Flume是一个分布式、可靠、和高可用的海量日志采集、聚合和传输的系统。
它可以采集文件,socket数据包等各种形式源数据,又可以将采集到的数据输出到HDFS、hbase、hive、kafka等众多外部存储系统中。
一、flume结构
Flume分布式系统中最核心的角色是agent,每一个agent相当于一个数据传递员,内部有三个组件:
Source: 采集源,用于跟数据源对接,以获取数据;
Channel : angent内部的数据传输通道,用于从source将数据传递到sink。
Sink::下沉地,采集数据的传送目的,用于往下一级agent传递数据或者往最终存储系统传递数据;
数据在flume内部以Event的封装形式存在。
flume的事务控制机制:
1、source到channel
2、channel到sink
二、Flume多个agent串联
三、Flume安装使用(未安装)
1、上传安装包,解压
2、执行脚本,模拟日志生产
while true; do echo 111111111111111111111111_$RANDOM >> access.log; sleep 0.2; done
案例一、采集端口数据
1、增加netcat-logger.conf
# Name the components on this agent
#给那三个组件取个名字
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
#类型, 从网络端口接收数据,在本机启动, 所以localhost, type=spoolDir采集目录源,目录里有就采
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
#下沉的时候是一批一批的, 下沉的时候是一个个eventChannel参数解释:
#capacity:默认该通道中最大的可以存储的event数量
#trasactionCapacity:每次最大可以从source中拿到或者送到sink中的event数量
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、启动
$ bin/flume-ng agent --conf conf --conf-file conf/netcat-logger.conf --name a1 -Dflume.root.logger=INFO,console
3、传入数据:
$ telnet localhost 44444
案例二、采集文件夹数据
1、增加spooldir-hdfs.conf
# Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
#监听目录,spoolDir指定目录, fileHeader要不要给文件夹前坠名
a1.sources.r1.type = spooldir
a1.sources.r1.spoolDir = /home/hadoop/flumespool
a1.sources.r1.fileHeader = true
# 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
2、启动
bin/flume-ng agent -c ./conf -f ./conf/spool-logger.conf -n a1 -Dflume.root.logger=INFO,console
3、传入数据:
往/home/hadoop/flumeSpool放文件
案例三:采集文件数据(方式一)
exec source 适用于监控一个实时追加的文件,没有偏移量,会出现数据丢失情况;
1、增加tail-hdfs.conf
# Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# exec 指的是命令
a1.sources.r1.channels = c1
a1.sources.r1.type = exec
# F根据文件名追中, f根据文件的nodeid追踪,即使换了文件名,也能跟踪到
a1.sources.r1.command = tail -F /home/hadoop/log/test.log
#下沉目标
a1.sinks.k1.channel = c1
a1.sinks.k1.type = hdfs
# 指定目录, flum帮做目的替换
a1.sinks.k1.hdfs.path = /flume/events/%y-%m-%d/%H%M/
#文件的命名, 前缀
a1.sinks.k1.hdfs.filePrefix = events-
#10 分钟就改目录
a1.sinks.k1.hdfs.round = true
a1.sinks.k1.hdfs.roundValue = 10
a1.sinks.k1.hdfs.roundUnit = minute
#文件滚动之前的等待时间(秒)
a1.sinks.k1.hdfs.rollInterval = 120
#文件滚动的大小限制(bytes)
a1.sinks.k1.hdfs.rollSize = 268435456
#写入多少个event数据后滚动文件(事件个数)
a1.sinks.k1.hdfs.rollCount = 20
#1000个事件就往里面写入
a1.sinks.k1.hdfs.batchSize = 1000
#用本地时间格式化目录
a1.sinks.k1.hdfs.useLocalTimeStamp = true
#下沉后, 生成的文件类型,默认是Sequencefile,可用DataStream,则为普通文本
a1.sinks.k1.hdfs.fileType = DataStream
# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 1000
2、启动命令
bin/flume-ng agent -c conf -f conf/tail-hdfs.conf -n a1
案例四:采集文件数据(方式二)
taildir Source 既能够实现断点续传,又可以保证数据不丢失,还能够进行实时监控。为了防止00:00的时候,今日的数据写到明日,在sink处增加拦截器,给数据一个时间戳,不使用节点机器上时间。
tail dir 是根据通配符监视多个文件,即使文件改了名,也不会重复采集,它是根据偏移量进行跟踪的;
1、增加tail-hdfs.conf
# Name the components on this agent
a1.sources = r1
a1.sinks = k1
a1.channels = c1
a1.sources.r1.channels = c1
a1.sources.r1.type = TAILDIR
a1.sources.r1.filegroups=g1
a1.sources.r1.filegroups.g1= /logdata/a.*
a1.sources.r1.fileHeader = true
# 加入拦截器
a1.sources.s1.interceptors = i1
a1.sources.s1.interceptors.i1.type = timestamp
a1.sources.s1.interceptors.i1.headerName= timestamp
#下沉目标
a1.sinks.k1.channel = c1
a1.sinks.k1.type = hdfs
