flume采集数据到kafka和hive

本文主要是介绍flume采集数据到kafka和hive，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

1. 构建ftp服务
   在安装flume的机器上添加sftp服务

useradd flumetest
passwd flumetest

#ubuntu-查看所有用户
cat /etc/shadow

apt-get install vsftpd
#查看
service vsftpd status
#创建接受数据目录
mkdir /home/flumetest/alarm

在vsftpd服务配置文件中设置：

# Allowanonymous FTP? (Disabled by default)
anonymous_enable=NO#Uncomment this to enable any form of FTP write command.
write_enable=YES#chroot_local_user=YES
chroot_list_enable=YES
#(default follows)
chroot_list_file=/etc/vsftpd.chroot_list

1. 配置kafka
   参考网址：http://kafka.apache.org/quickstart

#启动kafka，kafka节点都需要启动
nohup sh bin/kafka-server-start.sh config/server.properties > /dev/null2>&1 &#创建topic
bin/kafka-topics.sh --create --zookeeper 116.62.*.*:2181--replication-factor 2 --partitions 2 --topic alarm#查看topic List
bin/kafka-topics.sh --list --zookeeper 116.62.*.*:2181启动consumer，查看数据
bin/kafka-console-consumer.sh --bootstrap-server 116.62.*.*:9092--topic alarm --from-beginning#删除topic
bin/kafka-topics.sh --zookeeper 116.62.*.*:2181 --delete --topic alarm
#如果kafaka启动时加载的配置文件中server.properties没有配置delete.topic.enable=true，那么此时的删除并不是真正的删除，而是把topic标记为：marked for deletion，此时你若想真正删除它，可以如下操作：
#（1）登录zookeeper客户端：命令：./bin/zookeeper-client
#（2）找到topic所在的目录：ls /brokers/topics
#（3）找到要删除的topic，执行命令：rmr /brokers/topics/【topic name】即可，此时topic被彻底删除。

3 创建hive表

create table if not EXISTS alarm_no_partition(alm_timestring,alm_timeMs int,tag_NameID string,alm_Type int,priID int,alm_Ack_Timestring,alm_Ack_TimeMs int,alm_Group int,alm_Sub_Area int,tag_Data_Typestring,alm_Ack_Flg string,alm_Remove_Flg string,alm_Remove_Timestring,alm_Remove_TimeMs int,alarm_date string) clusteredby (priID) into 2 buckets stored as orc TBLPROPERTIES("transactional"="true");

4 配置flume
参考网址：http://flume.apache.org/FlumeUserGuide.html
配置flume-conf.properties文件：一个source，多个sink、

channel
agent1.sources= alarms1
agent1.channels= alarmc1 alarmc2 alarmc3
agent1.sinks= alarmk1 alarmk2 alarmk3#SpoolingDirectory
#setalarms1
agent1.sources.alarms1.type=spooldir
agent1.sources.alarms1.spoolDir=/home/flumetest/alarm/
agent1.sources.alarms1.channels=alarmc1alarmc2 alarmc3
agent1.sources.alarms1.fileHeader= false#setalarmk1
agent1.sinks.alarmk1.channel=alarmc1
agent1.sinks.alarmk1.type= org.apache.flume.sink.kafka.KafkaSink
agent1.sinks.alarmk1.topic= alarm
agent1.sinks.alarmk1.kafka.bootstrap.servers= 116.62.*.*:9092;116.62.*.*:9092;116.62.*.*:9092
agent1.sinks.alarmk1.kafka.flumeBatchSize= 20
agent1.sinks.alarmk1.kafka.producer.acks= 1
agent1.sinks.alarmk1.kafka.producer.linger.ms= 1
agent1.sinks.alarmk1.kafka.producer.compression.type= snappy
#setalarmk2
agent1.sinks.alarmk2.channel=alarmc2
agent1.sinks.alarmk2.type=hive
agent1.sinks.alarmk2.hive.metastore= thrift://127.0.0.1:9083
agent1.sinks.alarmk2.hive.database= alarm
agent1.sinks.alarmk2.hive.table= alarm_no_partition
#agent1.sinks.alarmk2.hive.partition=%{alarm_date}
agent1.sinks.alarmk2.useLocalTimeStamp= false
#agent1.sinks.alarmk2.roundValue= 10
#agent1.sinks.alarmk2.roundUnit= minute
agent1.sinks.alarmk2.serializer= DELIMITED
agent1.sinks.alarmk2.serializer.delimiter=,
agent1.sinks.alarmk2.serializer.serdeSeparator='\t'
agent1.sinks.alarmk2.serializer.fieldnames=alm_time,alm_timems,tag_nameid,alm_type,priid,alm_ack_time,alm_ack_timems,alm_group,alm_sub_area,tag_data_type,alm_ack_flg,alm_remove_flg,alm_remove_time,alm_remove_timems,alarm_date
#setalarmk3
#setalarmk3
#agent1.sinks.alarmk3.channel=alarmc3
#agent1.sinks.alarmk3.type=hbase
#agent1.sinks.alarmk3.table=alarm_test
#agent1.sinks.alarmk3.columnFamily=
#agent1.sinks.alarmk3.serializer=org.apache.flume.sink.hbase.RegexHbaseEventSerializer#setalarmc1
agent1.channels.alarmc1.type= memory
agent1.channels.alarmc1.capacity= 1000 
agent1.channels.alarmc1.transactionCapacity= 100
#setalarmc2
agent1.channels.alarmc2.type= memory
agent1.channels.alarmc2.capacity= 1000 
agent1.channels.alarmc2.transactionCapacity= 100
#setalarmc3
agent1.channels.alarmc3.type= memory
agent1.channels.alarmc3.capacity= 1000 
agent1.channels.alarmc3.transactionCapacity= 100

启动flume

nohup bin/flume-ng agent -cconf -f conf/flume-conf.properties -n agent1 -Dflume.root.logger=INFO,LOGFILE-Dflume.log.dir=logs >> /dev/null 2>&1

5 加载数据
向目录中添加文件加载完成文件后缀添加.COMPLETED

这篇关于flume采集数据到kafka和hive的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！

---