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。支持在日志系统中定制各类数据发送方,用于收集数
据;同时,Flume提供对数据进行简单处理,并写到各种数据接受方(比如文本、HDFS、Hbase等)的能力 。
什么是Flume?
flume 作为 cloudera 开发的实时日志收集系统,受到了业界的认可与广泛应用。Flume 初始的发行版本目前被统称
为 Flume OG(original generation),属于 cloudera。但随着 FLume 功能的扩展,Flume OG 代码工程臃肿、核心组件
设计不合理、核心配置不标准等缺点暴露出来,尤其是在 Flume OG 的最后一个发行版本 0.94.0 中,日志传输不稳定的现象
尤为严重,为了解决这些问题,2011 年 10 月 22 号,cloudera 完成了 Flume-728,对 Flume 进行了里程碑式的
改动:重构核心组件、核心配置以及代码架构,重构后的版本统称为 Flume NG(next generation);改动的另一原因是
将 Flume 纳入 apache 旗下,cloudera Flume 改名为 Apache Flume。
flume的特点:
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。支持在日志系统中定制各类数据发送方,用于
收集数据;同时,Flume提供对数据进行简单处理,并写到各种数据接受方(比如文本、HDFS、Hbase等)的能力 。
flume的数据流由事件(Event)贯穿始终。事件是Flume的基本数据单位,它携带日志数据(字节数组形式)并且携带有头信息,
这些Event由Agent外部的Source生成,当Source捕获事件后会进行特定的格式化,然后Source会把事件推入(单个或多个)
Channel中。你可以把Channel看作是一个缓冲区,它将保存事件直到Sink处理完该事件。Sink负责持久化日志或者把事件推向
另一个Source。
flume的可靠性
当节点出现故障时,日志能够被传送到其他节点上而不会丢失。Flume提供了三种级别的可靠性保障,从强到弱依次分别
为:end-to-end(收到数据agent首先将event写到磁盘上,当数据传送成功后,再删除;如果数据发送失败,可以重新发送。),
Store on failure(这也是scribe采用的策略,当数据接收方crash时,将数据写到本地,待恢复后,继续发送),
Besteffort(数据发送到接收方后,不会进行确认)。
flume的可恢复性
还是靠Channel。推荐使用FileChannel,事件持久化在本地文件系统里(性能较差)。
flume的一些核心概念
Agent使用JVM 运行Flume。每台机器运行一个agent,但是可以在一个agent中包含多个sources和sinks。
Client生产数据,运行在一个独立的线程。
Source从Client收集数据,传递给Channel。
Sink从Channel收集数据,运行在一个独立线程。
Channel连接 sources 和 sinks ,这个有点像一个队列。
Events可以是日志记录、 avro 对象等。
Flume以agent为最小的独立运行单位。一个agent就是一个JVM。单agent由Source、Sink和Channel三大组件构成,如下图:
值得注意的是,Flume提供了大量内置的Source、Channel和Sink类型。不同类型的Source,Channel和Sink可以自由组合。
组合方式基于用户设置的配置文件,非常灵活。比如:Channel可以把事件暂存在内存里,也可以持久化到本地硬盘上。
Sink可以把日志写入HDFS, HBase,甚至是另外一个Source等等。Flume支持用户建立多级流,也就是说,多个agent可以协同
工作,并且支持Fan-in、Fan-out、Contextual Routing、Backup Routes,这也正是NB之处。如下图所示:
flume的官方网站在哪里?
