共计 12318 个字符，预计需要花费 31 分钟才能阅读完成。
最近一段时间，负责公司的产品日志埋点与收集工作，搭建了基于 Flume+HDFS+Hive 日志搜集系统。
一、日志搜集系统架构：
简单画了一下日志搜集系统的架构图，可以看出，flume 承担了 agent 与 collector 角色，HDFS 承担了数据持久化存储的角色。
作者搭建的服务器是个 demo 版，只用到了一个 flume_collector，数据只存储在 HDFS。当然高可用的日志搜集处理系统架构是需要多台 flume collector 做负载均衡与容错处理的。
二、日志产生：
1、log4j 配置，每隔 1 分钟 roll 一个文件，如果 1 分钟之内文件大于 5M，则再生成一个文件。
<!-- 产品数据分析日志 按分钟分 -->
<RollingRandomAccessFile name="RollingFile_product_minute"
fileName="${STAT_LOG_HOME}/${SERVER_NAME}_product.log"
filePattern="${STAT_LOG_HOME}/${SERVER_NAME}_product.log.%d{yyyy-MM-dd-HH-mm}-%i">
<PatternLayout charset="UTF-8"
pattern="%d{yyyy-MM-dd HH:mm:ss.SSS} %level - %msg%xEx%n" />
<Policies>
<TimeBasedTriggeringPolicy interval="1"
modulate="true" />
<SizeBasedTriggeringPolicy size="${EVERY_FILE_SIZE}" />
</Policies>
<Filters>
<ThresholdFilter level="INFO" onMatch="ACCEPT"
onMismatch="NEUTRAL" />
</Filters>
</RollingRandomAccessFile> roll 后的文件格式如下
2、日志内容
json 格式文件，最外层 json 按顺序为:tableName,logRequest,timestamp,statBody,logResponse,resultCode,resultMsg
2016-11-30 09:18:21.916 INFO - {"tableName": "ReportView",
"logRequest": {***
},
"timestamp": 1480468701432,
"statBody": {***
},
"logResponse": {***
},
"resultCode": 1,
"resultFailMsg": ""
} 三、flume 配置
虚拟机环境，请参考 http://www.linuxidc.com/Linux/2016-12/137955.htm
Hadoop 环境，请参考 http://www.linuxidc.com/Linux/2016-12/137957.htm
此处 flume 环境是
CentOS1:flume-agent
centos2:flume-collector
1、flume agent 配置，conf 文件
a1.sources = linuxidcSource
a1.channels = linuxidcChannel
a1.sinks = linuxidcSink
a1.sources.linuxidcSource.type = spooldir
a1.sources.linuxidcSource.channels = linuxidcChannel
#日志目录
a1.sources.linuxidcSource.spoolDir = /opt/flumeSpool
a1.sources.linuxidcSource.fileHeader = true
#日志内容处理完后，会生成.COMPLETED 后缀的文件，同时.log 文件每一分钟 roll 一个，此处忽略.log 文件与.COMPLETED 文件
a1.sources.linuxidcSource.ignorePattern=([^_]+)|(.*(\.log)$)|(.*(\.COMPLETED)$)
a1.sources.linuxidcSource.basenameHeader=true
a1.sources.linuxidcSource.deserializer.maxLineLength=102400
#自定义拦截器，对 json 格式的源日志进行字段分隔，并添加 timestamp，为后面的 hdfsSink 做处理，拦截器代码见后面
a1.sources.linuxidcSource.interceptors=i1
a1.sources.linuxidcSource.interceptors.i1.type=com.linuxidc.flume_interceptor.HiveLogInterceptor2$Builder
a1.sinks.linuxidcSink.type = avro
a1.sinks.linuxidcSink.channel = linuxidcChannel
a1.sinks.linuxidcSink.hostname = centos2
a1.sinks.linuxidcSink.port = 4545
#此处配置 deflate 压缩后，hive collector 那边一定也要相应配置解压缩
a1.sinks.linuxidcSink.compression-type=deflate
a1.channels.linuxidcChannel.type=memory
a1.channels.linuxidcChannel.capacity=10000
a1.channels.linuxidcChannel.transactionCapacity=1000 2、flume collector 配置
a1.sources = avroSource
a1.channels = memChannel
a1.sinks = hdfsSink
a1.sources.avroSource.type = avro
a1.sources.avroSource.channels = memChannel
a1.sources.avroSource.bind=centos2
a1.sources.avroSource.port=4545
#与 flume agent 配置对应
a1.sources.avroSource.compression-type=deflate
a1.sinks.hdfsSink.type = hdfs
a1.sinks.hdfsSink.channel = memChannel
# linuxidc_hive_log 为 hive 表，按年 - 月 - 日分区存储，a1.sinks.hdfsSink.hdfs.path=hdfs://centos1:9000/flume/linuxidc_hive_log/dt=%Y-%m-%d
a1.sinks.hdfsSink.hdfs.batchSize=10000
a1.sinks.hdfsSink.hdfs.fileType=DataStream
a1.sinks.hdfsSink.hdfs.writeFormat=Text
a1.sinks.hdfsSink.hdfs.rollSize=10240000
