Flume使用

一、Flume简介

1. Flume提供一个分布式的，可靠的，对大数据量的日志进行高效收集、聚集、移动的服务，Flume只能在Unix环境下运行。
2. Flume基于流式架构，容错性强，也很灵活简单。
3. Flume、Kafka用来实时进行数据收集，Spark、Storm用来实时处理数据，impala用来实时查询。

二、Flume角色

2.1、Source
用于采集数据，Source是产生数据流的地方，同时Source会将产生的数据流传输到Channel，这个有点类似于Java IO部分的Channel。
2.2、Channel
用于桥接Sources和Sinks，类似于一个队列。
2.3、Sink
从Channel收集数据，将数据写到目标源(可以是下一个Source，也可以是HDFS或者HBase)。
2.4、Event
传输单元，Flume数据传输的基本单元，以事件的形式将数据从源头送至目的地。

三、Flume传输过程

source监控某个文件或数据流，数据源产生新的数据，拿到该数据后，将数据封装在一个Event中，并put到channel后commit提交，channel队列先进先出，sink去channel队列中拉取数据，然后写入到HDFS中。

四、Flume部署及使用

4.1、文件配置

flume-env.sh涉及修改项：

export JAVA_HOME=/home/admin/modules/jdk1.8.0_121

4.2、案例

4.2.1、案例一：监控端口数据

目标：Flume监控一端Console，另一端Console发送消息，使被监控端实时显示。
分步实现：

1. 安装telnet工具

$ sudo rpm -ivh xinetd-2.3.14-40.el6.x86_64.rpm
$ sudo rpm -ivh telnet-0.17-48.el6.x86_64.rpm
$ sudo rpm -ivh telnet-server-0.17-48.el6.x86_64.rpm

2. 创建Flume Agent配置文件flume-telnet.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 = 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

3. 判断44444端口是否被占用

$ netstat -tunlp | grep 44444

4. 先开启flume先听端口

$ bin/flume-ng agent --conf conf/ --name a1 --conf-file job/flume-telnet.conf -Dflume.root.logger==INFO,console

5. 使用telnet工具向本机的44444端口发送内容

$ telnet localhost 44444

4.2.2、案例二：实时读取本地文件到HDFS

目标：实时监控hive日志，并上传到HDFS中
分步实现：

1. 拷贝Hadoop相关jar到Flume的lib目录下（要学会根据自己的目录和版本查找jar包）

$ cp share/hadoop/common/lib/hadoop-auth-2.5.0-cdh5.3.6.jar ./lib/
$ cp share/hadoop/common/lib/commons-configuration-1.6.jar ./lib/
$ cp share/hadoop/mapreduce1/lib/hadoop-hdfs-2.5.0-cdh5.3.6.jar ./lib/
$ cp share/hadoop/common/hadoop-common-2.5.0-cdh5.3.6.jar ./lib/

$ cp ./share/hadoop/hdfs/lib/htrace-core-3.1.0-incubating.jar ./lib/
$ cp ./share/hadoop/hdfs/lib/commons-io-2.4.jar ./lib/

尖叫提示：标红的jar为1.99版本flume必须引用的jar
2) 创建flume-hdfs.conf文件

# Name the components on this agent
a2.sources = r2
a2.sinks = k2
a2.channels = c2
# Describe/configure the source
a2.sources.r2.type = exec
a2.sources.r2.command = tail -F /home/admin/modules/apache-hive-1.2.2-bin/hive.log
a2.sources.r2.shell = /bin/bash -c