官方网站:http://flume.apache.org/
如何安装
wget http://mirrors.hust.edu.cn/apache/flume/1.6.0/apache-flume-1.6.0-bin.tar.gz
tar -zxvf apache-flume-1.6.0-bin.tar.gz
mkdir /usr/local/flume
mv apache-flume-1.6.0 /usr/local/flume
cd /usr/local/flume
cp conf/flume-conf.properties.template conf/flume-conf.properties #配置文件
mkdir -p /tmp/log/flume 建立输出目录
编辑配置 conf/flume-conf.properties
vim conf/flume-conf.properties
agent.sources = r1
agent.channels = c1
agent.sinks = s1
agent.sources.r1.type = netcat
agent.sources.r1.bind = localhost
agent.sources.r1.port = 8888
agent.sources.r1.channels = c1
agent.sinks.s1.type = file_roll
agent.sinks.s1.sink.directory = /tmp/log/flume
agent.sinks.s1.channel = c1
agent.channels.c1.type = memory
agent.channels.c1.capacity = 100
启动 flume服务
bin/flume-ng agent --conf conf -f conf/flume-conf.properties -n agent& #启动服务
注:运行日志位于logs目录,或者启动时添加-Dflume.root.logger=INFO,console 选项前台启动,输出打印日志,
查看具体运行日志,服务异常时查原因。
安装验证
bin/flume-ng version #查看flume版本
telnet localhost 8888 #发送数据
输入数据
hello world!
hello Flume!
查看 /tmp/log/flume文件中文件内容
flume配置文件说明
Avro模式
Avro可以发送一个给定的文件给Flume,Avro 源使用AVRO RPC机制。
创建agent配置文件
vim /usr/local/flume/cong/avro.conf
-------------------------------------------avro.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = avro
a1.sources.r1.channels = c1
a1.sources.r1.bind = 0.0.0.0
a1.sources.r1.port = 4141
# 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
------------------------------------------- end ------------------------------------------------
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/avro.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
echo "hello world" > /usr/local/flume/log/log.00 #创建指定文件
/usr/local/flume/bin/flume-ng avro-client -c . -H m1 -p 4141 -F /usr/local/flume/log/log.00 #使用avro-client发送文件
控制台信息输出
Spool模式
Spool监测配置的目录下新增的文件,并将文件中的数据读取出来。需要注意两点:
1) 拷贝到spool目录下的文件不可以再打开编辑。
2) spool目录下不可包含相应的子目录
创建agent配置文件
vim /usr/local/flume/cong/spool.conf
-------------------------------------------spool.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = spooldir
a1.sources.r1.channels = c1
a1.sources.r1.spoolDir = /usr/local/flume/log
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
------------------------------------------------------------------------------------------------------
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/spool.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a2
echo "spool test1" > /usr/local/flume/log/spool_text.log #内容追加到指定目录
控制台信息输出
Exec模式
EXEC执行一个给定的命令获得输出的源,如果要使用tail命令,必选使得file足够大才能看到输出内容
创建agent配置文件
vim /usr/local/flume/cong/exec_tail.conf
-------------------------------------------exec_tail.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = exec
a1.sources.r1.channels = c1
a1.sources.r1.command = tail -F /usr/local/flume/log/log_exec_tail
# 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
----------------------------------------------------------------------------------------------------------
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/exec_tail.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a3
for i in {1..100};do echo "exec tail$i" >> /usr/local/flume/log/log_exec_tail;echo $i;sleep 0.1;done #内容追加到指定目录
控制台信息输出
Syslogtcp模式
1.Syslogtcp监听TCP的端口做为数据源
2.Syslogudp监听UDP的端口做为数据源
创建agent配置文件
vim /usr/local/flume/cong/syslog_tcp.conf
-------------------------------------------syslog_tcp.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5140
a1.sources.r1.host = localhost
a1.sources.r1.channels = c1
# 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
-----------------------------------------------------------------------------------------------------------
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/syslog_tcp.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a3
echo "hello idoall.org syslog" | nc localhost 5140 #测试产生syslog
控制台信息输出
JSONHandler模式
创建agent配置文件
vim /usr/local/flume/cong/post_json.conf
-------------------------------------------post_json.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = org.apache.flume.source.http.HTTPSource
a1.sources.r1.port = 8888
