
             e.printStackTrace();
        }
    }

    private List<String> m_replicaBrokers = new ArrayList<String>();

    public SimpleExample() {
        m_replicaBrokers = new ArrayList<String>();
    }

    public void run(long a_maxReads, String a_topic, int a_partition, List<String> a_seedBrokers, int a_port) throws Exception {
        // find the meta data about the topic and partition we are interested in
        //获取指定Topic partition的元数据 
        PartitionMetadata metadata = findLeader(a_seedBrokers, a_port, a_topic, a_partition);
        if (metadata == null) {
            LOG.error("Can't find metadata for Topic and Partition. Exiting");
            return;
        }
        if (metadata.leader() == null) {
            LOG.error("Can't find Leader for Topic and Partition. Exiting");
            return;
        }
        String leadBroker = metadata.leader().host();
        String clientName = "Client_" + a_topic + "_" + a_partition;
        SimpleConsumer consumer = new SimpleConsumer(leadBroker, a_port, 100000, 64 * 1024, clientName);
        long readOffset = getLastOffset(consumer,a_topic, a_partition, kafka.api.OffsetRequest.EarliestTime(), clientName);

        int numErrors = 0;
        while (a_maxReads > 0) {
            if (consumer == null) {
                consumer = new SimpleConsumer(leadBroker, a_port, 100000, 64 * 1024, clientName);
            }
            FetchRequest req = new FetchRequestBuilder()
                    .clientId(clientName)
                    .addFetch(a_topic, a_partition, readOffset, 100000) // Note: this fetchSize of 100000 might need to be increased if large batches are written to Kafka
                    .build();
            FetchResponse fetchResponse = consumer.fetch(req);

            if (fetchResponse.hasError()) {
                numErrors++;
                // Something went wrong!
                short code = fetchResponse.errorCode(a_topic, a_partition);
                LOG.error("Error fetching data from the Broker:" + leadBroker + " Reason: " + code);
                if (numErrors > 5) break;
                if (code == ErrorMapping.OffsetOutOfRangeCode())  {
                    // We asked for an invalid offset. For simple case ask for the last element to reset
                    readOffset = getLastOffset(consumer,a_topic, a_partition, kafka.api.OffsetRequest.LatestTime(), clientName);
                    continue;
                }
                consumer.close();
                consumer = null;
                leadBroker = findNewLeader(leadBroker, a_topic, a_partition, a_port);
                continue;
            }
            numErrors = 0;

            long numRead = 0;
            for (MessageAndOffset messageAndOffset : fetchResponse.messageSet(a_topic, a_partition)) {
                long currentOffset = messageAndOffset.offset();
                if (currentOffset < readOffset) {
                    LOG.error("Found an old offset: " + currentOffset + " Expecting: " + readOffset);
                    continue;
                }
                readOffset = messageAndOffset.nextOffset();
                ByteBuffer payload = messageAndOffset.message().payload();

                byte[] bytes = new byte[payload.limit()];
                payload.get(bytes);
                LOG.info("the messag's offset is :"+String.valueOf(messageAndOffset.offset()) + " and the value is :" + new String(bytes, "UTF-8"));
                numRead++;
                a_maxReads--;
            }

            if (numRead == 0) {
                try {
                    Thread.sleep(1000);
                } catch (InterruptedException ie) {
                }
            }
        }
        if (consumer != null) consumer.close();
    }

    public static long getLastOffset(SimpleConsumer consumer, String topic, int partition,
                                     long whichTime, String clientName) {
        TopicAndPartition topicAndPartition = new TopicAndPartition(topic, partition);
        Map<TopicAndPartition, PartitionOffsetRequestInfo> requestInfo = new HashMap<TopicAndPartition, PartitionOffsetRequestInfo>();
        requestInfo.put(topicAndPartition, new PartitionOffsetRequestInfo(whichTime, 1));
        kafka.javaapi.OffsetRequest request = new kafka.javaapi.OffsetRequest(
                requestInfo, kafka.api.OffsetRequest.CurrentVersion(), clientName);
        OffsetResponse response = consumer.getOffsetsBefore(request);