# Describe the sink
a2.sinks.k2.type = hdfs
a2.sinks.k2.hdfs.path = hdfs://linux01:8020/flume/%Y%m%d/%H
#上传文件的前缀
a2.sinks.k2.hdfs.filePrefix = logs-
#是否按照时间滚动文件夹
a2.sinks.k2.hdfs.round = true
#多少时间单位创建一个新的文件夹
a2.sinks.k2.hdfs.roundValue = 1
#重新定义时间单位
a2.sinks.k2.hdfs.roundUnit = hour
#是否使用本地时间戳
a2.sinks.k2.hdfs.useLocalTimeStamp = true
#积攒多少个Event才flush到HDFS一次
a2.sinks.k2.hdfs.batchSize = 1000
#设置文件类型，可支持压缩
a2.sinks.k2.hdfs.fileType = DataStream
#多久生成一个新的文件
a2.sinks.k2.hdfs.rollInterval = 600
#设置每个文件的滚动大小
a2.sinks.k2.hdfs.rollSize = 134217700
#文件的滚动与Event数量无关
a2.sinks.k2.hdfs.rollCount = 0
#最小冗余数
a2.sinks.k2.hdfs.minBlockReplicas = 1

# Use a channel which buffers events in memory
a2.channels.c2.type = memory
a2.channels.c2.capacity = 1000
a2.channels.c2.transactionCapacity = 100

# Bind the source and sink to the channel
a2.sources.r2.channels = c2
a2.sinks.k2.channel = c2

3. 执行监控配置

$ bin/flume-ng agent --conf conf/ --name a2 --conf-file job/flume-hdfs.conf

4.2.3、案例三：实时读取目录文件到HDFS

目标：使用flume监听整个目录的文件
分步实现：

1. 创建配置文件flume-dir.conf

a3.sources = r3
a3.sinks = k3
a3.channels = c3

# Describe/configure the source
a3.sources.r3.type = spooldir
a3.sources.r3.spoolDir = /opt/module/flume-1.7.0/upload
a3.sources.r3.fileSuffix = .COMPLETED
a3.sources.r3.fileHeader = true
#忽略所有以.tmp结尾的文件，不上传
a3.sources.r3.ignorePattern = ([^ ]*\.tmp)

# Describe the sink
a3.sinks.k3.type = hdfs
a3.sinks.k3.hdfs.path = hdfs://hadoop102:9000/flume/upload/%Y%m%d/%H
#上传文件的前缀
a3.sinks.k3.hdfs.filePrefix = upload-
#是否按照时间滚动文件夹
a3.sinks.k3.hdfs.round = true
#多少时间单位创建一个新的文件夹
a3.sinks.k3.hdfs.roundValue = 1
#重新定义时间单位
a3.sinks.k3.hdfs.roundUnit = hour
#是否使用本地时间戳
a3.sinks.k3.hdfs.useLocalTimeStamp = true
#积攒多少个Event才flush到HDFS一次
a3.sinks.k3.hdfs.batchSize = 100
#设置文件类型，可支持压缩
a3.sinks.k3.hdfs.fileType = DataStream
#多久生成一个新的文件
a3.sinks.k3.hdfs.rollInterval = 60
#设置每个文件的滚动大小大概是128M
a3.sinks.k3.hdfs.rollSize = 134217700
#文件的滚动与Event数量无关
a3.sinks.k3.hdfs.rollCount = 0
#最小冗余数
a3.sinks.k3.hdfs.minBlockReplicas = 1

# Use a channel which buffers events in memory
a3.channels.c3.type = memory
a3.channels.c3.capacity = 1000
a3.channels.c3.transactionCapacity = 100

# Bind the source and sink to the channel
a3.sources.r3.channels = c3
a3.sinks.k3.channel = c3

2. 执行测试：执行如下脚本后，请向upload文件夹中添加文件试试

$ bin/flume-ng agent --conf conf/ --name a3 --conf-file job/flume-dir.conf

尖叫提示： 在使用Spooling Directory Source时

1. 不要在监控目录中创建并持续修改文件
2. 上传完成的文件会以.COMPLETED结尾
3. 被监控文件夹每600毫秒扫描一次文件变动

4.2.4、案例四：Flume与Flume之间数据传递：单Flume多Channel、Sink，

目标：使用flume-1监控文件变动，flume-1将变动内容传递给flume-2，flume-2负责存储到HDFS。同时flume-1将变动内容传递给flume-3，flume-3负责输出到。
local filesystem。
分步实现：

