【log收集】:
机器名称        服务名称    用户
flume-agent-01: namenode    hdfs
flume-agent-02: datanode    hdfs
flume-agent-03: datanode    hdfs

【log聚合】:
机器名称                                用户
sht-sgmhadoopcm-01(172.16.101.54)       root

【sink到hdfs】:
 hdfs://172.16.101.56:8020/testwjp/

1.下载apache-flume-1.7.0-bin.tar.gz
[hdfs@flume-agent-01 tmp]$ wget http://www-eu.apache.org/dist/flume/1.7.0/apache-flume-1.7.0-bin.tar.gz
--2017-01-04 20:40:10--  http://www-eu.apache.org/dist/flume/1.7.0/apache-flume-1.7.0-bin.tar.gz
Resolving www-eu.apache.org... 88.198.26.2, 2a01:4f8:130:2192::2
Connecting to www-eu.apache.org|88.198.26.2|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 55711670 (53M) [application/x-gzip]
Saving to: “apache-flume-1.7.0-bin.tar.gz”

100%[===============================================================================================================================================================================================>] 55,711,670   473K/s   in 74s    

2017-01-04 20:41:25 (733 KB/s) - “apache-flume-1.7.0-bin.tar.gz” saved [55711670/55711670]

2.解压重命名
[hdfs@flume-agent-01 tmp]$
[hdfs@flume-agent-01 tmp]$ tar -xzvf apache-flume-1.7.0-bin.tar.gz
[hdfs@flume-agent-01 tmp]$ mv apache-flume-1.7.0-bin flume-ng
[hdfs@flume-agent-01 tmp]$ cd flume-ng/conf

3.拷贝flume环境配置和agent配置文件
[hdfs@flume-agent-01 tmp]$ cp flume-env.sh.template flume-env.sh
[hdfs@flume-agent-01 tmp]$ cp flume-conf.properties.template exec_memory_avro.properties

4.添加hdfs用户的环境变量文件
[hdfs@flume-agent-01 tmp]$ cd
[hdfs@flume-agent-01 ~]$ ls -la
total 24
drwxr-xr-x   3 hdfs hadoop 4096 Jul  8 14:05 .
drwxr-xr-x. 35 root root   4096 Dec 10  2015 ..
-rw-------   1 hdfs hdfs   4471 Jul  8 17:22 .bash_history
drwxrwxrwt   2 hdfs hadoop 4096 Nov 19  2014 cache
-rw-------   1 hdfs hdfs   3131 Jul  8 14:05 .viminfo
[hdfs@flume-agent-01 ~]$ cp /etc/skel/.* ./
cp: omitting directory `/etc/skel/.'
cp: omitting directory `/etc/skel/..'
[hdfs@flume-agent-01 ~]$ ls -la
total 36
drwxr-xr-x   3 hdfs hadoop 4096 Jan  4 20:49 .
drwxr-xr-x. 35 root root   4096 Dec 10  2015 ..
-rw-------   1 hdfs hdfs   4471 Jul  8 17:22 .bash_history
-rw-r--r--   1 hdfs hdfs     18 Jan  4 20:49 .bash_logout
-rw-r--r--   1 hdfs hdfs    176 Jan  4 20:49 .bash_profile
-rw-r--r--   1 hdfs hdfs    124 Jan  4 20:49 .bashrc
drwxrwxrwt   2 hdfs hadoop 4096 Nov 19  2014 cache
-rw-------   1 hdfs hdfs   3131 Jul  8 14:05 .viminfo

5.添加flume的环境变量
[hdfs@flume-agent-01 ~]$ vi .bash_profile

export FLUME_HOME=/tmp/flume-ng
export FLUME_CONF_DIR=$FLUME_HOME/conf
export PATH=$PATH:$FLUME_HOME/bin
[hdfs@flume-agent-01 ~]$ . .bash_profile

6.修改flume环境配置文件
[hdfs@flume-agent-01 conf]$ vi flume-env.sh
export JAVA_HOME=/usr/java/jdk1.7.0_25

7.将基于Flume-ng Exec Source开发自定义插件AdvancedExecSource的AdvancedExecSource.jar包上传到$FLUME_HOME/lib/
http://blog.itpub.net/30089851/viewspace-2131995/

[hdfs@LogshedNameNodeLogcollector lib]$ pwd
/tmp/flume-ng/lib
[hdfs@LogshedNameNodeLogcollector lib]$ ll AdvancedExecSource.jar
-rw-r--r-- 1 hdfs hdfs 10618 Jan  5 23:50 AdvancedExecSource.jar
[hdfs@LogshedNameNodeLogcollector lib]$