        if (response.hasError()) {
            LOG.error("Error fetching data Offset Data the Broker. Reason: " + response.errorCode(topic, partition) );
            return 0;
        }
        long[] offsets = response.offsets(topic, partition);
        return offsets[0];
    }
    /**
 * 找一个leader broker
 * 遍历每个broker，取出该topic的metadata，然后再遍历其中的每个partition metadata，如果找到我们要找的partition就返回
 * 根据返回的PartitionMetadata.leader().host()找到leader broker
 * @param a_oldLeader
 * @param a_topic
 * @param a_partition
 * @param a_port
 * @return
 * @throws Exception
 */
    private String findNewLeader(String a_oldLeader, String a_topic, int a_partition, int a_port) throws Exception {
        for (int i = 0; i < 3; i++) {
            boolean goToSleep = false;
            PartitionMetadata metadata = findLeader(m_replicaBrokers, a_port, a_topic, a_partition);
            if (metadata == null) {
                goToSleep = true;
            } else if (metadata.leader() == null) {
                goToSleep = true;
            } else if (a_oldLeader.equalsIgnoreCase(metadata.leader().host()) && i == 0) {
                // first time through if the leader hasn't changed give ZooKeeper a second to recover
                // second time, assume the broker did recover before failover, or it was a non-Broker issue
                //
                goToSleep = true;
            } else {
                return metadata.leader().host();
            }
            if (goToSleep) {
                try {
                    Thread.sleep(1000);
                } catch (InterruptedException ie) {
                }
            }
        }
        LOG.error("Unable to find new leader after Broker failure. Exiting");
        throw new Exception("Unable to find new leader after Broker failure. Exiting");
    }
    /**
 * 
 * @param a_seedBrokers
 * @param a_port
 * @param a_topic
 * @param a_partition
 * @return
 */
    private PartitionMetadata findLeader(List<String> a_seedBrokers, int a_port, String a_topic, int a_partition) {
        PartitionMetadata returnMetaData = null;

        loop:
        for (String seed : a_seedBrokers) { //遍历每个broker 
            SimpleConsumer consumer = null;
            try {
                // 创建Simple Consumer，
                consumer = new SimpleConsumer(seed, a_port, 100000, 64 * 1024, "leaderLookup");
                List<String> topics = Collections.singletonList(a_topic);
                TopicMetadataRequest req = new TopicMetadataRequest(topics);
                //发送TopicMetadata Request请求
                kafka.javaapi.TopicMetadataResponse resp = consumer.send(req);
                //取到Topic的Metadata 
                List<TopicMetadata> metaData = resp.topicsMetadata();
                //遍历每个partition的metadata
                for (TopicMetadata item : metaData) {
                    for (PartitionMetadata part : item.partitionsMetadata()) {
                        // 判断是否是要找的partition
                        if (part.partitionId() == a_partition) {
                            returnMetaData = part;
                            //找到就返回
                            break loop;
                        }
                    }
                }
            } catch (Exception e) {
                LOG.info("Error communicating with Broker [" + seed + "] to find Leader for [" + a_topic
                        + ", " + a_partition + "] Reason: " + e);
            } finally {
                if (consumer != null) consumer.close();
            }
        }
        if (returnMetaData != null) {
            m_replicaBrokers.clear();
            for (kafka.cluster.BrokerEndPoint replica : returnMetaData.replicas()) {
                m_replicaBrokers.add(replica.host());
            }
        }
        return returnMetaData;
    }
}

四、 使用NewConsumer API

(一)、自动提交offset偏移量

Properties props = new Properties();
//brokerServer(kafka)ip地址,不需要把所有集群中的地址都写上，可是一个或一部分
props.put("bootstrap.servers", "172.16.49.173:9092");
//设置consumer group name,必须设置
props.put("group.id", a_groupId);
//设置自动提交偏移量(offset),由auto.commit.interval.ms控制提交频率
props.put("enable.auto.commit", "true");
//偏移量(offset)提交频率
props.put("auto.commit.interval.ms", "1000");
//设置使用最开始的offset偏移量为该group.id的最早。如果不设置，则会是latest即该topic最新一个消息的offset
//如果采用latest，消费者只能得道其启动后，生产者生产的消息
props.put("auto.offset.reset", "earliest");
//设置心跳时间
props.put("session.timeout.ms", "30000");
//设置key以及value的解析（反序列）类
props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
props.put("value.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
KafkaConsumer<String, String> consumer = new KafkaConsumer<>(props);
//订阅topic
consumer.subscribe(Arrays.asList("topic_test"));
while (true) {
    //每次取100条信息
    ConsumerRecords<String, String> records = consumer.poll(100);
    for (ConsumerRecord<String, String> record : records)
    System.out.printf("offset = %d, key = %s, value = %s", record.offset(), record.key(), record.value());
 }