1. 创建flume-1.conf，用于监控hive.log文件的变动，同时产生两个channel和两个sink分别输送给flume-2和flume3：

# 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/hive.log
a1.sources.r1.shell = /bin/bash -c

# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = hadoop102
a1.sinks.k1.port = 4141

a1.sinks.k2.type = avro
a1.sinks.k2.hostname = hadoop102
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

2. 创建flume-2.conf，用于接收flume-1的event，同时产生1个channel和1个sink，将数据输送给hdfs：

# 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 = hadoop102
a2.sources.r1.port = 4141

# Describe the sink
a2.sinks.k1.type = hdfs
a2.sinks.k1.hdfs.path = hdfs://hadoop102:9000/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 = 60
#设置每个文件的滚动大小大概是128M
a2.sinks.k1.hdfs.rollSize = 134210000
#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


# 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

3. 创建flume-3.conf，用于接收flume-1的event，同时产生1个channel和1个sink，将数据输送给本地目录：

# 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 = hadoop102
a3.sources.r1.port = 4142

# Describe the sink
a3.sinks.k1.type = file_roll
a3.sinks.k1.sink.directory = /opt/module/flume1.7.0/job/flume-only/output

# 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

尖叫提示：输出的本地目录必须是已经存在的目录，如果该目录不存在，并不会创建新的目录。
4) 执行测试：分别开启对应flume-job（依次启动flume-1，flume-2，flume-3），同时产生文件变动并观察结果：

$ bin/flume-ng agent --conf conf/ --name a1 --conf-file job/group-job1/flume-1.conf
$ bin/flume-ng agent --conf conf/ --name a2 --conf-file job/group-job1/flume-2.conf
$ bin/flume-ng agent --conf conf/ --name a3 --conf-file job/group-job1/flume-3.conf

4.2.5、案例五：Flume与Flume之间数据传递，多Flume汇总数据到单Flume

目标：flume-1监控文件hive.log，flume-2监控某一个端口的数据流，flume-1与flume-2将数据发送给flume-3，flume3将最终数据写入到HDFS。
分步实现：

1. 创建flume-1.conf，用于监控hive.log文件，同时sink数据到flume-3：

# 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 /home/admin/modules/apache-hive-1.2.2-bin/hive.log
a1.sources.r1.shell = /bin/bash -c

# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = linux01
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

2. 创建flume-2.conf，用于监控端口44444数据流，同时sink数据到flume-3：

# 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 = linux01
a2.sources.r1.port = 44444

# Describe the sink
a2.sinks.k1.type = avro
a2.sinks.k1.hostname = linux01
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

3. 创建flume-3.conf，用于接收flume-1与flume-2发送过来的数据流，最终合并后sink到HDFS：

# 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 = linux01
a3.sources.r1.port = 4141

# Describe the sink
a3.sinks.k1.type = hdfs
a3.sinks.k1.hdfs.path = hdfs://linux01:8020/flume3/%Y%m%d/%H
#上传文件的前缀
a3.sinks.k1.hdfs.filePrefix = flume3-
#是否按照时间滚动文件夹
a3.sinks.k1.hdfs.round = true
#多少时间单位创建一个新的文件夹
a3.sinks.k1.hdfs.roundValue = 1
#重新定义时间单位
a3.sinks.k1.hdfs.roundUnit = hour
#是否使用本地时间戳
a3.sinks.k1.hdfs.useLocalTimeStamp = true
#积攒多少个Event才flush到HDFS一次
a3.sinks.k1.hdfs.batchSize = 100
#设置文件类型，可支持压缩
a3.sinks.k1.hdfs.fileType = DataStream
#多久生成一个新的文件
a3.sinks.k1.hdfs.rollInterval = 600
#设置每个文件的滚动大小大概是128M
a3.sinks.k1.hdfs.rollSize = 134217700
#文件的滚动与Event数量无关
a3.sinks.k1.hdfs.rollCount = 0
#最小冗余数
a3.sinks.k1.hdfs.minBlockReplicas = 1