# 指定目录, flum帮做目的替换
a1.sinks.k1.hdfs.path = /flume/events/%y-%m-%d/%H%M/
#文件的命名, 前缀
a1.sinks.k1.hdfs.filePrefix = events-
#10 分钟就改目录
a1.sinks.k1.hdfs.round = true
a1.sinks.k1.hdfs.roundValue = 10
a1.sinks.k1.hdfs.roundUnit = minute
#文件滚动之前的等待时间(秒)
a1.sinks.k1.hdfs.rollInterval = 120
#文件滚动的大小限制(bytes)
a1.sinks.k1.hdfs.rollSize = 268435456
#写入多少个event数据后滚动文件(事件个数)
a1.sinks.k1.hdfs.rollCount = 20
#1000个事件就往里面写入
a1.sinks.k1.hdfs.batchSize = 1000
#用本地时间格式化目录
a1.sinks.k1.hdfs.useLocalTimeStamp = false
#下沉后, 生成的文件类型,默认是Sequencefile,可用DataStream,则为普通文本
a1.sinks.k1.hdfs.fileType = DataStream
# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 1000
2、启动命令
bin/flume-ng agent -c conf -f conf/tail-hdfs.conf -n a1
案例五、flume级联操作
使用前景:复杂的网络或者日志服务器特别多,每台服务器流量不多,需要进行汇集;
需要写两个配置文件,分别放在两个机器上,一个当发送者,一个当收集者(Kafka为例)
1、编写tail-avro-.conf
a1.sources = r1
a1.sinks = k1
a1.channels = c1
a1.sources.r1.channels = c1
a1.sources.r1.type = TAILDIR
a1.sources.r1.filegroups = g1
a1.sources.r1.filegroups.g1 = /logdata/a.*
a1.sources.r1.fileHeader = false
a1.channels.c1.type = file
a1.sinks.k1.channel = c1
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = doitedu02
a1.sinks.k1.port = 4444
2、编写avro-fakfa.conf
a1.sources = r1
a1.sinks = k1
a1.channels = c1
a1.sources.r1.channels = c1
a1.sources.r1.type = avro
a1.sources.r1.bind = doitedu02
a1.sources.r1.port = 4444
a1.sources.r1.batchSize = 100
a1.channels.c1.type = file
a1.sinks.k1.channel = c1
a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink
a1.sinks.k1.kafka.bootstrap.servers = doitedu01:9092,doitedu02:9092,doitedu03:9092
a1.sinks.k1.kafka.topic = doitedu17
a1.sinks.k1.kafka.producer.acks = 1
2、先启动avro-fakfa.conf,再启动tail-avro-.conf
bin/flume-ng agent -c conf -f conf/avro-fakfa.conf -n al -Dflume.root.logger=INFO,console
bin/flume-ng agent -c conf -f conf/tail-avro-.conf -n a1
3、kafka基本命令
## topic查看
bin/kafka-topics.sh --list --zookeeper doitedu01:2181
## topic创建
bin/kafka-topics.sh --create --topic topic2 --partitions 2 --replication-factor 2 --zookeeper doitedu01:2181
## 启动一个控制台生产者来生产数据
bin/kafka-console-producer.sh --broker-list doitedu01:9092,doitedu02:9092,doitedu03:9092 --topic topic2
>hello tom
## 启动一个控制台消费者来消费数据
bin/kafka-console-consumer.sh --bootstrap-server doitedu01:9092,doitedu02:9092,doitedu03:9092 --topic topic2 --from-beginning
案例六:flume选择器
一个 source 可以对接多个 channel,那么,source 的数据如何在多个 channel 之间传递,就由 selector 来控制,配置应该挂载到 source 组件
1、复制选择器
一个连hdfs, 一个连kafka
a1.sources = r1
a1.channels = c1 c2
a1.sinks = k1 k2
a1.sources.r1.channels = c1 c2
a1.sources.r1.type = TAILDIR
a1.sources.r1.filegroups = g1
a1.sources.r1.filegroups.g1 = /logdata/a.*
a1.sources.r1.fileHeader = false
a1.sources.r1.selector.type = replicating
a1.sources.r1.interceptors = i1
a1.sources.r1.interceptors.i1.type = timestamp
a1.sources.r1.interceptors.i1.headerName = timestamp
a1.channels.c1.type = memory
a1.channels.c2.type = memory
a1.sinks.k1.channel = c1
a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink
a1.sinks.k1.kafka.bootstrap.servers = doitedu01:9092,doitedu02:9092,doitedu03:9092
a1.sinks.k1.kafka.topic = doitedu17
a1.sinks.k1.kafka.producer.acks = 1
a1.sinks.k2.channel = c2
a1.sinks.k2.type = hdfs