需要注意的:

group.id :必须设置
auto.offset.reset：如果想获得消费者启动前生产者生产的消息，则必须设置为earliest；如果只需要获得消费者启动后生产者生产的消息，则不需要设置该项
enable.auto.commit(默认值为true)：如果使用手动commit offset则需要设置为false，并再适当的地方调用consumer.commitSync()，否则每次启动消费折后都会从头开始消费信息(在auto.offset.reset=earliest的情况下);

（二）、 自己控制偏移量提交

很多时候，我们是希望在获得消息并经过一些逻辑处理后，才认为该消息已被消费，这可以通过自己控制偏移量提交来实现。

示例1：批量提交偏移量

import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
import java.util.Properties;

import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;
import org.apache.kafka.clients.consumer.KafkaConsumer;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;



/**
 * 手动批量提交偏移量
 * @author lxh
 *
 */
public class ManualOffsetConsumer {
    private static Logger LOG = LoggerFactory.getLogger(ManualOffsetConsumer.class);
    public ManualOffsetConsumer() {
        // TODO Auto-generated constructor stub
    }

    public static void main(String[] args) {
        // TODO Auto-generated method stub
        Properties props = new Properties();
        //props.put("bootstrap.servers", bootstrapServers);//"172.16.49.173:9092;172.16.49.173:9093");
        //设置brokerServer(kafka)ip地址
        props.put("bootstrap.servers", "172.16.49.173:9092");
        //设置consumer group name
        props.put("group.id","manual_g1");

        props.put("enable.auto.commit", "false");

        //设置使用最开始的offset偏移量为该group.id的最早。如果不设置，则会是latest即该topic最新一个消息的offset
        //如果采用latest，消费者只能得道其启动后，生产者生产的消息
        props.put("auto.offset.reset", "earliest");
        //
        props.put("session.timeout.ms", "30000");
        props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        props.put("value.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        KafkaConsumer<String ,String> consumer = new KafkaConsumer<String ,String>(props);
        consumer.subscribe(Arrays.asList("producer_test"));
        final int minBatchSize = 5;  //批量提交数量
         List<ConsumerRecord<String, String>> buffer = new ArrayList<>();
         while (true) {
             ConsumerRecords<String, String> records = consumer.poll(100);
             for (ConsumerRecord<String, String> record : records) {
                 LOG.info("consumer message values is "+record.value()+" and the offset is "+ record.offset());
                 buffer.add(record);
             }
             if (buffer.size() >= minBatchSize) {
                 LOG.info("now commit offset");
                 consumer.commitSync();
                 buffer.clear();
             }
         }
    }

}

示例2：消费完一个分区后手动提交偏移量

package com.goodix.kafka;

import java.util.Arrays;
import java.util.Collections;
import java.util.List;
import java.util.Properties;

import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;
import org.apache.kafka.clients.consumer.KafkaConsumer;
import org.apache.kafka.clients.consumer.OffsetAndMetadata;
import org.apache.kafka.common.TopicPartition;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
 * 消费完一个分区后手动提交偏移量
 * @author lxh
 *
 */
public class ManualCommitPartion {
    private static Logger LOG = LoggerFactory.getLogger(ManualCommitPartion.class);
    public ManualCommitPartion() {
        // TODO Auto-generated constructor stub
    }

    public static void main(String[] args) {
        // TODO Auto-generated method stub
        Properties props = new Properties();
        //props.put("bootstrap.servers", bootstrapServers);//"172.16.49.173:9092;172.16.49.173:9093");
        //设置brokerServer(kafka)ip地址
        props.put("bootstrap.servers", "172.16.49.173:9092");
        //设置consumer group name
        props.put("group.id","manual_g2");

        props.put("enable.auto.commit", "false");

