参考资料:http://blog.csdn.net/honglei915/article/details/37563647
参数说明:http://ju.outofmemory.cn/entry/119243
参数说明/Demo:http://www.aboutyun.com/thread-9906-1-1.html
Kafka+Spark:
http://shiyanjun.cn/archives/1097.html
http://ju.outofmemory.cn/entry/84636
1. Kafka启动:
1. 先启动所有节点的zookeeper : 进入ZOOKEEPER_HOME/bin 执行./zkServer.sh start
2. 启动所有节点的kafka:进入 KAFKA_HOME/bin 执行 ./kafka-server-start.sh config/server.properties &
2. 参数说明
2.0 boker参数说明 (配置文件位于config/server.properties)
name | 默认值 | 描述 |
---|---|---|
broker.id | none | 每一个boker都有一个唯一的id作为它们的名字。 这就允许boker切换到别的主机/端口上, consumer依然知道 |
enable.zookeeper | true | 允许注册到zookeeper |
log.flush.interval.messages | Long.MaxValue | 在数据被写入到硬盘和消费者可用前最大累积的消息的数量 |
log.flush.interval.ms | Long.MaxValue | 在数据被写入到硬盘前的最大时间 |
log.flush.scheduler.interval.ms | Long.MaxValue | 检查数据是否要写入到硬盘的时间间隔。 |
log.retention.hours | 168 | 控制一个log保留多长个小时 |
log.retention.bytes | -1 | 控制log文件最大尺寸 |
log.cleaner.enable | false | 是否log cleaning |
log.cleanup.policy | delete | delete还是compat. 其它控制参数还包括log.cleaner.threads,log.cleaner.io.max.bytes.per.second, log.cleaner.dedupe.buffer.size,log.cleaner.io.buffer.size,log.cleaner.io.buffer.load.factor, log.cleaner.backoff.ms,log.cleaner.min.cleanable.ratio,log.cleaner.delete.retention.ms |
log.dir | /tmp/kafka-logs | 指定log文件的根目录 |
log.segment.bytes | 110241024*1024 | 单一的log segment文件大小 |
log.roll.hours | 24 * 7 | 开始一个新的log文件片段的最大时间 |
message.max.bytes | 1000000 + MessageSet.LogOverhead | 一个socket 请求的最大字节数 |
num.network.threads | 3 | 处理网络请求的线程数 |
num.io.threads | 8 | 处理IO的线程数 |
background.threads | 10 | 后台线程序 |
num.partitions | 1 | 默认分区数 |
socket.send.buffer.bytes | 102400 | socket SO_SNDBUFF参数 |
socket.receive.buffer.bytes | 102400 | socket SO_RCVBUFF参数 |
zookeeper.connect | localhost:2182/kafka | 指定zookeeper连接字符串, 格式如hostname:port/chroot。chroot是一个namespace |
zookeeper.connection.timeout.ms | 6000 | 指定客户端连接zookeeper的最大超时时间 |
zookeeper.session.timeout.ms | 6000 | 连接zk的session超时时间 |
zookeeper.sync.time.ms | 2000 | zk follower落后于zk leader的最长时间 |
2.1 producer参数说明(配置文件位于config/producer.properties或者在程序内定义)
#指定kafka节点列表,用于获取metadata,不必全部指定
metadata.broker.list=192.168.2.105:,192.168.2.106: # 指定分区处理类。默认kafka.producer.DefaultPartitioner,表通过key哈希到对应分区
#partitioner.class=com.meituan.mafka.client.producer.CustomizePartitioner # 是否压缩,默认0表示不压缩,1表示用gzip压缩,2表示用snappy压缩。压缩后消息中会有头来指明消息压缩类型,故在消费者端消息解压是透明的无需指定。
compression.codec=none # 指定序列化处理类(mafka client API调用说明-->.序列化约定wiki),默认为kafka.serializer.DefaultEncoder,即byte[]
