Flume安装及几个入门案例

安装

1、下载
http://www.apache.org/dyn/closer.lua/flume/1.8.0/apache-flume-1.8.0-bin.tar.gz
2、上传到指定的服务器(master)中的某个目录
3、解压
tar -xvf apache-flume-1.8.0-bin.tar.gz
4、cd apache-flume-1.8.0-bin/conf
5、cp flume-conf.properties.template flume-conf.properties
6、vi flume-conf.properties

几个小案例带你了解flume

一个理论:flume每一次接收到的数据称为一个事件
说在前面telnet 需要自己去装,然后启动:telnet localhost 44445

案例一

接收netcat的数据展示到console

## 定义 sources、channels 以及 sinks
# 数据源
agent1.sources = netcatSrc
# 收集的数据放在那里,memoryChannel这个是放在内存中
agent1.channels = memoryChannel
# 数据放在那里loggerSink这个是以log的方式打印在控制台上
agent1.sinks = loggerSink

## netcatSrc 的配置
# 启动的服务
agent1.sources.netcatSrc.type = netcat
# Ip是什么
agent1.sources.netcatSrc.bind = localhost
# 在哪个端口上
agent1.sources.netcatSrc.port = 44445

## loggerSink 的配置
# 以log的方式写到控制台上
agent1.sinks.loggerSink.type = logger

## memoryChannel 的配置
agent1.channels.memoryChannel.type = memory
# 能装100条记录
agent1.channels.memoryChannel.capacity = 100

## 通过 memoryChannel 连接 netcatSrc 和 loggerSink
agent1.sources.netcatSrc.channels = memoryChannel
agent1.sinks.loggerSink.channel = memoryChannel


# agent 一个代理  conf用conf的方式启动 agent1就是配置里的那个
# -Dflume.root.logger=INFO,console   info的方式打印在控制台
bin/flume-ng agent --conf conf --conf-file conf/flume-conf.properties --name agent1 -Dflume.root.logger=INFO,console
接收到的数据先放内存,再打印到控制台上

案例二

接收telnet的数据,每五次事件存一次hdfs

## 定义 sources、channels 以及 sinks
agent1.sources = netcatSrc
# memoryChannel 可能会丢失数据  应该把它换成文件的形式
agent1.channels = memoryChannel
# 这个是把它放到hdfs去
agent1.sinks = hdfsSink

## netcatSrc 的配置
agent1.sources.netcatSrc.type = netcat
agent1.sources.netcatSrc.bind = localhost
agent1.sources.netcatSrc.port = 44445

## hdfsSink 的配置
# 保存到hdfs上去
agent1.sinks.hdfsSink.type = hdfs
# 放到这个目录下去
agent1.sinks.hdfsSink.hdfs.path = hdfs://master:8020/user/hadoop-jrq/spark-course/steaming/flume/%y-%m-%d
# 每5次flume事件写一次HDFS
agent1.sinks.hdfsSink.hdfs.batchSize = 5
# 这个一定位true,否则会报错
agent1.sinks.hdfsSink.hdfs.useLocalTimeStamp = true

## memoryChannel 的配置
agent1.channels.memoryChannel.type = memory
agent1.channels.memoryChannel.capacity = 100

## 通过 memoryChannel 连接 netcatSrc 和 hdfsSink
agent1.sources.netcatSrc.channels = memoryChannel
agent1.sinks.hdfsSink.channel = memoryChannel

启动:
bin/flume-ng agent --conf conf --conf-file conf/flume-conf.properties --name agent1 

案例三

监控一个文件,每五次事件写一次hdfs

## 定义 sources、channels 以及 sinks
agent1.sources = logSrc
# 数据接收到后放在文件中去
agent1.channels = fileChannel
agent1.sinks = hdfsSink

## logSrc 的配置
agent1.sources.logSrc.type = exec
# -F参数会有一个小bug因此,我使用-f0
agent1.sources.logSrc.command = tail -f0 /home/hadoop-jrq/spark-course/steaming/flume-course/demo3/logs/webserver.log

## hdfsSink 的配置
agent1.sinks.hdfsSink.type = hdfs
agent1.sinks.hdfsSink.hdfs.path = hdfs://master:8020/user/hadoop-jrq/spark-course/steaming/flume/%y-%m-%d
agent1.sinks.hdfsSink.hdfs.batchSize = 5
agent1.sinks.hdfsSink.hdfs.useLocalTimeStamp = true

