Flume-NG 使用首记
第一步,配置Cloudera-CDH4软件源,见Cloudera官网文档。略……
第二步,安装Flume客户端
yum -y install flume-ng
第三步,配置Flume
1、Flume使用经典的alternatives进行多版本管理,查看/添加配置路径:
# alternatives --display flume-ng-conf
flume-ng-conf - status is auto.
link currently points to /etc/flume-ng/conf.empty
/opt/cloudera/parcels/CDH-4.2.0-1.cdh4.2.0.p0.10/etc/flume-ng/conf.empty - priority 10
/etc/flume-ng/conf.empty - priority 30
Current `best' version is /etc/flume-ng/conf.empty.
2、更新客户端配置文件 /etc/flume-ng/conf/flume.conf :
# 定义Agent1对应的属性名称
agent1.sources = ngrinder
agent1.sinks = hdfs4log
agent1.channels = memory4log
# 定义日志的来源,下例为Ngrinder的启动日志
agent1.sources.ngrinder.type = exec
agent1.sources.ngrinder.command = tail -F /data/log/ngrinder/startup.log
agent1.sources.ngrinder.channels = momery01
# 指定日志存储的目标地址,下例为直接写到HDFS中
agent1.sinks.hdfs4log.type = hdfs
agent1.sinks.hdfs4log.hdfs.path = hdfs://hdfs.kisops.org:8020/flume/ngrinder
agent1.sinks.hdfs4log.channel = momery4log
# 定义日志缓冲区,方便在网络阻塞时日志可以延时输出
agent1.channels.memory4log.type = memory
agent1.channels.memory4log.capacity = 1000
agent1.channels.memory4log.transactionCapacity = 100
3、启动Flume客户端:
su -m -c 'flume-ng agent -n agent1 -f /etc/flume-ng/conf/flume.conf' hdfs
4、验证:重启Ngrinder;查看HDFS对应目录是否有数据生成。
5、测试环境
# flume-ng version
Flume 1.3.0-cdh4.2.1
Source code repository: https://git-wip-us.apache.org/repos/asf/flume.git
Revision: 68867f41ba0906cc4050efade51afca563f0e65b
Compiled by jenkins on Mon Apr 22 13:07:14 PDT 2013
From source with checksum a341b20667cda79e8561f388a94a5168# uname -a
Linux CentOS6.4 2.6.32-358.2.2.el6.x86_64