serializer.class=com.meituan.mafka.client.codec.MafkaMessageEncoder
# serializer.class=kafka.serializer.DefaultEncoder
# serializer.class=kafka.serializer.StringEncoder # 如果要压缩消息,这里指定哪些topic要压缩消息,默认empty,表示不压缩。
#compressed.topics= ########### request ack ###############
# producer接收消息ack的时机.默认为0.
# : producer不会等待broker发送ack
# : 当leader接收到消息之后发送ack
# : 当所有的follower都同步消息成功后发送ack.
request.required.acks= # 在向producer发送ack之前,broker允许等待的最大时间
# 如果超时,broker将会向producer发送一个error ACK.意味着上一次消息因为某种
# 原因未能成功(比如follower未能同步成功)
request.timeout.ms=
########## end ##################### # 同步还是异步发送消息,默认“sync”表同步,"async"表异步。异步可以提高发送吞吐量,
# 也意味着消息将会在本地buffer中,并适时批量发送,但是也可能导致丢失未发送过去的消息
producer.type=sync ############## 异步发送 (以下四个异步参数可选) ####################
# 在async模式下,当message被缓存的时间超过此值后,将会批量发送给broker,默认为5000ms
# 此值和batch.num.messages协同工作.
queue.buffering.max.ms = # 在async模式下,producer端允许buffer的最大消息量
# 无论如何,producer都无法尽快的将消息发送给broker,从而导致消息在producer端大量沉积
# 此时,如果消息的条数达到阀值,将会导致producer端阻塞或者消息被抛弃,默认为10000
queue.buffering.max.messages= # 如果是异步,指定每次批量发送数据量,默认为200
batch.num.messages= # 当消息在producer端沉积的条数达到"queue.buffering.max.meesages"后
# 阻塞一定时间后,队列仍然没有enqueue(producer仍然没有发送出任何消息)
# 此时producer可以继续阻塞或者将消息抛弃,此timeout值用于控制"阻塞"的时间
# -: 无阻塞超时限制,消息不会被抛弃
# :立即清空队列,消息被抛弃
queue.enqueue.timeout.ms=-
################ end ############### # 当producer接收到error ACK,或者没有接收到ACK时,允许消息重发的次数
# 因为broker并没有完整的机制来避免消息重复,所以当网络异常时(比如ACK丢失)
# 有可能导致broker接收到重复的消息,默认值为3.
message.send.max.retries= # producer刷新topic metada的时间间隔,producer需要知道partition leader的位置,以及当前topic的情况
# 因此producer需要一个机制来获取最新的metadata,当producer遇到特定错误时,将会立即刷新
# (比如topic失效,partition丢失,leader失效等),此外也可以通过此参数来配置额外的刷新机制,默认值600000
topic.metadata.refresh.interval.ms=
2.2 consumer参数说明(配置文件位于config/consumer.properties或者在程序内定义)
# zookeeper连接服务器地址,此处为线下测试环境配置(kafka消息服务-->kafka broker集群线上部署环境wiki)
# 配置例子:"127.0.0.1:3000,127.0.0.1:3001,127.0.0.1:3002"
zookeeper.connect=192.168.2.225:,192.168.2.225:,192.168.2.225:/config/mobile/mq/mafka # zookeeper的session过期时间,默认5000ms,用于检测消费者是否挂掉,当消费者挂掉,其他消费者要等该指定时间才能检查到并且触发重新负载均衡
zookeeper.session.timeout.ms=
zookeeper.connection.timeout.ms= # 指定多久消费者更新offset到zookeeper中。注意offset更新时基于time而不是每次获得的消息。一旦在更新zookeeper发生异常并重启,将可能拿到已拿到过的消息
zookeeper.sync.time.ms= #指定消费组
group.id=xxx # 当consumer消费一定量的消息之后,将会自动向zookeeper提交offset信息
# 注意offset信息并不是每消费一次消息就向zk提交一次,而是现在本地保存(内存),并定期提交,默认为true
auto.commit.enable=true # 自动更新时间。默认60 *
auto.commit.interval.ms= # 当前consumer的标识,可以设定,也可以有系统生成,主要用来跟踪消息消费情况,便于观察
conusmer.id=xxx # 消费者客户端编号,用于区分不同客户端,默认客户端程序自动产生
client.id=xxxx # 最大取多少块缓存到消费者(默认10)
queued.max.message.chunks= # 当有新的consumer加入到group时,将会reblance,此后将会有partitions的消费端迁移到新
