flume进阶（事务，原理，自定义，案例）

Flume 事务

Flume Agent 内部原理

Flume 拓扑结构

1.简单串联

2.副本和多路复用（多路复用要配合拦截器使用）

3.负载均衡和故障转移

Flume支持使用将多个sink逻辑上分到一个sink组，sink组配合不同的SinkProcessor可以实现负载均衡和错误恢复的功能。

4.聚合

Flume 企业开发案例

1.副本和多路复用

单数据源多出口案例（副本）

在/opt/module/datas/目录下创建flume3文件夹
[root@hadoop003 datas]$ mkdir flume3

在flume/job目录下创建group1文件夹
[root@hadoop003 job]$ mkdir group1

配置监控日志变化的flume
[root@hadoop003 group1]$ touch flume-file-flume.conf
[root@hadoop003 group1]$ vim flume-file-flume.conf

添加内容

# Name the components on this agent
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1 c2
# 将数据流复制给多个channel
a1.sources.r1.selector.type = replicating

# Describe/configure the source
a1.sources.r1.type = exec
a1.sources.r1.command = tail -F /opt/module/hive/logs/hive.log
a1.sources.r1.shell = /bin/bash -c

# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = hadoop003
a1.sinks.k1.port = 4141

a1.sinks.k2.type = avro
a1.sinks.k2.hostname = hadoop003
a1.sinks.k2.port = 4142

# Describe the channel
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100

a1.channels.c2.type = memory
a1.channels.c2.capacity = 1000
a1.channels.c2.transactionCapacity = 100

# Bind the source and sink to the channel
a1.sources.r1.channels = c1 c2
a1.sinks.k1.channel = c1
a1.sinks.k2.channel = c2

配置上级flume输出的source，输出到hdfs
[root@hadoop003 group1]$ touch flume-flume-hdfs.conf
[root@hadoop003 group1]$ vim flume-flume-hdfs.conf
添加内容

# Name the components on this agent
a2.sources = r1
a2.sinks = k1
a2.channels = c1

# Describe/configure the source
a2.sources.r1.type = avro
a2.sources.r1.bind = hadoop003
a2.sources.r1.port = 4141

# Describe the sink
a2.sinks.k1.type = hdfs
a2.sinks.k1.hdfs.path = hdfs://hadoop003:9000/flume2/%Y%m%d/%H
#上传文件的前缀
a2.sinks.k1.hdfs.filePrefix = flume2-
#是否按照时间滚动文件夹
a2.sinks.k1.hdfs.round = true
#多少时间单位创建一个新的文件夹
a2.sinks.k1.hdfs.roundValue = 1
#重新定义时间单位
a2.sinks.k1.hdfs.roundUnit = hour
#是否使用本地时间戳
a2.sinks.k1.hdfs.useLocalTimeStamp = true
#积攒多少个Event才flush到HDFS一次
a2.sinks.k1.hdfs.batchSize = 100
#设置文件类型，可支持压缩
a2.sinks.k1.hdfs.fileType = DataStream
#多久生成一个新的文件
a2.sinks.k1.hdfs.rollInterval = 30
#设置每个文件的滚动大小大概是128M
a2.sinks.k1.hdfs.rollSize = 134217700
#文件的滚动与Event数量无关
a2.sinks.k1.hdfs.rollCount = 0
#最小冗余数
a2.sinks.k1.hdfs.minBlockReplicas = 1
#访问hdfs的超时时间 默认10 000 毫秒
a2.sinks.k1.hdfs.callTimeout = 1000000


# Describe the channel
a2.channels.c1.type = memory
a2.channels.c1.capacity = 1000
a2.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a2.sources.r1.channels = c1
a2.sinks.k1.channel = c1

配置上级flume输出的source，输出到本地目录。
输出的本地目录必须是已经存在的目录。
[root@hadoop003 group1]$ touch flume-flume-dir.conf
[root@hadoop003 group1]$ vim flume-flume-dir.conf
添加内容