# 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

4. 执行测试：分别开启对应flume-job（依次启动flume-3，flume-2，flume-1），同时产生文件变动并观察结果：

$ bin/flume-ng agent --conf conf/ --name a3 --conf-file job/group-job2/flume-3.conf
$ bin/flume-ng agent --conf conf/ --name a2 --conf-file job/group-job2/flume-2.conf
$ bin/flume-ng agent --conf conf/ --name a1 --conf-file job/group-job2/flume-1.conf

尖叫提示：测试时记得启动hive产生一些日志，同时使用telnet向44444端口发送内容，如：

$ bin/hive 
$ telnet linux01 44444

五、Flume监控之Ganglia

5.1 Ganglia的安装与部署

1. 安装httpd服务与php

# yum -y install httpd php

2. 安装其他依赖

# yum -y install rrdtool perl-rrdtool rrdtool-devel
# yum -y install apr-devel

3. 安装ganglia

# rpm -Uvh http://dl.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm
# yum -y install ganglia-gmetad 
# yum -y install ganglia-web
# yum install -y ganglia-gmond

4. 修改配置文件
   文件ganglia.conf：

# vi /etc/httpd/conf.d/ganglia.conf
修改为：
#
# Ganglia monitoring system php web frontend
#

Alias /ganglia /usr/share/ganglia

<Location /ganglia>
  Order deny,allow
  Deny from all
  Allow from all
  # Allow from 127.0.0.1
  # Allow from ::1
  # Allow from .example.com
</Location>

文件gmetad.conf：

# vi /etc/ganglia/gmetad.conf
修改为：
data_source "linux" 192.168.216.20

文件gmond.conf：

# vi /etc/ganglia/gmond.conf
修改为：
cluster {
  name = "linux"
  owner = "unspecified"
  latlong = "unspecified"
  url = "unspecified"
}
udp_send_channel {
  #bind_hostname = yes # Highly recommended, soon to be default.
                       # This option tells gmond to use a source address
                       # that resolves to the machine's hostname.  Without
                       # this, the metrics may appear to come from any
                       # interface and the DNS names associated with
                       # those IPs will be used to create the RRDs.
  # mcast_join = 239.2.11.71
  host = 192.168.216.20
  port = 8649
  ttl = 1
}
udp_recv_channel {
  # mcast_join = 239.2.11.71
  port = 8649
  bind = 192.168.216.20
  retry_bind = true
  # Size of the UDP buffer. If you are handling lots of metrics you really
  # should bump it up to e.g. 10MB or even higher.
  # buffer = 10485760
}

文件config：

# vi /etc/selinux/config
修改为：
# This file controls the state of SELinux on the system.
# SELINUX= can take one of these three values:
#     enforcing - SELinux security policy is enforced.
#     permissive - SELinux prints warnings instead of enforcing.
#     disabled - No SELinux policy is loaded.
SELINUX=disabled
# SELINUXTYPE= can take one of these two values:
#     targeted - Targeted processes are protected,
#     mls - Multi Level Security protection.
SELINUXTYPE=targeted

尖叫提示：selinux本次生效关闭必须重启，如果此时不想重启，可以临时生效之：

$ sudo setenforce 0

5. 启动ganglia

$ sudo service httpd start
$ sudo service gmetad start
$ sudo service gmond start

6. 打开网页浏览ganglia页面

http://192.168.216.20/ganglia 

尖叫提示：如果完成以上操作依然出现权限不足错误，请修改/var/lib/ganglia目录的权限：

$ sudo chmod -R 777 /var/lib/ganglia

5.2 操作Flume测试监控

1. 修改flume-env.sh配置：

JAVA_OPTS="-Dflume.monitoring.type=ganglia
-Dflume.monitoring.hosts=192.168.216.20:8649
-Xms100m
-Xmx200m"

2. 启动flume任务

$ bin/flume-ng agent \
--conf conf/ \
--name a1 \
--conf-file job/group-job0/flume-telnet.conf \
-Dflume.root.logger==INFO,console \
-Dflume.monitoring.type=ganglia \
-Dflume.monitoring.hosts=192.168.216.20:8649

3. 发送数据观察ganglia监测图

$ telnet localhost 44444

样式如图：