# 的consumer上,如果一个consumer获得了某个partition的消费权限,那么它将会向zk注册
# "Partition Owner registry"节点信息,但是有可能此时旧的consumer尚没有释放此节点,
# 此值用于控制,注册节点的重试次数.
rebalance.max.retries= # 获取消息的最大尺寸,broker不会像consumer输出大于此值的消息chunk
# 每次feth将得到多条消息,此值为总大小,提升此值,将会消耗更多的consumer端内存
fetch.min.bytes= # 当消息的尺寸不足时,server阻塞的时间,如果超时,消息将立即发送给consumer
fetch.wait.max.ms=
socket.receive.buffer.bytes= # 如果zookeeper没有offset值或offset值超出范围。那么就给个初始的offset。有smallest、largest、
# anything可选,分别表示给当前最小的offset、当前最大的offset、抛异常。默认largest
auto.offset.reset=smallest # 指定序列化处理类(mafka client API调用说明-->.序列化约定wiki),默认为kafka.serializer.DefaultDecoder,即byte[]
derializer.class=com.meituan.mafka.client.codec.MafkaMessageDecoder
3. 例:
接口 KafkaProperties.java
public interface KafkaProperties {
final static String zkConnect = "192.168.1.160:2181";
final static String groupId = "group1";
final static String topic = "topic1";
// final static String kafkaServerURL = "192.168.1.160";
// final static int kafkaServerPort = 9092;
// final static int kafkaProducerBufferSize = 64 * 1024;
// final static int connectionTimeOut = 20000;
// final static int reconnectInterval = 10000;
// final static String topic2 = "topic2";
// final static String topic3 = "topic3";
// final static String clientId = "SimpleConsumerDemoClient";
}
生产者 KafkaProducer.java
import java.util.Properties; import kafka.producer.KeyedMessage;
import kafka.producer.ProducerConfig; public class KafkaProducer extends Thread {
private final kafka.javaapi.producer.Producer<Integer, String> producer;
private final String topic;
private final Properties props = new Properties(); public KafkaProducer(String topic) {
props.put("serializer.class", "kafka.serializer.StringEncoder");
props.put("metadata.broker.list", "192.168.1.160:9092"); // 配置kafka端口
producer = new kafka.javaapi.producer.Producer<Integer, String>(new ProducerConfig(props));
this.topic = topic;
} @Override
public void run() {
int messageNo = 1;
while (true) {
String messageStr = new String("This is a message, number: " + messageNo);
System.out.println("Send:" + messageStr);
producer.send(new KeyedMessage<Integer, String>(topic, messageStr));
messageNo++;
try {
sleep(1000);
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
} }
消费者 KafkaConsumer.java
import java.util.Properties; import kafka.consumer.ConsumerConfig;
import kafka.consumer.ConsumerIterator;
import kafka.consumer.KafkaStream;
import kafka.javaapi.consumer.ConsumerConnector; public class KafkaConsumer extends Thread {
private final ConsumerConnector consumer;
private final String topic; public KafkaConsumer(String topic) {
consumer = kafka.consumer.Consumer.createJavaConsumerConnector(createConsumerConfig());
this.topic = topic;
} private static ConsumerConfig createConsumerConfig() {
Properties props = new Properties();
props.put("zookeeper.connect", KafkaProperties.zkConnect); // zookeeper的地址
props.put("group.id", KafkaProperties.groupId); // 组ID //zk连接超时
props.put("zookeeper.session.timeout.ms", "40000");
props.put("zookeeper.sync.time.ms", "200");
props.put("auto.commit.interval.ms", "1000"); return new ConsumerConfig(props);
} @Override
public void run() {
Map<String, Integer> topicCountMap = new HashMap<String, Integer>();
topicCountMap.put(topic, new Integer(1)); Map<String, List<KafkaStream<byte[], byte[]>>> consumerMap = consumer.createMessageStreams(topicCountMap); KafkaStream<byte[], byte[]> stream = consumerMap.get(topic).get(0);
ConsumerIterator<byte[], byte[]> it = stream.iterator();
while (it.hasNext()) {
System.out.println("receive:" + new String(it.next().message()));
try {
sleep(1000);
} catch (InterruptedException e) {
e.printStackTrace();
}
}
}
}
执行函数 KafkaConsumerProducerDemo.java
public class KafkaConsumerProducerDemo {
public static void main(String[] args) {
KafkaProducer producerThread = new KafkaProducer(KafkaProperties.topic);
producerThread.start(); KafkaConsumer consumerThread = new KafkaConsumer(KafkaProperties.topic);
consumerThread.start();
}
}
-----------------------------
另一个例子:http://www.cnblogs.com/sunxucool/p/3913919.html
Producer端代码
1) producer.properties文件:此文件放在/resources目录下
#partitioner.class=
metadata.broker.list=127.0.0.1:9092,127.0.0.1:9093
##,127.0.0.1:9093
producer.type=sync
compression.codec=0
serializer.class=kafka.serializer.StringEncoder
##在producer.type=async时有效
#batch.num.messages=100
2) LogProducer.java代码样例
package com.test.kafka; import java.util.ArrayList;