# Name the components on this agent
a3.sources = r1
a3.sinks = k1
a3.channels = c2

# Describe/configure the source
a3.sources.r1.type = avro
a3.sources.r1.bind = hadoop003
a3.sources.r1.port = 4142

# Describe the sink
a3.sinks.k1.type = file_roll
a3.sinks.k1.sink.directory = /opt/module/datas/flume3

# Describe the channel
a3.channels.c2.type = memory
a3.channels.c2.capacity = 1000
a3.channels.c2.transactionCapacity = 100

# Bind the source and sink to the channel
a3.sources.r1.channels = c2
a3.sinks.k1.channel = c2

开启hadoop
[root@hadoop003 hadoop-2.7.2]$ sbin/start-dfs.sh
[root@hadoop003 hadoop-2.7.2]$ sbin/start-yarn.sh

开启监控
先开启和关闭流程都是先接收方，后发送方
[root@hadoop003 flume]$ bin/flume-ng agent --conf conf/ --name a3 --conf-file job/group1/flume-flume-dir.conf
[root@hadoop003 flume]$ bin/flume-ng agent --conf conf/ --name a2 --conf-file job/group1/flume-flume-hdfs.conf
[root@hadoop003 flume]$ bin/flume-ng agent --conf conf/ --name a1 --conf-file job/group1/flume-file-flume.conf
开启hive
[root@hadoop003 hive]$ bin/hive
hive (default)>
查看hdfs和本地的数据

2.负载均衡和故障转移

故障转移案例

flume-1

# Name the components on this agent
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1
a1.sinkgroups = g1

a1.sources.r1.type = netcat
a1.sources.r1.bind = hadoop003
a1.sources.r1.port = 44444

a1.sinks.k1.type = avro
a1.sinks.k1.hostname = hadoop003
a1.sinks.k1.port = 4545
a1.sinks.k2.type = avro
a1.sinks.k2.hostname = hadoop003
a1.sinks.k2.port = 4546

a1.sinkgroups.g1.sinks = k1 k2
a1.sinkgroups.g1.processor.type = failover
a1.sinkgroups.g1.processor.priority.k1 = 5
a1.sinkgroups.g1.processor.priority.k2 = 10
a1.sinkgroups.g1.processor.maxpenalty = 10000

a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100

a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1
a1.sinks.k2.channel = c1

flume-2

# Name the components on this agent
a2.sources = r1
a2.sinks = k1
a2.channels = c1

# Describe/configure the source
a2.sources.r1.type = avro
a2.sources.r1.bind = hadoop003
a2.sources.r1.port = 4545
# Describe the sink
a2.sinks.k1.type = logger

# Use a channel which buffers events in memory
a2.channels.c1.type = memory
a2.channels.c1.capacity = 1000
a2.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a2.sources.r1.channels = c1
a2.sinks.k1.channel = c1

flume-3

# Name the components on this agent
a3.sources = r1
a3.sinks = k1
a3.channels = c1

# Describe/configure the source
a3.sources.r1.type = avro
a3.sources.r1.bind = hadoop003
a3.sources.r1.port = 4546
# Describe the sink
a3.sinks.k1.type = logger

# Use a channel which buffers events in memory
a3.channels.c1.type = memory
a3.channels.c1.capacity = 1000
a3.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a3.sources.r1.channels = c1
a3.sinks.k1.channel = c1

负载均衡

# Name the components on this agent
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1
a1.sinkgroups = g1

a1.sources.r1.type = netcat
a1.sources.r1.bind = hadoop003
a1.sources.r1.port = 44444

a1.sinks.k1.type = avro
a1.sinks.k1.hostname = hadoop003
a1.sinks.k1.port = 4545
a1.sinks.k2.type = avro
a1.sinks.k2.hostname = hadoop003
a1.sinks.k2.port = 4546

a1.sinkgroups.g1.sinks = k1 k2
a1.sinkgroups.g1.processor.type = load_balance
a1.sinkgroups.g1.processor.backoff = true
a1.sinkgroups.g1.processor.selector = random

a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100

a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1
a1.sinks.k2.channel = c1

3.聚合（多数据源汇总案例）

本案例代码实现以三个flume都在hadoop003一台主机上完成，为了方便。
以后可以按下图在三台主机上完成，但都要安装flume

在hadoop003的/opt/module/flume/job目录下创建一个group2文件夹
[root@hadoop003 job]$ mkdir group2
配置flume1监控hive.log，配置sink输出数据到下一级flume3。
[root@hadoop003 group2]$ touch flume-file.conf
[root@hadoop003 group2]$ vim flume1.conf
添加配置内容

# Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1

# Describe/configure the source
a1.sources.r1.type = exec
a1.sources.r1.command = tail -F /opt/module/hive/logs/hive.log
a1.sources.r1.shell = /bin/bash -c

# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = hadoop003
a1.sinks.k1.port = 4141

# Describe the channel
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1

配置flume2监控端口44444数据流，配置sink数据到下一级flume3：
[root@hadoop003 group2]# vim flume2.conf
添加配置内容

#Name the components on this agent
a2.sources = r1
a2.sinks = k1
a2.channels = c1

# Describe/configure the source
a2.sources.r1.type = netcat
a2.sources.r1.bind = hadoop003
a2.sources.r1.port = 44444

# Describe the sink
a2.sinks.k1.type = avro
a2.sinks.k1.hostname = hadoop003
a2.sinks.k1.port = 4141

# Use a channel which buffers events in memory
a2.channels.c1.type = memory
a2.channels.c1.capacity = 1000
a2.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a2.sources.r1.channels = c1
a2.sinks.k1.channel = c1

配置flume3用于接收flume1与flume2发送过来的数据流，最终合并后sink到控制台。
[root@hadoop003 group2]# vim flume3.conf
添加配置内容

#Name the components on this agent
a3.sources = r1
a3.sinks = k1
a3.channels = c1

# Describe/configure the source
a3.sources.r1.type = avro
a3.sources.r1.bind = hadoop003
a3.sources.r1.port = 4141

# Describe the sink
# Describe the sink
a3.sinks.k1.type = logger

# Describe the channel
a3.channels.c1.type = memory
a3.channels.c1.capacity = 1000
a3.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a3.sources.r1.channels = c1
a3.sinks.k1.channel = c1

分别开启对应配置文件：flume3.conf，flume2.conf，flume1.conf。

开启flume3
bin/flume-ng agent --conf conf/ --name a3 --conf-file job/group2/flume3.conf -Dflume.root.logger=INFO,console
开启flume2
bin/flume-ng agent --conf conf/ --name a2 --conf-file job/group2/flume2.conf
登录到hadoop003向44444端口发送数据asd

查看flume3的监控

开启flume1
bin/flume-ng agent --conf conf/ --name a1 --conf-file job/group2/flume1.conf
[root@hadoop003 logs]# echo ‘111111’ >> hive.log

查看flume3的监控

4.自定义拦截器interceptor（多路复用）

Interceptor和Multiplexing ChannelSelector结合使用
实现：为每一个evnet加上header（kv对），在编写flume的conf文件中，在ChannelSelector位置让每个v分发到不同的channel

编写拦截器类，然后打jar包上传到虚拟机的flume/lib下
导入依赖

<dependencies>
        <dependency>
            <groupId>org.apache.flume</groupId>
            <artifactId>flume-ng-core</artifactId>
            <version>1.7.0</version>
        </dependency>
    </dependencies>

编写类

package com.demo;

import org.apache.flume.Context;
import org.apache.flume.Event;
import org.apache.flume.interceptor.Interceptor;

import java.util.ArrayList;
import java.util.List;
import java.util.Map;

/**
 * @author ljh
 * @create 2021-02-20 18:20
 */
public class TypeInterceptor implements Interceptor {
    //创建存放事件的集合
    List<Event> typeEvents;
    //初始化集合
    public void initialize() {
        typeEvents=new ArrayList<Event>();
    }
    //单个事件拦截
    public Event intercept(Event event) {
        //获取事件的头信息和body信息
        Map<String, String> headers = event.getHeaders();
        String body = new String(event.getBody());
        //根据body中数据是否包含hello来决定添加怎样的头信息
        if(body.contains("hello")){
            headers.put("type","good");
        }else{
            headers.put("type","bad");
        }
        return event;
    }
    //批量事件拦截
    public List<Event> intercept(List<Event> list) {
        //清空集合中上次传来的event
        typeEvents.clear();
        //给每一个event添加头信息
        for (Event event:list){
            Event e = intercept(event);
            typeEvents.add(e);
        }
        return typeEvents;
    }