a1.sinks.hdfsSink.hdfs.rollCount=0
a1.sinks.hdfsSink.hdfs.rollInterval=300
a1.channels.memChannel.type=memory
a1.channels.memChannel.capacity=100000
a1.channels.memChannel.transactionCapacity=10000 四、hive 表创建与分区
1、hive 表创建
在 hive 中执行建表语句后，hdfs://centos1:9000/flume/ 目录下新生成了 linuxidc_hive_log 目录。（建表语句里面有 location 关键字）
\u0001 表示 hive 通过该分隔符进行字段分离，该字符在 linux 用 vim 编辑器打开是 ^A。
由于日志格式是 JSON 格式，因为需要将 JSON 格式转换成 \u0001 字符分隔，并通过 dt 进行分区。这一步通过 flume 自定义拦截器来完成。
CREATE TABLE `linuxidc_hive_log`(
`tableNmae` string,
`logRequest` string,
`timestamp` bigint,
`statBody` string,
`logResponse` string,
`resultCode` int,
`resultFailMsg` string
)
PARTITIONED BY (`dt` string)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY '\u0001'
STORED AS INPUTFORMAT
'org.apache.hadoop.mapred.TextInputFormat'
OUTPUTFORMAT
'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
LOCATION
'hdfs://centos1:9000/flume/linuxidc_hive_log'; 2、hive 表分区
for ((i=-1;i<=365;i++))
do
dt=$(date -d "$(date +%F) ${i} days" +%Y-%m-%d)
echo date=$dt
hive -e "ALTER TABLE linuxidc_hive_log ADD PARTITION(dt='${dt}')" >> logs/init_linuxidc_hive_log.out 2>>logs/init_linuxidc_hive_log.err
done 五、自定义 flume 拦截器
新建 maven 工程，拦截器 HiveInterceptor2 代码如下。
package com.linuxidc.flume_interceptor;
import java.util.ArrayList;
import java.util.List;
import java.util.Map;
import org.apache.flume.Context;
import org.apache.flume.Event;
import org.apache.flume.interceptor.Interceptor;
import org.apache.flume.interceptor.TimestampInterceptor.Constants;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.alibaba.fastjson.JSONObject;
import com.google.common.base.Charsets;
import com.google.common.base.Joiner;
public class HiveLogInterceptor2 implements Interceptor
{private static Logger logger = LoggerFactory.getLogger(HiveLogInterceptor2.class);
public static final String HIVE_SEPARATOR = "\001";
public void close()
{// TODO Auto-generated method stub
}
public void initialize()
{// TODO Auto-generated method stub
}
public Event intercept(Event event)
{String orginalLog = new String(event.getBody(), Charsets.UTF_8);
try
{String log = this.parseLog(orginalLog);
// 设置时间, 用于 hdfsSink
long now = System.currentTimeMillis();
Map headers = event.getHeaders();
headers.put(Constants.TIMESTAMP, Long.toString(now));
event.setBody(log.getBytes());
} catch (Throwable throwable)
{logger.error(("errror when intercept,log [" + orginalLog + "]"), throwable);
return null;
}
return event;
}
public List<Event> intercept(List<Event> list)
{List<Event> events = new ArrayList<Event>();
for (Event event : list)
{Event interceptedEvent = this.intercept(event);
if (interceptedEvent != null)
{events.add(interceptedEvent);
}
}
return events;
}
private static String parseLog(String log)
{List<String> logFileds = new ArrayList<String>();
String dt = log.substring(0, 10);
String keyStr = "INFO -";
int index = log.indexOf(keyStr);
String content = "";
if (index != -1)
{content = log.substring(index + keyStr.length(), log.length());
}
//针对不同 OS，使用不同回车换行符号
content = content.replaceAll("\r", "");
content = content.replaceAll("\n", "\\\\" + System.getProperty("line.separator"));
JSONObject jsonObj = JSONObject.parseObject(content);
String tableName = jsonObj.getString("tableName");
String logRequest = jsonObj.getString("logRequest");