8.修改flume的agent配置文件
[hdfs@flume-agent-01 conf]$ vi exec_memory_avro.properties
# Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1

# Describe/configure the custom exec source
a1.sources.r1.type = com.onlinelog.analysis.AdvancedExecSource
a1.sources.r1.command = tail -f /var/log/hadoop-hdfs/hadoop-cmf-hdfs1-NAMENODE-flume-agent-01.log.out
a1.sources.r1.hostname = flume-agent-01
a1.sources.r1.servicename = namenode

# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = 172.16.101.54
a1.sinks.k1.port = 4545

# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.keep-alive = 60
a1.channels.c1.capacity = 1000000
a1.channels.c1.transactionCapacity = 2000

# Bind the source and sink to the channel
a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1

9.将flume-agent-01的flume-ng打包,scp到flume-agent-02/03 和 sht-sgmhadoopcm-01(172.16.101.54)
[hdfs@flume-agent-01 tmp]$ zip -r flume-ng.zip flume-ng/*

[jpwu@flume-agent-01 ~]$ scp /tmp/flume-ng.zip flume-agent-02:/tmp/
[jpwu@flume-agent-01 ~]$ scp /tmp/flume-ng.zip flume-agent-03:/tmp/
[jpwu@flume-agent-01 ~]$ scp /tmp/flume-ng.zip sht-sgmhadoopcm-01:/tmp/

10.在flume-agent-02配置hdfs用户环境变量和解压，修改agent配置文件
[hdfs@flume-agent-02 ~]$ cp /etc/skel/.* ./
cp: omitting directory `/etc/skel/.'
cp: omitting directory `/etc/skel/..'
[hdfs@flume-agent-02 ~]$ vi .bash_profile
export FLUME_HOME=/tmp/flume-ng
export FLUME_CONF_DIR=$FLUME_HOME/conf
export PATH=$PATH:$FLUME_HOME/bin
[hdfs@flume-agent-02 ~]$ . .bash_profile

[hdfs@flume-agent-02 tmp]$ unzip flume-ng.zip
[hdfs@flume-agent-02 tmp]$ cd flume-ng/conf

##修改以下参数即可
[hdfs@flume-agent-02 conf]$ vi exec_memory_avro.properties
a1.sources.r1.command = tail -f /var/log/hadoop-hdfs/hadoop-cmf-hdfs1-DATANODE-flume-agent-02.log.out
a1.sources.r1.hostname = flume-agent-02
a1.sources.r1.servicename = datanode

###要检查flume-env.sh的JAVA_HOME目录是否存在

11.在flume-agent-03配置hdfs用户环境变量和解压，修改agent配置文件
[hdfs@flume-agent-03 ~]$ cp /etc/skel/.* ./
cp: omitting directory `/etc/skel/.'
cp: omitting directory `/etc/skel/..'
[hdfs@flume-agent-03 ~]$ vi .bash_profile
export FLUME_HOME=/tmp/flume-ng
export FLUME_CONF_DIR=$FLUME_HOME/conf
export PATH=$PATH:$FLUME_HOME/bin
[hdfs@flume-agent-03 ~]$ . .bash_profile

[hdfs@flume-agent-03 tmp]$ unzip flume-ng.zip
[hdfs@flume-agent-03 tmp]$ cd flume-ng/conf

##修改以下参数即可
[hdfs@flume-agent-03 conf]$ vi exec_memory_avro.properties
a1.sources.r1.command = tail -f /var/log/hadoop-hdfs/hadoop-cmf-hdfs1-DATANODE-flume-agent-03.log.out
a1.sources.r1.hostname = flume-agent-03
a1.sources.r1.servicename = datanode

###要检查flume-env.sh的JAVA_HOME目录是否存在

12.聚合端 sht-sgmhadoopcm-01，配置root用户环境变量和解压，修改agent配置文件
[root@sht-sgmhadoopcm-01 tmp]# vi /etc/profile
export JAVA_HOME=/usr/java/jdk1.7.0_67-cloudera
export FLUME_HOME=/tmp/flume-ng
export FLUME_CONF_DIR=$FLUME_HOME/conf

export PATH=$FLUME_HOME/bin:$JAVA_HOME/bin:$PATH
[root@sht-sgmhadoopcm-01 tmp]# source /etc/profile
[root@sht-sgmhadoopcm-01 tmp]#

[root@sht-sgmhadoopcm-01 tmp]# unzip flume-ng.zip
[root@sht-sgmhadoopcm-01 tmp]# cd flume-ng/conf

[root@sht-sgmhadoopcm-01 conf]# vi flume-env.sh
export JAVA_HOME=/usr/java/jdk1.7.0_67-cloudera

