Hadoop生态圈-Flume的组件之自定义拦截器（interceptor）
作者：尹正杰
版权声明：原创作品，谢绝转载！否则将追究法律责任。
本篇博客只是举例了一个自定义拦截器的方法，测试字节传输速度。
1>.自定义interceptor方法
1 /* 2 @author :yinzhengjie 3 Blog:http://www.cnblogs.com/yinzhengjie/tag/Hadoop%E7%94%9F%E6%80%81%E5%9C%88/ 4 EMAIL:y1053419035@qq.com 5 */ 6 package cn.org.yinzhengjie.interceptor; 7 8 import org.apache.flume.Context; 9 import org.apache.flume.Event; 10 import org.apache.flume.interceptor.Interceptor; 11 12 import java.util.List; 13 14 /** 15 * 设置限速拦截器 16 * <p> 17 * 当 字节/时间，即同一时刻，如果进入的字节过多 18 * 则休眠一会 19 */ 20 public class MyInterceptor implements Interceptor { 21 22 private int speed; 23 24 25 //构造 26 private MyInterceptor(int speed) { 27 this.speed = speed; 28 } 29 30 31 //do nothing 32 public void initialize() { 33 34 } 35 36 /** 37 * 1、拿出上一个event的时间，和当前时间进行相减，得出上一个event的时间间隔 38 * 2、得到上一个event的body字节数 39 * 3、相除得到上一个event的速度，并在此event中先进行停留，再返回event 40 * 41 * @param event 42 * @return 43 */ 44 45 long lastTime = -1; 46 long lastBodySize = 0; 47 48 public Event intercept(Event event) { 49 50 51 byte[] body = event.getBody(); 52 int len = body.length; 53 54 55 long current = System.nanoTime(); 56 57 //第一个event 58 if (lastTime == -1) { 59 lastTime = current; 60 lastBodySize = len; 61 } 62 63 //非第一个event 64 else { 65 //计算上一个event停留的时间 66 long interval = current - lastTime; 67 System.out.println("=========================" + current + "/" + lastTime + "/" + interval + "========================="); 68 //上一个event的速度 69 int now_speed = (int) ((double) lastBodySize / interval * 1000); 70 if (now_speed > speed) { 71 System.out.println("=========================" + now_speed + "========================="); 72 //计算需要停留多少秒 线程休眠，时间 = shouldTime - interval 73 try { 74 Thread.sleep((lastBodySize / speed) * 1000 - interval); 75 } catch (InterruptedException e) { 76 e.printStackTrace(); 77 } 78 } 79 lastBodySize = len; 80 lastTime = System.currentTimeMillis(); 81 82 } 83 return event; 84 85 } 86 87 //迭代List<Event>，将所有Event交给intercept(Event)进行处理 88 public List<Event> intercept(List<Event> events) { 89 for (Event event : events) { 90 intercept(event); 91 } 92 return events; 93 } 94 95 //do nothing 96 public void close() { 97 98 } 99 100 public static class Builder implements Interceptor.Builder { 101 102 private int speed; 103 104 public void configure(Context context) { 105 speed = context.getInteger(Constants.SPEED, Constants.DEFAULT_SPEED); 106 107 } 108 109 public Interceptor build() { 110 return new MyInterceptor(speed); 111 } 112 } 113 114 public static class Constants { 115 public static String SPEED = "speed"; 116 public static int DEFAULT_SPEED = 1; 117 118 } 119 }
2>.打包并将其发送到 /soft/flume/lib下
[yinzhengjie@s101 ~]$ cd /soft/flume/lib/ [yinzhengjie@s101 lib]$ [yinzhengjie@s101 lib]$ ll | grep MyFlume -rw-r--r-- 1 yinzhengjie yinzhengjie 5231 Jun 20 18:53 MyFlume-1.0-SNAPSHOT.jar [yinzhengjie@s101 lib]$ [yinzhengjie@s101 lib]$ rm -rf MyFlume-1.0-SNAPSHOT.jar [yinzhengjie@s101 lib]$ [yinzhengjie@s101 lib]$ rz[yinzhengjie@s101 lib]$ [yinzhengjie@s101 lib]$ ll | grep MyFlume -rw-r--r-- 1 yinzhengjie yinzhengjie 8667 Jun 20 21:02 MyFlume-1.0-SNAPSHOT.jar [yinzhengjie@s101 lib]$ [yinzhengjie@s101 lib]$
3>.编写agent的配置文件
[yinzhengjie@s101 ~]$ more /soft/flume/conf/yinzhengjie_myInterceptor.conf # Name the components on this agent a1.sources = r1 a1.sinks = k1 a1.channels = c1# 定义源: seq a1.sources.r1.type = seq # 定义一次RPC产生的批次数量 a1.sources.r1.batchSize = 1024# 指定添加拦截器 a1.sources.r1.interceptors = i1 a1.sources.r1.interceptors.i1.type = cn.org.yinzhengjie.interceptor.MyInterceptor$Builder a1.sources.r1.interceptors.i1.speed = 1# Describe the sink a1.sinks.k1.type = logger# Use a channel which buffers events in memory a1.channels.c1.type = memory a1.channels.c1.capacity = 100000 a1.channels.c1.transactionCapacity = 10000# Bind the source and sink to the channel a1.sources.r1.channels = c1 a1.sinks.k1.channel = c1 [yinzhengjie@s101 ~]$
4>.启动flume并测试
[yinzhengjie@s101 ~]$ flume-ng agent -f /soft/flume/conf/yinzhengjie_myInterceptor.conf -n a1
下图是运行agent部分的输出内容
