集群规划

Flume集群，负载均衡和故障转移模式，笔者准备了3台机器安装了flume，其中webapp200是应用服务器，flume安装在这里，目的是收集应用服务器上的日志，通过2个avro sink分别对接到flume130和flume131机器；再通过flume130和flume131将数据传输到HDFS。(注：吞吐量大的channels可以换成kafka)。

Webapp200：TAILDIR -> file -> avro

Flume130：avro -> file -> hdfs

Flume131：avro -> file -> hdfs

流程图

下载安装

下载地址

官网：http://flume.apache.org/

解压

解压到/opt/module/目录

$ tar -zxvf apache-flume-1.9.0-bin.tar.gz -C /opt/module/

配置环境变量

配置JAVA_HOME

修改配置文件名称

$ mv flume-env.sh.template flume-env.sh

修改Flume-env.sh

$ vi conf/flume-env.sh

修改JAVA_HOME，修改成自己的JAVA_HOME

export JAVA_HOME=/opt/module/jdk1.8.0_221

配置完后，将flume分发到其他机器

配置Agent

webapp200中的Agent

创建taildir-file-hdfs.conf，并添加如下内容：

# Name the components on this agenta1.sources = r1a1.sinks = k1 k2a1.channels = c1 # Describe/configure the sourcea1.sources.r1.type = TAILDIRa1.sources.r1.channels = c1a1.sources.r1.positionFile = /opt/module/apache-flume-1.9.0-bin/position/taildir_position.jsona1.sources.r1.filegroups = f1a1.sources.r1.filegroups.f1 = /opt/logs/info*.log* # Describe the sinkgroupsa1.sinkgroups = g1a1.sinkgroups.g1.sinks = k1 k2 k3a1.sinkgroups.g1.processor.type = load_balancea1.sinkgroups.g1.processor.backoff = truea1.sinkgroups.g1.processor.selector = round_robina1.sinkgroups.g1.processor.selector.maxTimeOut=10000 #Define the sink k1a1.sinks.k1.type = avroa1.sinks.k1.channel = c1a1.sinks.k1.hostname = flume130a1.sinks.k1.port = 4545 #Define the sink k2a1.sinks.k2.type = avroa1.sinks.k2.channel = c1a1.sinks.k2.hostname = flume131a1.sinks.k2.port = 4545 # Use a channel which buffers events in memorya1.channels.c1.type = filea1.channels.c1.checkpointDir=/opt/module/apache-flume-1.9.0-bin/data/checkpoint/balancea1.channels.c1.dataDirs=/opt/module/apache-flume-1.9.0-bin/data/balancea1.channels.c1.capacity = 1000a1.channels.c1.transactionCapacity = 100 # Bind the source and sink to the channela1.sources.r1.channels = c1a1.sinks.k1.channel = c1a1.sinks.k2.channel = c1

flume130和flume131中的Agent

# Name the components on this agenta1.sources = r1a1.sinks = k1a1.channels = c1 # Describe/configure the sourcea1.sources.r1.type = avroa1.sources.r1.channels = c1a1.sources.r1.bind = 0.0.0.0a1.sources.r1.port = 4545 # Describe the sinka1.sinks.k1.type = hdfsa1.sinks.k1.channel = c1a1.sinks.k1.hdfs.path = hdfs://hadoop100:9000/flume/events/%y-%m-%d/%Ha1.sinks.k1.hdfs.useLocalTimeStamp = truea1.sinks.k1.hdfs.filePrefix = events-a1.sinks.k1.hdfs.batchSize=100a1.sinks.k1.hdfs.fileType=DataStreama1.sinks.k1.hdfs.rollInterval=0a1.sinks.k1.hdfs.rollSize=134217700a1.sinks.k1.hdfs.rollCount=0a1.sinks.k1.hdfs.round = truea1.sinks.k1.hdfs.roundValue = 1a1.sinks.k1.hdfs.roundUnit = hour # Use a channel which buffers events in memorya1.channels.c1.type = filea1.channels.c1.checkpointDir=/opt/module/apache-flume-1.9.0-bin/data/checkpoint/balancea1.channels.c1.dataDirs=/opt/module/apache-flume-1.9.0-bin/data/balancea1.channels.c1.capacity = 1000a1.channels.c1.transactionCapacity = 100 # Bind the source and sink to the channela1.sources.r1.channels = c1a1.sinks.k1.channel = c1

注意HDFS sink 需要放置相应的jar包和xml配置文件到flume目录下，下载和使用链接(笔者亲测有效，注意版本)：flume hdfs sink所需jar包(flume1.9.0 hadoop3.1.2)。

启动flume

启动命令

$ bin/flume-ng agent -n a1 -c conf -f job/taildir-file-avro.conf

后台启动，在结尾加上&

$ nohup bin/flume-ng agent -n a1 -c conf -f job/taildir-file-avro.conf &

再加上nohup可以把原本在console输出的运行日志输出在[当前运行目录]/nohup.out中

关闭flume

flume进程启动动没有关闭的命令，只能kill掉。

查看占用4545端口的进程ID

$ netstat -nap | grep 4545

或者直接jps找到flume进程，然后kill

$ kill [pid]