## fileChannel 的配置
agent1.channels.fileChannel.type = file
agent1.channels.fileChannel.checkpointDir = /home/hadoop-jrq/spark-course/steaming/flume-course/demo2-2/checkpoint
agent1.channels.fileChannel.dataDirs = /home/hadoop-jrq/spark-course/steaming/flume-course/demo2-2/data

## 通过 fileChannel 连接 logSrc 和 hdfsSink
agent1.sources.logSrc.channels = fileChannel
agent1.sinks.hdfsSink.channel = fileChannel


日志文件收集:实时的保存到hdfs中去
实时接收日志信息,然后保存到hdfs中去

bin/flume-ng agent --conf conf --conf-file conf/flume-conf.properties --name agent1 

echo testdata >> webserver.log

案例四,一个简单的模拟网站实时流处理

监控日志文件,写入kafka,然后再用spark streaming处理数据
flume配置:

agent1.sources = s1
agent1.channels = c1
agent1.sinks = k1

agent1.sources.s1.type = exec
agent1.sources.s1.command = tail -fn0 /home/hadoop-jrq/spark-course/steaming/weblog.log

agent1.channels.c1.type = memory
agent1.channels.c1.capacity = 1000
agent1.channels.c1.transactionCapacity = 100

agent1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink
# kafka的主题
agent1.sinks.k1.kafka.topic = pageview
# kafka集群的地址
agent1.sinks.k1.kafka.bootstrap.servers = master:9092,slave1:9092,slave2:9092
# 每20次事件发送一次
agent1.sinks.k1.kafka.flumeBatchSize = 20
# ack机制的选用,不懂的,可以去看我kafka的ack机制文章
agent1.sinks.k1.kafka.producer.acks = 1
agent1.sinks.k1.kafka.producer.linger.ms = 1
# 压缩机制
agent1.sinks.k1.kafka.producer.compression.type = snappy

agent1.sources.s1.channels = c1
agent1.sinks.k1.channel = c1

启动flume:bin/flume-ng agent --conf conf --conf-file conf/flume-conf.properties --name agent1 
# 模拟数据发送
echo testdata >> webserver.log

依赖的jar包:

<dependency>
            <groupId>org.apache.spark</groupId>
            <artifactId>spark-core_2.11</artifactId>
            <version>2.2.0</version>
<!--            <scope>provided</scope>-->
        </dependency>
        <dependency>
            <groupId>org.apache.spark</groupId>
            <artifactId>spark-streaming_2.11</artifactId>
            <version>2.2.0</version>
<!--            <scope>provided</scope>-->
        </dependency>
        <dependency>
            <groupId>org.apache.spark</groupId>
            <artifactId>spark-streaming-kafka-0-8_2.11</artifactId>
            <version>2.2.0</version>
        </dependency>
         <dependency>
            <groupId>redis.clients</groupId>
            <artifactId>jedis</artifactId>
            <version>2.5.2</version>
        </dependency>

spark streaming代码:

import com.jrq.streaming.InternalRedisClient
import kafka.serializer.StringDecoder
import org.apache.spark.streaming.kafka.KafkaUtils
import org.apache.spark.streaming.{Seconds, StreamingContext}
import org.apache.spark.{SparkConf, SparkContext}

object PageViewStream {
  def main(args: Array[String]) {
    if (args.length < 2) {
      printIn("参数不够")
      System.exit(1)
    }
    println("开始")
    val sparkConf = new SparkConf()
    if (!sparkConf.contains("spark.master")){
     // 这个是因为我本地的 环境有点问题,手动加上的Hadoop环境,否则会报错,如果你的不会报错,就不需要加上
      System.setProperty("hadoop.home.dir", "D:\\hadoop-common-2.2.0-bin-master")
        // 我这是本地启动,这个程序最少要5个core才能启动,3个core用于接收数据,2个core用于处理数据
      sparkConf.setMaster("local[6]")
    }
    sparkConf.setAppName("DirectKafkaWordCount")
    