String timestamp = jsonObj.getString("timestamp");
String statBody = jsonObj.getString("statBody");
String logResponse = jsonObj.getString("logResponse");
String resultCode = jsonObj.getString("resultCode");
String resultFailMsg = jsonObj.getString("resultFailMsg");
//字段分离
logFileds.add(tableName);
logFileds.add(logRequest);
logFileds.add(timestamp);
logFileds.add(statBody);
logFileds.add(logResponse);
logFileds.add(resultCode);
logFileds.add(resultFailMsg);
logFileds.add(dt);
return Joiner.on(HIVE_SEPARATOR).join(logFileds);
}
public static class Builder implements Interceptor.Builder
{public Interceptor build()
{return new HiveLogInterceptor2();}
public void configure(Context arg0)
{}}
} pom.xml 增加如下配置，将 flume 拦截器工程进行 maven 打包，jar 包与依赖包均拷到 ${flume-agent}/lib 目录
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-dependency-plugin</artifactId>
<configuration>
<outputDirectory>
${project.build.directory}
</outputDirectory>
</configuration>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-dependency-plugin</artifactId>
<executions>
<execution>
<id>copy-dependencies</id>
<phase>prepare-package</phase>
<goals>
<goal>copy-dependencies</goal>
</goals>
<configuration>
<outputDirectory>${project.build.directory}/lib</outputDirectory>
<overWriteReleases>true</overWriteReleases>
<overWriteSnapshots>true</overWriteSnapshots>
<overWriteIfNewer>true</overWriteIfNewer>
</configuration>
</execution>
</executions>
</plugin>
</plugins>
</build> 对日志用分隔符 ”\001″ 进行分隔，。经拦截器处理后的日志格式如下，^A 即是 ”\001″
ReportView^A{"request":{},"requestBody":{"detailInfos":[],"flag":"","reportId":7092,"pageSize":0,"searchs":[],"orders":[],"pageNum":1}}^A1480468701432^A{"sourceId":22745,"reportId":7092,"projectId":29355,"userId":2532}^A{"responseBody":{"statusCodeValue":200,"httpHeaders":{},"body":{"msg":" 请求成功 ","httpCode":200,"timestamp":1480468701849},"statusCode":"OK"},"response":{}}^A1^A^A2016-11-30至此，flume+Hdfs+Hive 的配置均已完成。
后续可以通过 mapreduce 或者 HQL 对数据进行分析。
六、启动运行与结果
1、启动 hadoop hdfs
参考前一篇文章：Hadoop 1.2 集群搭建与环境配置 http://www.linuxidc.com/Linux/2016-12/137957.htm
2、启动 flume_collector 和 flume_agent，由于 flume 启动命令参数太多，自己写了一个启动脚本
start-Flume.sh
#!/bin/bash
jps -l|grep org.apache.flume.node.Application|awk '{print $1}'|xargs kill -9 2>&1 >/dev/null
cd "$(dirname "$0")"
cd ..
nohup bin/flume-ng agent --conf conf --conf-file conf/flume-conf.properties --name a1 2>&1 > /dev/null &3、hdfs 查看数据
可以看到搜集的日志已经上传到 HDFS 上
[root@centos1 bin]# rm -rf FlumeData.1480587273016.tmp
[root@centos1 bin]# hadoop fs -ls /flume/linuxidc_hive_log/dt=2016-12-01/
Found 3 items
-rw-r--r-- 3 root supergroup 5517 2016-12-01 08:12 /flume/linuxidc_hive_log/dt=2016-12-01/FlumeData.1480608753042.tmp
-rw-r--r-- 3 root supergroup 5517 2016-12-01 08:40 /flume/linuxidc_hive_log/dt=2016-12-01/FlumeData.1480610453116
-rw-r--r-- 3 root supergroup 5517 2016-12-01 08:44 /flume/linuxidc_hive_log/dt=2016-12-01/FlumeData.1480610453117
[root@centos1 bin]# 4、启动 hive，查看数据，可以看到 hive 已经可以加载 hdfs 数据
[root@centos1 lib]# hive
Logging initialized using configuration in file:/root/apache-hive-1.2.1-bin/conf/hive-log4j.properties
hive> select * from linuxidc_hive_log limit 2;
OK