a1.sinks.k2.hdfs.path = hdfs://doitedu01:8020/flumedata/%Y-%m-%d/%H
a1.sinks.k2.hdfs.filePrefix = doitedu-log-
a1.sinks.k2.hdfs.fileSuffix = .log
a1.sinks.k2.hdfs.rollSize = 268435456
a1.sinks.k2.hdfs.rollInterval = 120
a1.sinks.k2.hdfs.rollCount = 0
a1.sinks.k2.hdfs.batchSize = 1000
a1.sinks.k2.hdfs.fileType = CompressedStream
a1.sinks.k2.hdfs.codeC = snappy
a1.sinks.k2.hdfs.useLocalTimeStamp = false
2、多路选择器
一个source里数据,可能有不同种类数据,需要使用拦截器,对数据进行区分,然后使用多路选择器插入到不同的channel里,一个写到kakfa,一个写到hdfs。
2.1 拦截器,并打包放到flume的lib下
package cn.doitedu.yiee.flume;
import org.apache.flume.Context;
import org.apache.flume.Event;
import org.apache.flume.interceptor.Interceptor;
import java.util.List;
public class MultiplexingInterceptor implements Interceptor {
private Integer flagfield = 0;
private Integer timestampfield = 0;
public MultiplexingInterceptor(Integer flagfield,Integer timestampfield) {
this.flagfield = flagfield;
this.timestampfield = timestampfield;
}
/**
* 拦截器构造实例后的初始化工作
*/
public void initialize() {
}
// 日志格式:
// u01,ev1,mall,1568738583468
public Event intercept(Event event) {
// 根据event的数据内容,以及参数中指定的标记字段,来产生不同的header值
byte[] body = event.getBody();
String line = new String(body);
String[] split = line.split(",");
// 切出业务标记,并添加到header
event.getHeaders().put("flag",split[flagfield]);
// 切出行为(事件)时间戳,并添加到header
event.getHeaders().put("timestamp",split[timestampfield]);
return event;
}
public List<Event> intercept(List<Event> list) {
for (Event event : list) {
intercept(event);
}
return list;
}
/**
* 拦截器销毁之前的一些清理工作
*/
public void close() {
}
public static class MultiplexingInterceptorBuilder implements Interceptor.Builder{
Integer flagfield = 0;
Integer timestampfield = 0;
/**
* 用户构建一个拦截器实例
* @return
*/
public Interceptor build() {
return new MultiplexingInterceptor(flagfield,timestampfield);
}
/**
* 获取参数的入口
* @param context
*/
public void configure(Context context) {
flagfield = context.getInteger("flagfield");
timestampfield = context.getInteger("timestampfield");
}
}
}
2.2 模拟日志生成脚本
while true
do
if [ $(($RANDOM % 2)) -eq 0 ]
then
echo "u$RANDOM,e1,waimai,`date +%s`000" >> a.log
else
echo "u$RANDOM,e1,mall,`date +%s`000" >> a.log
fi
sleep 0.2
done
2.3变成配置文件
1.sources = r1
a1.channels = c1 c2
a1.sinks = k1 k2
a1.sources.r1.channels = c1 c2
a1.sources.r1.type = TAILDIR
a1.sources.r1.filegroups = g1
a1.sources.r1.filegroups.g1 = /logdata/a.*
a1.sources.r1.fileHeader = false
a1.sources.r1.interceptors = i1
a1.sources.r1.interceptors.i1.type = cn.doitedu.yiee.flume.MultiplexingInterceptor$MultiplexingInterceptorBuilder
a1.sources.r1.interceptors.i1.flagfield = 2
a1.sources.r1.interceptors.i1.timestampfield = 3
a1.sources.r1.selector.type = multiplexing
a1.sources.r1.selector.header = flag
a1.sources.r1.selector.mapping.mall = c1
a1.sources.r1.selector.mapping.waimai = c2
a1.sources.r1.selector.default = c2
a1.channels.c1.type = memory
a1.channels.c1.capacity = 2000
a1.channels.c1.transactionCapacity = 1000
a1.channels.c2.type = memory
a1.channels.c2.capacity = 2000
a1.channels.c2.transactionCapacity = 1000
a1.sinks.k1.channel = c1
a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink
a1.sinks.k1.kafka.bootstrap.servers = doitedu01:9092,doitedu02:9092,doitedu03:9092