        //设置使用最开始的offset偏移量为该group.id的最早。如果不设置，则会是latest即该topic最新一个消息的offset
        //如果采用latest，消费者只能得道其启动后，生产者生产的消息
        props.put("auto.offset.reset", "earliest");
        //
        props.put("session.timeout.ms", "30000");
        props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        props.put("value.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        KafkaConsumer<String ,String> consumer = new KafkaConsumer<String ,String>(props);
        consumer.subscribe(Arrays.asList("producer_test"));
        while (true) {
            ConsumerRecords<String, String> records = consumer.poll(Long.MAX_VALUE);
            for (TopicPartition partition : records.partitions()) {
                List<ConsumerRecord<String, String>> partitionRecords = records.records(partition);
                for (ConsumerRecord<String, String> record : partitionRecords) {
                    LOG.info("now consumer the message it's offset is :"+record.offset() + " and the value is :" + record.value());
                }
                long lastOffset = partitionRecords.get(partitionRecords.size() - 1).offset();
                LOG.info("now commit the partition[ "+partition.partition()+"] offset");
                consumer.commitSync(Collections.singletonMap(partition, new OffsetAndMetadata(lastOffset + 1)));
            }
        }
    }

}

（三）、指定消费某个分区的消息

import java.util.Arrays;
import java.util.Properties;

import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;
import org.apache.kafka.clients.consumer.KafkaConsumer;
import org.apache.kafka.common.TopicPartition;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
 * 消费指定分区的消息
 * @author lxh
 *
 */
public class ManualPartion {
    private static Logger LOG = LoggerFactory.getLogger(ManualPartion.class);
    public ManualPartion() {
        // TODO Auto-generated constructor stub
    }

    public static void main(String[] args) {
        Properties props = new Properties();
        //设置brokerServer(kafka)ip地址
        props.put("bootstrap.servers", "172.16.49.173:9092");
        //设置consumer group name
        props.put("group.id", "manual_g4");
        //设置自动提交偏移量(offset),由auto.commit.interval.ms控制提交频率
        props.put("enable.auto.commit", "true");
        //偏移量(offset)提交频率
        props.put("auto.commit.interval.ms", "1000");
        //设置使用最开始的offset偏移量为该group.id的最早。如果不设置，则会是latest即该topic最新一个消息的offset
        //如果采用latest，消费者只能得道其启动后，生产者生产的消息
        props.put("auto.offset.reset", "earliest");
        //
        props.put("session.timeout.ms", "30000");
        props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        props.put("value.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        TopicPartition partition0 = new TopicPartition("producer_test", 0);
        TopicPartition partition1 = new TopicPartition("producer_test", 1);
        KafkaConsumer<String ,String> consumer = new KafkaConsumer<String ,String>(props);
        consumer.assign(Arrays.asList(partition0, partition1));
        while (true) {
              ConsumerRecords<String, String> records = consumer.poll(Long.MAX_VALUE);
              for (ConsumerRecord<String, String> record : records)
                  System.out.printf("offset = %d, key = %s, value = %s \r\n", record.offset(), record.key(), record.value());
              try {
                Thread.sleep(1000);
            } catch (InterruptedException e) {
                // TODO Auto-generated catch block
                e.printStackTrace();
            }
        }
    }

}

总结

使用newConsumer API 只需要引用kafka-clients即可
newConsumer API 更加易懂、易用

<dependency>
<groupId>org.apache.kafka</groupId>
<artifactId>kafka-clients</artifactId>
<version>0.10.0.0</version>
</dependency>

官方对于consumer与partition的建议

1. 如果consumer比partition多，是浪费，因为kafka的设计是在一个partition上是不允许并发的，所以consumer数不要大于partition数
2. 如果consumer比partition少，一个consumer会对应于多个partitions，这里主要合理分配consumer数和partition数，否则会导致partition里面的数据被取的不均匀。最好partiton数目是consumer数目的整数倍，所以partition数目很重要，比如取24，就很容易设定consumer数目
3. 如果consumer从多个partition读到数据，不保证数据间的顺序性，kafka只保证在一个partition上数据是有序的，但多个partition，根据你读的顺序会有不同
4. 增减consumer，broker，partition会导致rebalance，所以rebalance后consumer对应的partition会发生变化
5. High-level接口中获取不到数据的时候是会block的

参考文章：
Java线程池使用说明
Java并发编程：线程池的使用
如何确定Kafka的分区数、key和consumer线程数s
Kafka Consumer接口