a1.sources.r1.channels = c1
# 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
---------------------------------------------------------------------------------------------------------
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/post_json.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a3
curl -X POST -d '[{ "headers" :{"a" : "a1","b" : "b1"},"body" : "idoall.org_body"}]' http://localhost:8888 #测试产生syslog
控制台信息输出
Hadoop sink模式
其中关于hadoop2.2.0部分的安装部署,请参考其他文档
创建agent配置文件
vim /usr/local/flume/cong/hdfs_sink.conf
-------------------------------------------hdfs_sink.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5140
a1.sources.r1.host = localhost
a1.sources.r1.channels = c1
# Describe the sink
a1.sinks.k1.type = hdfs
a1.sinks.k1.channel = c1
a1.sinks.k1.hdfs.path = hdfs://m1:9000/user/flume/syslogtcp
a1.sinks.k1.hdfs.filePrefix = Syslog
a1.sinks.k1.hdfs.round = true
a1.sinks.k1.hdfs.roundValue = 10
a1.sinks.k1.hdfs.roundUnit = minute
# 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
---------------------------------------------------------------------------------------------------------
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/hdfs_sink.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a3
echo "hello idoall flume -> hadoop testing one" | nc localhost 5140 #测试产生syslog
控制台信息输出
File Roll Sink模式
创建agent配置文件
vim /usr/local/flume/cong/file_roll.conf
-------------------------------------------file_roll.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5555
a1.sources.r1.host = localhost
a1.sources.r1.channels = c1
# Describe the sink
a1.sinks.k1.type = file_roll
a1.sinks.k1.sink.directory = /usr/local/flume/log
# 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
--------------------------------------------------------------------------------------------------------
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/file_roll.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a3
echo "hello idoall.org syslog" | nc localhost 5555
echo "hello idoall.org syslog 2" | nc localhost 5555
查看/usr/local/flume/log下是否生成文件,默认每30秒生成一个新文件
Replicating Channel Selector模式
Flume支持Fan out流从一个源到多个通道。有两种模式的Fan out,分别是复制和复用。在复制的情况下,流的事件被发送到
所有的配置通道。在复用的情况下,事件被发送到可用的渠道中的一个子集。Fan out流需要指定源和Fan out通道的规则。
这次需要用到m1,m2两台机器
在m1创建replicating_Channel_Selector配置文件
vim /usr/local/flume/cong/replicating_Channel_Selector.conf
------------------------------------------- replicating_Channel_Selector.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1 c2
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5140
a1.sources.r1.host = localhost
a1.sources.r1.channels = c1 c2
a1.sources.r1.selector.type = replicating
# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.channel = c1
a1.sinks.k1.hostname = m1
a1.sinks.k1.port = 5555
a1.sinks.k2.type = avro
a1.sinks.k2.channel = c2
a1.sinks.k2.hostname = m2
a1.sinks.k2.port = 5555
# Use a channel which buffers events in memory
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
---------------------------------------------------------------------------------------------------------------------------------
vim /usr/local/flume/cong/replicating_Channel_Selector_avro.conf
------------------------------------------- replicating_Channel_Selector_avro.conf-----------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = avro
a1.sources.r1.channels = c1
a1.sources.r1.bind = 0.0.0.0
a1.sources.r1.port = 5555
# 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
---------------------------------------------------------------------------------------------------------------------------------
在m1上将2个配置文件复制到m2上一份
打开4个窗口,在m1和m2上同时启动两个flume agent
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/replicating_Channel_Selector_avro.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/replicating_Channel_Selector.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
然后在m1或m2的任意一台机器上,测试产生syslog
echo "hello idoall.org syslog" | nc localhost 5140
在m1和m2的sink窗口,分别可以看到以下信息,这说明信息得到了同步:
/12/28 14:08:18 INFO ipc.NettyServer: Connection to /192.168.1.51:46844 disconnected.