###测试: 先聚合, sink到hdfs端
[root@sht-sgmhadoopcm-01 conf]# vi avro_memory_hdfs.properties
# Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1

# Describe/configure the source
a1.sources.r1.type = avro
a1.sources.r1.bind = 172.16.101.54
a1.sources.r1.port = 4545

# Describe the sink
a1.sinks.k1.type = hdfs
a1.sinks.k1.hdfs.path = hdfs://172.16.101.56:8020/testwjp/
a1.sinks.k1.hdfs.filePrefix = logs
a1.sinks.k1.hdfs.inUsePrefix = .

a1.sinks.k1.hdfs.rollInterval = 0
### roll 16 m = 16777216 bytes
a1.sinks.k1.hdfs.rollSize = 1048576
a1.sinks.k1.hdfs.rollCount = 0
a1.sinks.k1.hdfs.batchSize = 6000

a1.sinks.k1.hdfs.writeFormat = text
a1.sinks.k1.hdfs.fileType = DataStream

# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.keep-alive = 90
a1.channels.c1.capacity = 1000000
a1.channels.c1.transactionCapacity = 6000

# Bind the source and sink to the channel
a1.sources.r1.channels = c1
a1.sinks.k1.channel = c1

13.后台启动
[root@sht-sgmhadoopcm-01 flume-ng]# source /etc/profile
[hdfs@flume-agent-01 flume-ng]$ . ~/.bash_profile
[hdfs@flume-agent-02 flume-ng]$ . ~/.bash_profile
[hdfs@flume-agent-03 flume-ng]$ . ~/.bash_profile

[root@sht-sgmhadoopnn-01 flume-ng]# nohup  flume-ng agent -c conf -f /tmp/flume-ng/conf/avro_memory_hdfs.properties -n a1 -Dflume.root.logger=INFO,console &

[hdfs@flume-agent-01 flume-ng]$ nohup  flume-ng agent -c /tmp/flume-ng/conf -f /tmp/flume-ng/conf/exec_memory_avro.properties -n a1 -Dflume.root.logger=INFO,console &
[hdfs@flume-agent-01 flume-ng]$ nohup  flume-ng agent -c /tmp/flume-ng/conf -f /tmp/flume-ng/conf/exec_memory_avro.properties -n a1 -Dflume.root.logger=INFO,console &
[hdfs@flume-agent-01 flume-ng]$ nohup  flume-ng agent -c /tmp/flume-ng/conf -f /tmp/flume-ng/conf/exec_memory_avro.properties -n a1 -Dflume.root.logger=INFO,console &

14.校验：将集群的日志下载到本地，打开查看即可(略)

------------------------------------------------------------------------------------------------------------------------------------------------

【备注】:
1.错误1 flume-ng安装的机器上没有hadoop环境，所以假如sink到hdfs话，需要用到hdfs的jar包
[ERROR - org.apache.flume.node.PollingPropertiesFileConfigurationProvider$FileWatcherRunnable.run(PollingPropertiesFileConfigurationProvider.java:146)] Failed to start agent
because dependencies were not found in classpath. Error follows.
java.lang.NoClassDefFoundError: org/apache/hadoop/io/SequenceFile$CompressionType

只需在其他安装hadoop机器上搜索以下5个jar包，拷贝到$FLUME_HOME/lib目录即可。
搜索方法: find $HADOOP_HOME/ -name commons-configuration*.jar

commons-configuration-1.6.jar
hadoop-auth-2.7.3.jar
hadoop-common-2.7.3.jar
hadoop-hdfs-2.7.3.jar
hadoop-mapreduce-client-core-2.7.3.jar
protobuf-java-2.5.0.jar
htrace-core-3.1.0-incubating.jar
commons-io-2.4.jar

2.错误2 无法加载自定义插件的类 Unable to load source type: com.onlinelog.analysis.AdvancedExecSource
2017-01-06 21:10:48,278 (conf-file-poller-0) [ERROR - org.apache.flume.node.PollingPropertiesFileConfigurationProvider$FileWatcherRunnable.run(PollingPropertiesFileConfigurationProvider.java:142)] Failed to load configuration data. Exception follows.
org.apache.flume.FlumeException: Unable to load source type: com.onlinelog.analysis.AdvancedExecSource, class: com.onlinelog.analysis.AdvancedExecSource

执行hdfs或者root用户的环境变量即可
[root@sht-sgmhadoopcm-01 flume-ng]# source /etc/profile
[hdfs@flume-agent-01 flume-ng]$ . ~/.bash_profile
[hdfs@flume-agent-02 flume-ng]$ . ~/.bash_profile
[hdfs@flume-agent-03 flume-ng]$ . ~/.bash_profile