a1.sinks.k1.kafka.topic = mall
a1.sinks.k1.kafka.producer.acks = 1
a1.sinks.k2.channel = c2
a1.sinks.k2.type = hdfs
a1.sinks.k2.hdfs.path = hdfs://doitedu01:8020/waimai/%Y-%m-%d/%H
a1.sinks.k2.hdfs.filePrefix = doitedu-log-
a1.sinks.k2.hdfs.fileSuffix = .log
a1.sinks.k2.hdfs.rollSize = 268435456
a1.sinks.k2.hdfs.rollInterval = 120
a1.sinks.k2.hdfs.rollCount = 0
a1.sinks.k2.hdfs.batchSize = 1000
a1.sinks.k2.hdfs.fileType = DataStream
a1.sinks.k2.hdfs.useLocalTimeStamp = false
案例七:自动失败切换
多个sink连接一个channel,默认不需要专门去配置的, 相当于负载均衡,或者failover sink processor 自动失败,需要将多个 sink 创建成 group。正常情况下,只运行一个sink,只有当它失败后,才切换到别的sink上。
默认是走兰色的线,若是兰色的机器挂掉,就走绿色的线;
1、级联高可用配置第一级
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1
a1.sources.r1.channels = c1
a1.sources.r1.type = TAILDIR
a1.sources.r1.filegroups = g1
a1.sources.r1.filegroups.g1 = /logdata/a.*
a1.sources.r1.fileHeader = false
a1.channels.c1.type = memory
a1.channels.c1.capacity = 2000
a1.channels.c1.transactionCapacity = 1000
a1.sinks.k1.channel = c1
a1.sinks.k1.type = avro
a1.sinks.k1.hostname = doitedu02
a1.sinks.k1.port = 4444
a1.sinks.k2.channel = c1
a1.sinks.k2.type = avro
a1.sinks.k2.hostname = doitedu03
a1.sinks.k2.port = 4444
a1.sinkgroups = g1
a1.sinkgroups.g1.sinks = k1 k2
a1.sinkgroups.g1.processor.type = failover
a1.sinkgroups.g1.processor.priority.k1 = 200
a1.sinkgroups.g1.processor.priority.k2 = 100
a1.sinkgroups.g1.processor.maxpenalty = 5000
2、级联高可用配置第2级(节点1)
a1.sources = r1
a1.sinks = k1
a1.channels = c1
a1.sources.r1.channels = c1
a1.sources.r1.type = avro
a1.sources.r1.bind = doitedu02
a1.sources.r1.port = 4444
a1.sources.r1.batchSize = 100
a1.channels.c1.type = memory
a1.channels.c1.capacity = 2000
a1.channels.c1.transactionCapacity = 1000
a1.sinks.k1.channel = c1
a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink
a1.sinks.k1.kafka.bootstrap.servers = doitedu01:9092,doitedu02:9092,doitedu03:9092
a1.sinks.k1.kafka.topic = failover
a1.sinks.k1.kafka.producer.acks = 1
3、级联高可用配置第2级(节点2)
a1.sources = r1
a1.sinks = k1
a1.channels = c1
a1.sources.r1.channels = c1
a1.sources.r1.type = avro
a1.sources.r1.bind = doitedu03
a1.sources.r1.port = 4444
a1.sources.r1.batchSize = 100
a1.channels.c1.type = memory
a1.channels.c1.capacity = 2000
a1.channels.c1.transactionCapacity = 1000
a1.sinks.k1.channel = c1
a1.sinks.k1.type = org.apache.flume.sink.kafka.KafkaSink
a1.sinks.k1.kafka.bootstrap.servers = doitedu01:9092,doitedu02:9092,doitedu03:9092
a1.sinks.k1.kafka.topic = failover
a1.sinks.k1.kafka.producer.acks = 1
四、flume监控
flume 在运行时,状态是否正常,吞吐量是否正常,可以使用ganglia 进行展现:
-Dflume.monitoring.type=ganglia -Dflume.monitoring.port=34890
Ganglia 是一个通用的集群运维监控系统;
它在各台需要监控状态信息的机器上安装“探针”,然后这些“探针”会收集所在机器上的各种状态
信息(cpu 负载,内存负载,磁盘 IO 负载,网络 IO 负载,以及各类应用软件的状态信息),然后汇
聚到它的中心汇聚点,并提供 web 页面进行图形可视化查看
五、监控flume进程、自动拉起
#!/bin/bash
export FLUME_HOME=/opt/apps/flume-1.9.0
while true
do
pc=`ps -ef | grep flume | grep -v "grep" | wc -l`
if [[ $pc -lt 1 ]]
then
echo "detected no flume process.... preparing to launch flume agent...... "
${FLUME_HOME}/bin/flume-ng agent -n a1 -c ${FLUME_HOME}/conf/ -f ${FLUME_HOME}/agentconf/failover.properties 1>/dev/null 2>&1 &
else
echo "detected flume process number is : $pc "
fi
sleep 1
done