/12/28 14:08:52 INFO ipc.NettyServer: [id: 0x90f8fe1f, /192.168.1.50:35873 => /192.168.1.50:5555] OPEN
/12/28 14:08:52 INFO ipc.NettyServer: [id: 0x90f8fe1f, /192.168.1.50:35873 => /192.168.1.50:5555] BOUND: /192.168.1.50:5555
/12/28 14:08:52 INFO ipc.NettyServer: [id: 0x90f8fe1f, /192.168.1.50:35873 => /192.168.1.50:5555] CONNECTED: /192.168.1.50:35873
/12/28 14:08:59 INFO ipc.NettyServer: [id: 0xd6318635, /192.168.1.51:46858 => /192.168.1.50:5555] OPEN
/12/28 14:08:59 INFO ipc.NettyServer: [id: 0xd6318635, /192.168.1.51:46858 => /192.168.1.50:5555] BOUND: /192.168.1.50:5555
/12/28 14:08:59 INFO ipc.NettyServer: [id: 0xd6318635, /192.168.1.51:46858 => /192.168.1.50:5555] CONNECTED: /192.168.1.51:46858
/12/28 14:09:20 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 68 65 6C 6C 6F 20 69 64 6F 61 6C 6C 2E 6F 72 67 hello idoall.org }
Multiplexing Channel Selector模式
在m1创建Multiplexing_Channel_Selector配置文件
vim /usr/local/flume/cong/Multiplexing_Channel_Selector.conf
------------------------------------------- Multiplexing_Channel_Selector.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1 c2
# Describe/configure the source
a1.sources.r1.type = org.apache.flume.source.http.HTTPSource
a1.sources.r1.port = 5140
a1.sources.r1.channels = c1 c2
a1.sources.r1.selector.type = multiplexing
a1.sources.r1.selector.header = type
#映射允许每个值通道可以重叠。默认值可以包含任意数量的通道。
a1.sources.r1.selector.mapping.baidu = c1
a1.sources.r1.selector.mapping.ali = c2
a1.sources.r1.selector.default = c1
# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.channel = c1
a1.sinks.k1.hostname = m1
a1.sinks.k1.port = 5555
a1.sinks.k2.type = avro
a1.sinks.k2.channel = c2
a1.sinks.k2.hostname = m2
a1.sinks.k2.port = 5555
# Use a channel which buffers events in memory
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
------------------------------------------------------------------------------------------------------------------------------
vim /usr/local/flume/cong/Multiplexing_Channel_Selector_avro.conf
------------------------------------------- Multiplexing_Channel_Selector_avro.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = avro
a1.sources.r1.channels = c1
a1.sources.r1.bind = 0.0.0.0
a1.sources.r1.port = 5555
# 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
-----------------------------------------------------------------------------------------------------------------------------------
在m1上将2个配置文件复制到m2上一份
打开4个窗口,在m1和m2上同时启动两个flume agent
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/Multiplexing_Channel_Selector.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/Multiplexing_Channel_Selector_avro.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
然后在m1或m2的任意一台机器上,测试产生syslog
curl -X POST -d '[{ "headers" :{"type" : "baidu"},"body" : "idoall_TEST1"}]' http://localhost:5140 && curl -X POST -d '[{ "headers" :{"type" : "ali"},"body" : "idoall_TEST2"}]' http://localhost:5140 && curl -X POST -d '[{ "headers" :{"type" : "qq"},"body" : "idoall_TEST3"}]' http://localhost:5140
----在m1的sink窗口,可以看到以下信息----->
16/12/28 14:32:21 INFO node.Application: Starting Sink k1
16/12/28 14:32:21 INFO node.Application: Starting Source r1
16/12/28 14:32:21 INFO source.AvroSource: Starting Avro source r1: { bindAddress: 0.0.0.0, port: 5555 }...
16/12/28 14:32:21 INFO instrumentation.MonitoredCounterGroup: Monitored counter group for type: SOURCE, name: r1: Successfully registered new MBean.
16/12/28 14:32:21 INFO instrumentation.MonitoredCounterGroup: Component type: SOURCE, name: r1 started
16/12/28 14:32:21 INFO source.AvroSource: Avro source r1 started.