    //    val sparkConf = new SparkConf().setAppName("DirectKafkaWordCount")
    val sc = new SparkContext(sparkConf)

    val Array(brokers, topics) = args

    // Create the context
    val ssc = new StreamingContext(sc, Seconds(1))
    // 这里是因为我配置了HA的名字,因此程序中需要这么写,否则会有bug,就是当你master挂掉的时候,你的程序会报错,当然在本地测试的时候会有问题,换成hdfs://master:8020/就行了
  ssc.checkpoint("hdfs://mycluster/user/hadoop-jrq/spark-course/streaming/checkpoint")
    // 从kafka中读取数据
    val topicsSet = topics.split(",").toSet
    val kafkaParams = Map[String, String]("metadata.broker.list" -> brokers)
    // 初始化一个kafka Direct 模式
    val messages = KafkaUtils.createDirectStream[String, String, StringDecoder, StringDecoder](
      ssc, kafkaParams, topicsSet)

    // 且将每一行的字符串转换成PageView对象
    val pageViews = messages.map { case (_, value) => PageView.fromString(value) }

    // 每隔3秒统计前50秒内每一个URL的访问错误数
    val badCount = pageViews.window(Seconds(50), Seconds(3))
      .filter(_.status != 200).count()

    // 每隔两秒统计前15秒内有多少个用户访问了url:https:foo.com/
    val fooCount = pageViews.window(Seconds(15), Seconds(2)).filter(_.url != "https://foo.com/").map(view => (view.userID, 1)).groupByKey().count()

    // 每隔5秒统计前30秒内有多少活跃用户
    val activeUserCount = pageViews.window(Seconds(30), Seconds(1))
      .map(view => (view.userID, 1))
      .groupByKey()
      .count()

    activeUserCount.print()

    activeUserCount.foreachRDD { rdd =>
      // 对每个分区进行遍历,写入redis
      rdd.foreachPartition { partitionRecords =>
        // redis 的连接池
        val jedis = InternalRedisClient.getPool.getResource
        partitionRecords.foreach { count =>
          jedis.set("active_user_count", count.toString)
        }
        // 写完后吧这个连接还给连接池
        InternalRedisClient.getPool.returnResource(jedis)
      }
    }

    ssc.start()
    ssc.awaitTermination()
  }
}

/*
  表示访问一个网站页面的行为日志数据,格式为:
  url status zipCode userID\n
 */
case class PageView(val url: String, val status: Int, val zipCode: Int, val userID: Int)
  extends Serializable

object PageView extends Serializable {
  def fromString(in: String): PageView = {
    val parts = in.split(" ")
    new PageView(parts(0), parts(1).toInt, parts(2).toInt, parts(3).toInt)
  }
}

Redis连接池代码:

import org.apache.commons.pool2.impl.GenericObjectPoolConfig
import redis.clients.jedis.JedisPool

/**
  * Internal Redis client for managing Redis connection {@link Jedis} based on {@link RedisPool}
  */
object InternalRedisClient extends Serializable {

  @transient private lazy val pool: JedisPool = {
    val maxTotal = 10
    val maxIdle = 10
    val minIdle = 1
    val redisHost = "192.168.240.186"
    val redisPort = 6379
    val redisTimeout = 30000
    val poolConfig = new GenericObjectPoolConfig()
    poolConfig.setMaxTotal(maxTotal)
    poolConfig.setMaxIdle(maxIdle)
    poolConfig.setMinIdle(minIdle)
    poolConfig.setTestOnBorrow(true)
    poolConfig.setTestOnReturn(false)
    poolConfig.setMaxWaitMillis(100000)


    val hook = new Thread{
      override def run = pool.destroy()
    }
    sys.addShutdownHook(hook.run)

    new JedisPool(poolConfig, redisHost, redisPort, redisTimeout)
  }

  def getPool: JedisPool = {
    assert(pool != null)
    pool
  }

  def main(args: Array[String]): Unit = {
    val p = getPool

    val j = p.getResource

    j.set("much", "")
  }
}

集群启动的方式

/home/hadoop-jrq/bigdata/spark-2.2 .0-bin-hadoop2 .7/bin/spark-submit
--class com.jrq.streaming.kafka.PageViewStream \
-- master spark :// master: 7077 \
-- deploy - mode client \
-- driver - memory 512 m \
-- executor - memory 512 m \
-- total - executor - cores 5 \
-- executor - cores 2 \
/home/hadoop-jrq/spark-course/steaming/spark-streaming-datasource-1.0-SNAPSHOT-jar-with-dependencies.jar \
master: 9092, slave1: 9092, slave2: 9092  pageview

模拟发送日志的方式:
echo http:// baidu.com 200 899787 22224 >> weblog.log

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值