    public void close() {

    }
    //类名随便起，使用时要写这个类名
    public static class Builder implements Interceptor.Builder{

        public Interceptor build() {
            return new TypeInterceptor();
        }

        public void configure(Context context) {

        }
    }
}

a1采集数据，多路复用传给a2,a3
此处的type值是根据自定义类的全类名+内部类的类名

# Name the components on this agent
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1 c2

# Describe/configure the source
a1.sources.r1.type = netcat
a1.sources.r1.bind = localhost
a1.sources.r1.port = 44444

a1.sources.r1.interceptors = i1
a1.sources.r1.interceptors.i1.type =com.demo.TypeInterceptor$Builder

a1.sources.r1.selector.type = multiplexing
a1.sources.r1.selector.header = type
a1.sources.r1.selector.mapping.good = c1
a1.sources.r1.selector.mapping.bad = c2
# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = hadoop003
a1.sinks.k1.port = 4545
a1.sinks.k2.type = avro
a1.sinks.k2.hostname = hadoop003
a1.sinks.k2.port = 4546


# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100

a1.channels.c2.type = memory
a1.channels.c2.capacity = 1000
a1.channels.c2.transactionCapacity = 100

# Bind the source and sink to the channel
a1.sources.r1.channels = c1 c2
a1.sinks.k1.channel = c1
a1.sinks.k2.channel = c2

a2打印到控制台（包含hello的）

a2.sources = r1
a2.sinks = k1
a2.channels = c1


a2.sources.r1.type = avro
a2.sources.r1.bind = hadoop003
a2.sources.r1.port = 4545

a2.sinks.k1.type = logger


a2.channels.c1.type = memory
a2.channels.c1.capacity = 1000
a2.channels.c1.transactionCapacity = 100


a2.sources.r1.channels = c1
a2.sinks.k1.channel = c1

a3打印到控制台（不包含hello的）

a3.sources = r1
a3.sinks = k1
a3.channels = c1


a3.sources.r1.type = avro
a3.sources.r1.bind = hadoop003
a3.sources.r1.port = 4546

a3.sinks.k1.type = logger


a3.channels.c1.type = memory
a3.channels.c1.capacity = 1000
a3.channels.c1.transactionCapacity = 100


a3.sources.r1.channels = c1
a3.sinks.k1.channel = c1

执行

flume-ng agent -n a1 -c /opt/module/flume-1.7.0/conf/ -f netcat-intercepter-avros.conf
flume-ng agent -n a2 -c /opt/module/flume-1.7.0/conf/ -f avro-flume-logs1.conf -Dflume.root.logger=INFO,console
flume-ng agent -n a3 -c /opt/module/flume-1.7.0/conf/ -f avro-flume-logs2.conf -Dflume.root.logger=INFO,console

5.自定义source

开发者文档http://flume.apache.org/releases/content/1.7.0/FlumeDeveloperGuide.html#source

实现：导入依赖，编写自定义source代码，打jar包放入flume/lib下，编写conf，在source部分引入自定义source全类名，就可以运行了。

需求：使用 flume 接收数据，并给每条数据添加前缀，输出到控制台。前缀可从 flume 配置文件中配置。
导入依赖

<dependencies>
<dependency>
<groupId>org.apache.flume</groupId>
<artifactId>flume-ng-core</artifactId>
<version>1.7.0</version>
</dependency>
</dependencies>

编写类

package com.demo;

import org.apache.flume.Context;
import org.apache.flume.Event;
import org.apache.flume.EventDeliveryException;
import org.apache.flume.PollableSource;
import org.apache.flume.conf.Configurable;
import org.apache.flume.event.SimpleEvent;
import org.apache.flume.source.AbstractSource;

/**
 * @author ljh
 * @create 2021-02-22 16:34
 */
public class MySource extends AbstractSource implements Configurable, PollableSource {
    private String prefix;
    private String subfix;
    //读取配置信息，给前后缀赋值
    public void configure(Context context) {
        prefix=context.getString("prefix");//读配置赋值
        subfix=context.getString("subfix","aaa");//给默认值（如果配置文件里赋值，先用配置文件的）
    }

