flume-kafka-spark streaming(pysp

学习了差不多一个星期，终于把flume-kafka-spark streaming贯通了，直接上流程图：

至于为什么要这样，当然是方便咯
参考某博客

一、环境部署

hadoop集群2.7.1
zookeerper集群
kafka集群：kafka_2.11-0.10.0.0
spark集群：spark-2.0.1-bin-hadoop2.7.tgz
flume1.7.0
环境搭建可参考我前面几篇文章。不再赘述
三台机器：master，slave1,slave2

二、配置flume

sink为kafka
source你随意

a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5140
a1.sources.r1.host = 192.168.31.131
a1.sources.r1.channels = c1
# Describe the sink
a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink
a1.sinks.k1.topic = test5
a1.sinks.k1.brokerList = 192.168.31.131:9092
a1.sinks.k1.requiredAcks = 1
a1.sinks.k1.batchSize = 20
a1.sinks.k1.channel = c1
# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100
#yi bai tiao jiu submit

三、编程,KafkaWordCount.py

编写spark steaming 代码，读取kafka流数据，并统计词频

# -*- coding: UTF-8 -*-
 ###spark streaming&&kafka
from pyspark import SparkContext
from pyspark.streaming import StreamingContext
from pyspark.streaming.kafka import KafkaUtils

sc=SparkContext("local[2]","KafkaWordCount")
#处理时间间隔为1s
ssc=StreamingContext(sc,2)
zookeeper="192.168.31.131:2181,192.168.31.132:2181,192.168.31.133:2181"
#打开一个TCP socket 地址 和 端口号
topic={"test5":0,"test5":1,"test5":2}
groupid="test-consumer-group"
lines = KafkaUtils.createStream(ssc, zookeeper,groupid,topic)
lines1=lines.map(lambda x:x[1])

#对1s内收到的字符串进行分割
words=lines1.flatMap(lambda line:line.split(" "))

#映射为（word，1）元祖
pairs=words.map(lambda word:(word,1))

wordcounts=pairs.reduceByKey(lambda x,y:x+y)

#输出文件，前缀+自动加日期
wordcounts.saveAsTextFiles("/tmp/fkafka")

wordcounts.pprint()

#启动spark streaming应用
ssc.start()
#等待计算终止
ssc.awaitTermination()

四、启动环境

1.启动hadoop集群

start-all.sh

2.启动spark集群

start-master.sh
start-slaves.sh

3.启动zookeeper集群

在三台机器下均输入以下命令

zkServer.sh start

4.启动kafka集群

在三台机器下均输入以下命令

kafka-server-start.sh -daemon ../config/server.properties

5.jps查看进程

master:

slave1与slave2一样:

确保所有进程启动

6.创建kafka topic

kafka-topics.sh --create --zookeeper 192.168.31.131:2181,192.168.31.132:2181,192.168.31.133:2181 --replication-factor 3 --partitions 3 --topic test5

7.启动flume agent

flume-ng agent --conf flume/conf/  -f /home/cms/flume/conf/flume-conf.properties -n a1 -Dflume.root.logger=INFO,console

五、测试

1.运行KafkaWordCount.py

在master下
运行

spark-submit --jars kafka/libs/spark-streaming-kafka-0-8-assembly_2.11-2.0.1.jar KafkaWordCount.py 2> error.txt

2.发送数据

echo "hello'\t'word" | nc 192.168.31.131 5140

3.观察终端输出

六、hdfs上查看输出

hadoop fs -ls /tmp/fkafka*

参考文档
(flume-kafka- spark streaming(pyspark) - redis 实时日志收集实时计算)[http://blog.csdn.net/zhong_han_jun/article/details/50721736]

http://spark.apache.org/docs/latest/streaming-kafka-0-8-integration.html