16/12/28 14:32:36 INFO ipc.NettyServer: [id: 0xcf00eea6, /192.168.1.50:35916 => /192.168.1.50:5555] OPEN
16/12/28 14:32:36 INFO ipc.NettyServer: [id: 0xcf00eea6, /192.168.1.50:35916 => /192.168.1.50:5555] BOUND: /192.168.1.50:5555
16/12/28 14:32:36 INFO ipc.NettyServer: [id: 0xcf00eea6, /192.168.1.50:35916 => /192.168.1.50:5555] CONNECTED: /192.168.1.50:35916
16/12/28 14:32:44 INFO ipc.NettyServer: [id: 0x432f5468, /192.168.1.51:46945 => /192.168.1.50:5555] OPEN
16/12/28 14:32:44 INFO ipc.NettyServer: [id: 0x432f5468, /192.168.1.51:46945 => /192.168.1.50:5555] BOUND: /192.168.1.50:5555
16/12/28 14:32:44 INFO ipc.NettyServer: [id: 0x432f5468, /192.168.1.51:46945 => /192.168.1.50:5555] CONNECTED: /192.168.1.51:46945
16/12/28 14:34:11 INFO sink.LoggerSink: Event: { headers:{type=baidu} body: 69 64 6F 61 6C 6C 5F 54 45 53 54 31 idoall_TEST1 }
16/12/28 14:34:57 INFO sink.LoggerSink: Event: { headers:{type=qq} body: 69 64 6F 61 6C 6C 5F 54 45 53 54 33 idoall_TEST3 }
----在m2的sink窗口,可以看到以下信息----->
16/12/28 14:32:27 INFO node.Application: Starting Sink k1
16/12/28 14:32:27 INFO node.Application: Starting Source r1
16/12/28 14:32:27 INFO source.AvroSource: Starting Avro source r1: { bindAddress: 0.0.0.0, port: 5555 }...
16/12/28 14:32:27 INFO instrumentation.MonitoredCounterGroup: Monitored counter group for type: SOURCE, name: r1: Successfully registered new MBean.
16/12/28 14:32:27 INFO instrumentation.MonitoredCounterGroup: Component type: SOURCE, name: r1 started
16/12/28 14:32:27 INFO source.AvroSource: Avro source r1 started.
16/12/28 14:32:36 INFO ipc.NettyServer: [id: 0x7c2f0aec, /192.168.1.50:38104 => /192.168.1.51:5555] OPEN
16/12/28 14:32:36 INFO ipc.NettyServer: [id: 0x7c2f0aec, /192.168.1.50:38104 => /192.168.1.51:5555] BOUND: /192.168.1.51:5555
16/12/28 14:32:36 INFO ipc.NettyServer: [id: 0x7c2f0aec, /192.168.1.50:38104 => /192.168.1.51:5555] CONNECTED: /192.168.1.50:38104
16/12/28 14:32:44 INFO ipc.NettyServer: [id: 0x3d36f553, /192.168.1.51:48599 => /192.168.1.51:5555] OPEN
16/12/28 14:32:44 INFO ipc.NettyServer: [id: 0x3d36f553, /192.168.1.51:48599 => /192.168.1.51:5555] BOUND: /192.168.1.51:5555
16/12/28 14:32:44 INFO ipc.NettyServer: [id: 0x3d36f553, /192.168.1.51:48599 => /192.168.1.51:5555] CONNECTED: /192.168.1.51:48599
16/12/28 14:34:33 INFO sink.LoggerSink: Event: { headers:{type=ali} body: 69 64 6F 61 6C 6C 5F 54 45 53 54 32 idoall_TEST2 }
可以看到,根据header中不同的条件分布到不同的channel上
Flume Sink Processors 模式
failover的机器是一直发送给其中一个sink,当这个sink不可用的时候,自动发送到下一个sink
在m1创建Flume_Sink_Processors配置文件
vim /usr/local/flume/cong/Flume_Sink_Processors.conf
------------------------------------------- Flume_Sink_Processors.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1 c2
#这个是配置failover的关键,需要有一个sink group
a1.sinkgroups = g1
a1.sinkgroups.g1.sinks = k1 k2
#处理的类型是failover
a1.sinkgroups.g1.processor.type = failover
#优先级,数字越大优先级越高,每个sink的优先级必须不相同
a1.sinkgroups.g1.processor.priority.k1 = 5
a1.sinkgroups.g1.processor.priority.k2 = 10
#设置为10秒,当然可以根据你的实际状况更改成更快或者很慢
a1.sinkgroups.g1.processor.maxpenalty = 10000
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5140
a1.sources.r1.channels = c1 c2
a1.sources.r1.selector.type = replicating
# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.channel = c1
a1.sinks.k1.hostname = m1
a1.sinks.k1.port = 5555
a1.sinks.k2.type = avro
a1.sinks.k2.channel = c2
a1.sinks.k2.hostname = m2
a1.sinks.k2.port = 5555
# Use a channel which buffers events in memory
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
-----------------------------------------------------------------------------------------------------------------------------------
vim /usr/local/flume/cong/Flume_Sink_Processors_avro.conf
------------------------------------------- Flume_Sink_Processors_avro.conf--------------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = avro
a1.sources.r1.channels = c1
a1.sources.r1.bind = 0.0.0.0
a1.sources.r1.port = 5555
# 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
------------------------------------------------------------------------------------------------------------------------------------
在m1上将2个配置文件复制到m2上一份
打开4个窗口,在m1和m2上同时启动两个flume agent
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/Flume_Sink_Processors_avro.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/Flume_Sink_Processors.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
然后在m1或m2的任意一台机器上,测试产生log
echo "idoall.org test1 failover" | nc localhost 5140
因为m2的优先级高,所以在m2的sink窗口,可以看到以下信息,而m1没有:
16/12/28 15:02:46 INFO ipc.NettyServer: Connection to /192.168.1.51:48692 disconnected.