ReportView {"request":{},"requestBody":{"detailInfos":[],"flag":"","reportId":7092,"pageSize":0,"searchs":[],"orders":[],"pageNum":1}} 1480468701432 {"sourceId":22745,"reportId":7092,"projectId":29355,"userId":2532} {"responseBody":{"statusCodeValue":200,"httpHeaders":{},"body":{"msg":"请求成功","httpCode":200,"timestamp":1480468701849},"statusCode":"OK"},"response":{}} 1 2016-12-01
ReportDesignResult {"request":{},"requestBody":{"sourceId":22745,"detailInfos":[{"colName":"月份 ","flag":"0","reportId":7092,"colCode":"col_2_22745","pageSize":20,"type":"1","pageNum":1,"rcolCode":"col_25538","colType":"string","formula":"","id":25538,"position":"row","colId":181664,"dorder":1,"pColName":" 月份 ","pRcolCode":"col_25538"},{"colName":" 综合利率 (合计)","flag":"1","reportId":7092,"colCode":"col_11_22745","pageSize":20,"type":"1","pageNum":1,"rcolCode":"sum_col_25539","colType":"number","formula":"sum","id":25539,"position":"group","colId":181673,"dorder":1,"pColName":" 综合利率 ","pRcolCode":"col_25539"}],"flag":"bar1","reportId":7092,"reportName":"iiiissszzzV","pageSize":100,"searchs":[],"orders":[],"pageNum":1,"projectId":29355}} 1480468703586{"reportType":"bar1","sourceId":22745,"reportId":7092,"num":5,"usedFields":" 月份 $$ 综合利率 (合计)$$","projectId":29355,"userId":2532} {"responseBody":{"statusCodeValue":200,"httpHeaders":{},"body":{"msg":" 请求成功","reportId":7092,"httpCode":200,"timestamp":1480468703774},"statusCode":"OK"},"response":{}} 1 2016-12-01
Time taken: 2.212 seconds, Fetched: 2 row(s)
hive>七、常见问题与处理方法
1、FATAL: Spool Directory source linuxidcSource: {spoolDir: /opt/flumeSpool}: Uncaught exception in SpoolDirectorySource thread. Restart or reconfigure Flume to continue processing.
java.nio.charset.MalformedInputException: Input length = 1
可能原因：
1、字符编码问题，spoolDir 目录下的日志文件必须是 UTF-8
2、使用 Spooling Directory Source 的时候，一定要避免同时读写一个文件的情况,conf 文件增加如下配置
a1.sources.linuxidcSource.ignorePattern=([^_]+)|(.*(\.log)$)|(.*(\.COMPLETED)$)
2、日志导入到 hadoop 目录，但是 hive 表查询无数据。如 hdfs://centos1:9000/flume/linuxidc_hive_log/dt=2016-12-01/ 下面有数据，
hive 查询 select * from linuxidc_hive_log 却无数据
可能原因：
1、建表的时候，没有建立分区。即使 flume 进行了配置（a1.sinks.hdfsSink.hdfs.path=hdfs://centos1:9000/flume/linuxidc_hive_log/dt=%Y-%m-%d），但是表的分区结构没有建立，因此文件导入到 HDFS 上后，HIVE 并不能读取。
解决方法：先创建分区，建立 shell 可执行文件，将该表的分区先建好
for ((i=-10;i<=365;i++))
do
dt=$(date -d "$(date +%F) ${i} days" +%Y-%m-%d)
echo date=$dt
hive -e "ALTER TABLE linuxidc_hive_log ADD PARTITION(dt='${dt}')" >> logs/init_linuxidc_hive_log.out 2>>logs/init_linuxidc_hive_log.err
done2、也可能是文件在 hdfs 上还是.tmp 文件，仍然被 hdfs 在写入。.tmp 文件 hive 暂时无法读取，只能读取非.tmp 文件。
解决方法：等待 hdfs 配置的 roll 间隔时间，或者达到一定大小后 tmp 文件重命名为 hdfs 上的日志文件后，再查询 hive，即可查到。
Hadoop 如何修改 HDFS 文件存储块大小 http://www.linuxidc.com/Linux/2013-09/90100.htm
将本地文件拷到 HDFS 中 http://www.linuxidc.com/Linux/2013-05/83866.htm
从 HDFS 下载文件到本地 http://www.linuxidc.com/Linux/2012-11/74214.htm
将本地文件上传至 HDFS http://www.linuxidc.com/Linux/2012-11/74213.htm
HDFS 基本文件常用命令 http://www.linuxidc.com/Linux/2013-09/89658.htm
Hadoop 中 HDFS 和 MapReduce 节点基本简介 http://www.linuxidc.com/Linux/2013-09/89653.htm
《Hadoop 实战》中文版 + 英文文字版 + 源码【PDF】http://www.linuxidc.com/Linux/2012-10/71901.htm
Hadoop: The Definitive Guide【PDF 版】http://www.linuxidc.com/Linux/2012-01/51182.htm
更多 Hadoop 相关信息见Hadoop 专题页面 http://www.linuxidc.com/topicnews.aspx?tid=13
本文永久更新链接地址：http://www.linuxidc.com/Linux/2016-12/137959.htm