    /**
     * 接收数据（此处我们用for循环造数据，以后可以通过io流和jdbc接受数据）
     * 将数据封装到event中
     * 把event传给channel
     * @return
     *
     */
    public Status process() throws EventDeliveryException {
        Status status=null;
        try {
            for (int i=0;i<5;i++){
                //构建事件
                SimpleEvent event = new SimpleEvent();//此行不能提到循环外，本来想提出去减少创建event对象，没想到不行
                //给event传入数据
                //前缀设置值，后缀用默认值
                event.setBody((prefix+"--"+i+"--"+subfix).getBytes());
                //将evnet传递给channel
                getChannelProcessor().processEvent(event);
                status=Status.READY;
            }
        } catch (Exception e) {
            e.printStackTrace();
            status=Status.BACKOFF;
        }
        //接受一次睡2s，免得发送太快不便于观察
        try {
            Thread.sleep(2000);
        } catch (InterruptedException e) {
            e.printStackTrace();
        }
        return status;
    }

    public long getBackOffSleepIncrement() {
        return 0;
    }

    public long getMaxBackOffSleepInterval() {
        return 0;
    }


}

配置文件

a1.sources = r1
a1.sinks = k1
a1.channels = c1

# Describe/configure the source
a1.sources.r1.type = com.demo.MySource
a1.sources.r1.prefix = hello
a1.sources.r1.subfix = ss


# Describe the sink
a1.sinks.k1.type = logger

# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1

执行
flume-ng agent -n a1 -c /opt/module/flume-1.7.0/conf/ -f mySource.conf -Dflume.root.logger=INFO,console
提示：loggersink每条日志的输出长度是有默认值的 官网：maxBytesToLog 16 Maximum number of bytes of the Event body to log

6.自定义 Sink

Sink 是完全事务性的。在从 Channel 批量删除数据之前，每个 Sink 用 Channel 启动一个事务。批量事件一旦成功写出到存储系统或下一个 Flume Agent，Sink 就利用 Channel 提交事务。事务一旦被提交，该 Channel 从自己的内部缓冲区删除事件。

需求：使用 flume 接收数据，并在 Sink 端给每条数据添加前缀和后缀，输出到控制台。前后缀可在 flume 任务配置文件中配置。

自定义类，打jar包上传

package com.demo;

import org.apache.flume.*;
import org.apache.flume.conf.Configurable;
import org.apache.flume.sink.AbstractSink;

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


/**
 * @author ljh
 * @create 2021-02-22 21:04
 */
public class MySink extends AbstractSink implements Configurable {
    //定义前后缀，来测试一下通过conf配置文件的kv赋值
    private String prefix;
    private String subfix;
    //创建logger对象
    private Logger logger = LoggerFactory.getLogger(MySink.class);
    public void configure(Context context) {
        prefix = context.getString("prefix");//从conf获取值
        subfix=context.getString("subfix");//给默认值

    }

    /**
     *获取channel
     * 从channel获取事务和event
     * 发送event
     */
    public Status process() throws EventDeliveryException {
        Status status= null;
        Channel channel = getChannel();
        Transaction transaction = channel.getTransaction();
        try {
            transaction.begin();
            Event event = channel.take();
            //处理event
            if(event!=null){
                String body = new String(event.getBody());
                //以logger形式输出，至于输出到控制台还是文件，要靠配置信息-Dflume.root.logger=INFO,console
                String tmp=prefix + body+subfix;//此处有问题，不出后缀，目前找不出问题所在
                logger.info(tmp);

            }


            transaction.commit();
            status=Status.READY;
        } catch (ChannelException e) {
            e.printStackTrace();
            transaction.rollback();
            status=Status.BACKOFF;
        }finally {
            transaction.close();
        }
        return status;
    }


}

编写conf

# Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1

# Describe/configure the source
a1.sources.r1.type = netcat
a1.sources.r1.bind = localhost
a1.sources.r1.port = 44444

# Describe the sink
a1.sinks.k1.type = com.demo.MySink
a1.sinks.k1.prefix=aaa
a1.sinks.k1.subfix=qq

# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100

# Bind the source and sink to the channel
a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1

启动agent
flume-ng agent -n a1 -c /opt/module/flume-1.7.0/conf/ -f mySink.conf -Dflume.root.logger=INFO,console