16/12/28 15:03:12 INFO ipc.NettyServer: [id: 0x09a14036, /192.168.1.51:48704 => /192.168.1.51:5555] OPEN
16/12/28 15:03:12 INFO ipc.NettyServer: [id: 0x09a14036, /192.168.1.51:48704 => /192.168.1.51:5555] BOUND: /192.168.1.51:5555
16/12/28 15:03:12 INFO ipc.NettyServer: [id: 0x09a14036, /192.168.1.51:48704 => /192.168.1.51:5555] CONNECTED: /192.168.1.51:48704
16/12/28 15:03:26 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 31 idoall.org test1 }
这时我们停止掉m2机器上的sink(ctrl+c),再次输出测试数据:
echo "idoall.org test2 failover" | nc localhost 5140
可以在m1的sink窗口,看到读取到了刚才发送的两条测试数据:
16/12/28 15:02:46 INFO ipc.NettyServer: Connection to /192.168.1.51:47036 disconnected.
16/12/28 15:03:12 INFO ipc.NettyServer: [id: 0xbcf79851, /192.168.1.51:47048 => /192.168.1.50:5555] OPEN
16/12/28 15:03:12 INFO ipc.NettyServer: [id: 0xbcf79851, /192.168.1.51:47048 => /192.168.1.50:5555] BOUND: /192.168.1.50:5555
16/12/28 15:03:12 INFO ipc.NettyServer: [id: 0xbcf79851, /192.168.1.51:47048 => /192.168.1.50:5555] CONNECTED: /192.168.1.51:47048
16/12/28 15:07:56 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 31 idoall.org test1 }
16/12/28 15:07:56 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 32 idoall.org test2 }
我们再在m2的sink窗口中,启动sink:
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/Flume_Sink_Processors_avro.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
输入两批测试数据
echo "idoall.org test3 failover" | nc localhost 5140 && echo "idoall.org test4 failover" | nc localhost 5140
在m2的sink窗口,我们可以看到以下信息,因为优先级的关系,log消息会再次落到m2上
16/12/28 15:09:47 INFO node.Application: Starting Sink k1
16/12/28 15:09:47 INFO node.Application: Starting Source r1
16/12/28 15:09:47 INFO source.AvroSource: Starting Avro source r1: { bindAddress: 0.0.0.0, port: 5555 }...
16/12/28 15:09:47 INFO instrumentation.MonitoredCounterGroup: Monitored counter group for type: SOURCE, name: r1: Successfully registered new MBean.
16/12/28 15:09:47 INFO instrumentation.MonitoredCounterGroup: Component type: SOURCE, name: r1 started
16/12/28 15:09:47 INFO source.AvroSource: Avro source r1 started.