import java.util.Collection;
import java.util.List;
import java.util.Properties; import kafka.javaapi.producer.Producer;
import kafka.producer.KeyedMessage;
import kafka.producer.ProducerConfig;
public class LogProducer { private Producer<String,String> inner;
public LogProducer() throws Exception{
Properties properties = new Properties();
properties.load(ClassLoader.getSystemResourceAsStream("producer.properties"));
ProducerConfig config = new ProducerConfig(properties);
inner = new Producer<String, String>(config);
} public void send(String topicName,String message) {
if(topicName == null || message == null){
return;
}
KeyedMessage<String, String> km = new KeyedMessage<String, String>(topicName,message);
inner.send(km);
} public void send(String topicName,Collection<String> messages) {
if(topicName == null || messages == null){
return;
}
if(messages.isEmpty()){
return;
}
List<KeyedMessage<String, String>> kms = new ArrayList<KeyedMessage<String, String>>();
for(String entry : messages){
KeyedMessage<String, String> km = new KeyedMessage<String, String>(topicName,entry);
kms.add(km);
}
inner.send(kms);
} public void close(){
inner.close();
} /**
* @param args
*/
public static void main(String[] args) {
LogProducer producer = null;
try{
producer = new LogProducer();
int i=0;
while(true){
producer.send("test-topic", "this is a sample" + i);
i++;
Thread.sleep(2000);
}
}catch(Exception e){
e.printStackTrace();
}finally{
if(producer != null){
producer.close();
}
} } }
五.Consumer端
1) consumer.properties:文件位于/resources目录下
zookeeper.connect=127.0.0.1:2181,127.0.0.1:2182,127.0.0.1:2183
##,127.0.0.1:2182,127.0.0.1:2183
# timeout in ms for connecting to zookeeper
zookeeper.connectiontimeout.ms=1000000
#consumer group id
group.id=test-group
#consumer timeout
#consumer.timeout.ms=5000
2) LogConsumer.java代码样例
package com.test.kafka; import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Properties;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors; import kafka.consumer.Consumer;
import kafka.consumer.ConsumerConfig;
import kafka.consumer.ConsumerIterator;
import kafka.consumer.KafkaStream;
import kafka.javaapi.consumer.ConsumerConnector;
import kafka.message.MessageAndMetadata;
public class LogConsumer { private ConsumerConfig config;
private String topic;
private int partitionsNum;
private MessageExecutor executor;
private ConsumerConnector connector;
private ExecutorService threadPool;
public LogConsumer(String topic,int partitionsNum,MessageExecutor executor) throws Exception{
Properties properties = new Properties();
properties.load(ClassLoader.getSystemResourceAsStream("consumer.properties"));
config = new ConsumerConfig(properties);
this.topic = topic;
this.partitionsNum = partitionsNum;
this.executor = executor;
} public void start() throws Exception{
connector = Consumer.createJavaConsumerConnector(config);
Map<String,Integer> topics = new HashMap<String,Integer>();
topics.put(topic, partitionsNum);
Map<String, List<KafkaStream<byte[], byte[]>>> streams = connector.createMessageStreams(topics);
List<KafkaStream<byte[], byte[]>> partitions = streams.get(topic);
threadPool = Executors.newFixedThreadPool(partitionsNum);
for(KafkaStream<byte[], byte[]> partition : partitions){
threadPool.execute(new MessageRunner(partition));
}
} public void close(){
try{
threadPool.shutdownNow();
}catch(Exception e){
//
}finally{
connector.shutdown();
} } class MessageRunner implements Runnable{
private KafkaStream<byte[], byte[]> partition; MessageRunner(KafkaStream<byte[], byte[]> partition) {
this.partition = partition;
} public void run(){
ConsumerIterator<byte[], byte[]> it = partition.iterator();
while(it.hasNext()){
MessageAndMetadata<byte[],byte[]> item = it.next();
System.out.println("partiton:" + item.partition());
System.out.println("offset:" + item.offset());
executor.execute(new String(item.message()));//UTF-8
}
}
} interface MessageExecutor { public void execute(String message);
} /**
* @param args
*/
public static void main(String[] args) {
LogConsumer consumer = null;
try{
MessageExecutor executor = new MessageExecutor() { public void execute(String message) {
System.out.println(message); }
};
consumer = new LogConsumer("test-topic", 2, executor);
consumer.start();
}catch(Exception e){
e.printStackTrace();
}finally{
// if(consumer != null){
// consumer.close();
// }
} } }