Kafka - 消费接口分析

1.概述

　　在 Kafka 中，官方对外提供了两种消费 API，一种是高等级消费 API，另一种是低等级的消费 API。在《高级消费 API》一文中，介绍了其高级消费的 API 实现。今天给大家介绍另一种消费 API。

2.内容

　　在使用过 Kafka 的高级消费 API 后，我们知道它是一种高度抽象的消费 API，使用起来简单，方便，但是对于某些特殊的需求我们可能要用到第二种更加底层的 API。那么，我们首先需要知道低级消费 API 的作用。它能帮助我们去做那些事情：

一个消息进行多次读取
在处理过程中只消费 Partition 其中的某一部分消息
添加事物管理机制以保证消息仅被处理一次

　　当然，在使用的过程当中也是有些弊端的，其内容如下：

必须在程序中跟踪 Offset 的值
必须找出指定的 Topic Partition 中的 Lead Broker
必须处理 Broker 的变动

　　使用其 API 的思路步骤如下所示：

从所有处于 Active 状态的 Broker 中找出哪个是指定 Topic Partition 中的 Lead Broker
找出指定 Topic Partition 中的所有备份 Broker
构造请求
发送请求并查询数据
处理 Leader Broker 的变动

3.代码实现

3.1 Java Project

　　若是使用 Java Project 工程去实现该部分代码，需要添加相关以来 JAR 文件，其内容包含如下：

scala-xml_${version}-${version}.jar
scala-library-${version}.jar
metrics-core-${version}.jar
kafka-client-${version}.jar
kafka_${version}-${version}.jar

　　针对 Java Project 工程，需要自己筛选 JAR 去添加。保证代码的顺利执行。

3.2 Maven Project

　　对 Maven 工程，在 pom.xml 文件中添加相应的依赖信息即可，简单方便。让 Maven 去管理相应的依赖 JAR 文件。内容如下所示：

<dependency>
    <groupId>org.apache.kafka</groupId>
    <artifactId>kafka_2.11</artifactId>
    <version>0.8.2.1</version>
    <exclusions>
        <exclusion>
            <groupId>org.apache.zookeeper</groupId>
            <artifactId>zookeeper</artifactId>
    </exclusion>
    <exclusion>
            <groupId>log4j</groupId>
            <artifactId>log4j</artifactId>
    </exclusion>
    </exclusions>
</dependency>

　　这样在 Maven 工程中相应的依赖 JAR 文件就添加完成了。

3.3 代码实现

　　在低级消费 API 中，实现代码如下所示：

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

/**
 * @Date Mar 2, 2016
 *
 * @Author dengjie
 *
 * @Note Simple consumer api
 */
public class SimpleKafkaConsumer {
    private static Logger log = LoggerFactory.getLogger(SimpleKafkaConsumer.class);
    private List<String> m_replicaBrokers = new ArrayList<String>();
 
    public SimpleKafkaConsumer() {
        m_replicaBrokers = new ArrayList<String>();
    }
 
    public static void main(String[] args) {
        SimpleKafkaConsumer example = new SimpleKafkaConsumer();
        // Max read number
        long maxReads = SystemConfig.getIntProperty("kafka.read.max");
        // To subscribe to the topic
        String topic = SystemConfig.getProperty("kafka.topic");
        // Find partition
        int partition = SystemConfig.getIntProperty("kafka.partition");
        // Broker node's ip
        List<String> seeds = new ArrayList<String>();
        String[] hosts = SystemConfig.getPropertyArray("kafka.server.host", ",");
        for (String host : hosts) {
            seeds.add(host);
        }
        int port = SystemConfig.getIntProperty("kafka.server.port");
        try {
            example.run(maxReads, topic, partition, seeds, port);
        } catch (Exception e) {
            log.error("Oops:" + e);
            e.printStackTrace();
        }
    }
 
    public void run(long a_maxReads, String a_topic, int a_partition, List<String> a_seedBrokers, int a_port)
            throws Exception {
        // Get point topic partition's meta
        PartitionMetadata metadata = findLeader(a_seedBrokers, a_port, a_topic, a_partition);
        if (metadata == null) {
            log.info("[SimpleKafkaConsumer.run()] - Can't find metadata for Topic and Partition. Exiting");
            return;
        }
        if (metadata.leader() == null) {
            log.info("[SimpleKafkaConsumer.run()] - 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).build();
            FetchResponse fetchResponse = consumer.fetch(req);
 
            if (fetchResponse.hasError()) {
                numErrors++;
                // Something went wrong!
                short code = fetchResponse.errorCode(a_topic, a_partition);
                log.info("[SimpleKafkaConsumer.run()] - 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.info("[SimpleKafkaConsumer.run()] - 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);
                System.out.println(String.valueOf(messageAndOffset.offset()) + ": " + new String(bytes, "UTF-8")); // Message deal enter
                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.info("[SimpleKafkaConsumer.getLastOffset()] - Error fetching data Offset Data the Broker. Reason: "
                    + response.errorCode(topic, partition));
            return 0;
        }
        long[] offsets = response.offsets(topic, partition);
        return offsets[0];
    }
 
    /**
     * @param a_oldLeader
     * @param a_topic
     * @param a_partition
     * @param a_port
     * @return String
     * @throws Exception
     *             find next leader broker
     */
    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) {
                }
            }
        }
        throw new Exception("Unable to find new leader after Broker failure. Exiting");
    }
 
    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) {
            SimpleConsumer consumer = null;
            try {
                consumer = new SimpleConsumer(seed, a_port, 100000, 64 * 1024, "leaderLookup");
                List<String> topics = Collections.singletonList(a_topic);
                TopicMetadataRequest req = new TopicMetadataRequest(topics);
                kafka.javaapi.TopicMetadataResponse resp = consumer.send(req);
 
                List<TopicMetadata> metaData = resp.topicsMetadata();
                for (TopicMetadata item : metaData) {
                    for (PartitionMetadata part : item.partitionsMetadata()) {
                        if (part.partitionId() == a_partition) {
                            returnMetaData = part;
                            break loop;
                        }
                    }
                }
            } catch (Exception e) {
                log.error("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.Broker replica : returnMetaData.replicas()) {
                m_replicaBrokers.add(replica.host());
            }
        }
        return returnMetaData;
    }
}