16/12/28 15:09:54 INFO ipc.NettyServer: [id: 0x96615732, /192.168.1.51:48741 => /192.168.1.51:5555] OPEN
16/12/28 15:09:54 INFO ipc.NettyServer: [id: 0x96615732, /192.168.1.51:48741 => /192.168.1.51:5555] BOUND: /192.168.1.51:5555
16/12/28 15:09:54 INFO ipc.NettyServer: [id: 0x96615732, /192.168.1.51:48741 => /192.168.1.51:5555] CONNECTED: /192.168.1.51:48741
16/12/28 15:09:57 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 32 idoall.org test2 }
16/12/28 15:10:43 INFO ipc.NettyServer: [id: 0x12621f9a, /192.168.1.50:38166 => /192.168.1.51:5555] OPEN
16/12/28 15:10:43 INFO ipc.NettyServer: [id: 0x12621f9a, /192.168.1.50:38166 => /192.168.1.51:5555] BOUND: /192.168.1.51:5555
16/12/28 15:10:43 INFO ipc.NettyServer: [id: 0x12621f9a, /192.168.1.50:38166 => /192.168.1.51:5555] CONNECTED: /192.168.1.50:38166
16/12/28 15:10:43 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 33 idoall.org test3 }
16/12/28 15:10:43 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 34 idoall.org test4 }
Load balancing Sink Processor模式
load balance type和failover不同的地方是,load balance有两个配置,一个是轮询,一个是随机。两种情况下如果被选择的sink不可用,就会自动尝试发送到下一个可用的sink上面。
在m1创建Load_balancing_Sink_Processors配置文件
vim /usr/local/flume/cong/Load_balancing_Sink_Processors.conf
------------------------------------------- Load_balancing_Sink_Processors.conf------------------------------------------------
a1.sources = r1
a1.sinks = k1 k2
a1.channels = c1
#这个是配置Load balancing的关键,需要有一个sink group
a1.sinkgroups = g1
a1.sinkgroups.g1.sinks = k1 k2
a1.sinkgroups.g1.processor.type = load_balance
a1.sinkgroups.g1.processor.backoff = true
a1.sinkgroups.g1.processor.selector = round_robin
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5140
a1.sources.r1.channels = c1
# Describe the sink
a1.sinks.k1.type = avro
a1.sinks.k1.channel = c1
a1.sinks.k1.hostname = m1
a1.sinks.k1.port = 5555
a1.sinks.k2.type = avro
a1.sinks.k2.channel = c1
a1.sinks.k2.hostname = m2
a1.sinks.k2.port = 5555
# Use a channel which buffers events in memory
a1.channels.c1.type = memory
a1.channels.c1.capacity = 1000
a1.channels.c1.transactionCapacity = 100
---------------------------------------------------------------------------------------------------------------------------------
在m1创建Load_balancing_Sink_Processors_avro配置文件
vim /usr/local/flume/cong/Load_balancing_Sink_Processors_avro.conf
------------------------------------------- Load_balancing_Sink_Processors_avro.conf----------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = avro
a1.sources.r1.channels = c1
a1.sources.r1.bind = 0.0.0.0
a1.sources.r1.port = 5555
# 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
----------------------------------------------------------------------------------------------------------------------------------
在m1上将2个配置文件复制到m2上一份
打开4个窗口,在m1和m2上同时启动两个flume agent
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/Load_balancing_Sink_Processors_avro.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/Load_balancing_Sink_Processors.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
然后在m1或m2的任意一台机器上,测试产生log,一行一行输入,输入太快,容易落到一台机器上
echo "idoall.org test1" | nc localhost 5140
echo "idoall.org test2" | nc localhost 5140
echo "idoall.org test3" | nc localhost 5140
echo "idoall.org test4" | nc localhost 5140
在m1的sink窗口,可以看到以下信息:
16/12/28 15:35:29 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 32 idoall.org test2 }
16/12/28 15:35:33 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 34 idoall.org test4 }
在m2的sink窗口,可以看到以下信息:
16/12/28 15:35:27 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 31 idoall.org test1 }
16/12/28 15:35:29 INFO sink.LoggerSink: Event: { headers:{Severity=0, flume.syslog.status=Invalid, Facility=0} body: 69 64 6F 61 6C 6C 2E 6F 72 67 20 74 65 73 74 33 idoall.org test3 }
说明轮询模式起到了作用
Hbase sink模式
在测试之前,请先参考《ubuntu12.04+hadoop2.2.0+zookeeper3.4.5+hbase0.96.2+hive0.13.1分布式环境部署》将hbase启动
然后将以下文件复制到flume中:
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/protobuf-java-2.5.0.jar /usr/local/flume/lib
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/hbase-client-0.96.2-hadoop2.jar /usr/local/flume/lib
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/hbase-common-0.96.2-hadoop2.jar /usr/local/flume/lib
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/hbase-protocol-0.96.2-hadoop2.jar /usr/local/flume/lib
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/hbase-server-0.96.2-hadoop2.jar /usr/local/flume/lib
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/hbase-hadoop2-compat-0.96.2-hadoop2.jar /usr/local/flume/lib
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/hbase-hadoop-compat-0.96.2-hadoop2.jar /usr/local/flume/lib@@@
cp /home/hadoop/hbase-0.96.2-hadoop2/lib/htrace-core-2.04.jar /usr/local/flume/lib
确保test_idoall_org表在hbase中已经存在,test_idoall_org表的格式以及字段请参考《ubuntu12.04+hadoop2.2.0+zookeeper3.4.5+hbase0.96.2+hive0.13.1分布式环境部署》中关于hbase部分的建表代码。
在m1创建hbase_simple配置文件
vim /usr/local/flume/cong/hbase_simple.conf
------------------------------------------- hbase_simple.conf------------------------------------------------------
a1.sources = r1
a1.sinks = k1
a1.channels = c1
# Describe/configure the source
a1.sources.r1.type = syslogtcp
a1.sources.r1.port = 5140
a1.sources.r1.host = localhost
a1.sources.r1.channels = c1
# Describe the sink
a1.sinks.k1.type = logger
a1.sinks.k1.type = hbase
a1.sinks.k1.table = test_idoall_org
a1.sinks.k1.columnFamily = name
a1.sinks.k1.column = idoall
a1.sinks.k1.serializer = org.apache.flume.sink.hbase.RegexHbaseEventSerializer
a1.sinks.k1.channel = memoryChannel
# 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
-------------------------------------------------------------------------------------------------------------------
启动flume agent
/usr/local/flume/bin/flume-ng agent -c . -f /usr/local/flume/conf/hbase_simple.conf -n a1 -Dflume.root.logger=INFO,console #启动flume agent a1
测试产生syslog
echo "hello idoall.org from flume" | nc localhost 5140
这时登录到hbase中,可以发现新数据已经插入
Kafka模式
1创建kafka.conf配置文件
vim /usr/local/flume/cong/kafka.conf
------------------------------------------- kafka.conf------------------------------------------------------
agent.sources = r1
agent.channels = c1
agent.sinks = s1
# For each one of the sources, the type is defined
agent.sources.r1.type = syslogudp
agent.sources.r1.bind = 192.168.201.73
agent.sources.r1.port = 7520
#消息平均分布在多个分区上 官方自带插件
agent.sources.r1.interceptors = i2
agent.sources.r1.interceptors.i2.type=org.apache.flume.sink.solr.morphline.UUIDInterceptor$Builder
agent.sources.r1.interceptors.i2.headerName=key
agent.sources.r1.interceptors.i2.preserveExisting=false
# The channel can be defined as follows.
agent.sources.r1.channels = c1
# Each sink's type must be defined
agent.sinks.s1.type = org.apache.flume.sink.kafka.KafkaSink
agent.sinks.s1.topic=php #kafka消息池
agent.sinks.s1.brokerList=192.168.201.73:9092 #kafka主机和端口
agent.sinks.s1.requiredAcks = 1
agent.sinks.s1.batchSize = 20
agent.sinks.s1.channel = c1
#agent.sinks.s1.requireAcks 0不需要确认消息是否成功 -1 确认消息成功 还需备份
#agent.sinks.s1.batchSize 配置每次下沉多少条消息 每次下沉的数量越多延迟越高
# Each channel's type is defined.
agent.channels.c1.type = memory
# Other config values specific to each type of channel(sink or source)
# can be defined as well
# In this case, it specifies the capacity of the memory channel
agent.channels.c1.capacity = 10000
-----------------------------------------------------------------------------------------------------------------
启动flume /usr/local/flume/bin/flume-ng agent --conf conf -f conf/kafka.conf -n agent&