logstash一次同步Mysql多张表到ES深入详解

题记

一次同步多张表是开发中的一般需求。之前研究了很久找到方法，但没有详细总结。
博友前天在线提问，说明这块理解的还不够透彻。
我整理下，
一是为了尽快解决博友问题，
二是加深记忆，便于未来产品开发中快速上手。

1、同步原理

原有ES专栏中有详解，不再赘述。详细请参考我的专栏：
深入详解Elasticsearch
以下是通过ES5.4.0， logstash5.4.1 验证成功。
可以确认的是2.X版本同样可以验证成功。

2、核心配置文件

input {
  stdin {
  }

  jdbc {
  type => "cxx_article_info"
  # mysql jdbc connection string to our backup databse 后面的test对应mysql中的test数据库
  jdbc_connection_string => "jdbc:mysql://110.10.15.37:3306/cxxwb"
  # the user we wish to excute our statement as
  jdbc_user => "root"
  jdbc_password => "xxxxx"

  record_last_run => "true"
  use_column_value => "true"
  tracking_column => "id"
  last_run_metadata_path => "/opt/logstash/bin/logstash_xxy/cxx_info"
  clean_run => "false"

  # the path to our downloaded jdbc driver
  jdbc_driver_library => "/opt/elasticsearch/lib/mysql-connector-java-5.1.38.jar"
  # the name of the driver class for mysql
  jdbc_driver_class => "com.mysql.jdbc.Driver"
  jdbc_paging_enabled => "true"
  jdbc_page_size => "500"
  statement => "select * from cxx_article_info where id > :sql_last_value"
#定时字段 各字段含义（由左至右）分、时、天、月、年，全部为*默认含义为每分钟都更新
  schedule => "* * * * *"
#设定ES索引类型
  }

  jdbc {
  type => "cxx_user"
  # mysql jdbc connection string to our backup databse 后面的test对应mysql中的test数据库
  jdbc_connection_string => "jdbc:mysql://110.10.15.37:3306/cxxwb"
  # the user we wish to excute our statement as
  jdbc_user => "root"
  jdbc_password => "xxxxxx"

  record_last_run => "true"
  use_column_value => "true"
  tracking_column => "id"
  last_run_metadata_path => "/opt/logstash/bin/logstash_xxy/cxx_user_info"
  clean_run => "false"

  # the path to our downloaded jdbc driver
  jdbc_driver_library => "/opt/elasticsearch/lib/mysql-connector-java-5.1.38.jar"
  # the name of the driver class for mysql
  jdbc_driver_class => "com.mysql.jdbc.Driver"
  jdbc_paging_enabled => "true"
  jdbc_page_size => "500"
  statement => "select * from cxx_user_info where id > :sql_last_value"
#以下对应着要执行的sql的绝对路径。
#statement_filepath => "/opt/logstash/bin/logstash_mysql2es/department.sql"
#定时字段 各字段含义（由左至右）分、时、天、月、年，全部为*默认含义为每分钟都更新
schedule => "* * * * *"
#设定ES索引类型
  }

}

filter {
mutate {
  convert => [ "publish_time", "string" ]
 }

date {
  timezone => "Europe/Berlin"
  match => ["publish_time" , "ISO8601", "yyyy-MM-dd HH:mm:ss"]
}
#date {
 # match => [ "publish_time", "yyyy-MM-dd HH:mm:ss,SSS" ]
  # remove_field => [ "publish_time" ]
  # }
json {
  source => "message"
  remove_field => ["message"]
  }
}

output {

if [type]=="cxxarticle_info" {
  elasticsearch {
#ESIP地址与端口
  hosts => "10.100.11.231:9200"
#ES索引名称（自己定义的）
  index => "cxx_info_index"
#自增ID编号
 # document_id => "%{id}"
  }
}

if [type]=="cxx_user" {
  elasticsearch {
#ESIP地址与端口
  hosts => "10.100.11.231:9200"
#ES索引名称（自己定义的）
  index => "cxx_user_index"
#自增ID编号
 # document_id => "%{id}"
  }
}

}

3、同步成功结果

[2017-07-19T15:08:05,438][INFO ][logstash.pipeline ] Pipeline main started
The stdin plugin is now waiting for input:
[2017-07-19T15:08:05,491][INFO ][logstash.agent ] Successfully started Logstash API endpoint {:port=>9600}
[2017-07-19T15:09:00,721][INFO ][logstash.inputs.jdbc ](0.007000s) SELECT count(*) AS `count` FROM (select * from cxx_article_info where id > 0) AS `t1` LIMIT 1
[2017-07-19T15:09:00,721][INFO ][logstash.inputs.jdbc ](0.008000s) SELECT count(*) AS `count` FROM (select * from cxx_user_info where id > 0) AS `t1` LIMIT 1
[2017-07-19T15:09:00,730][INFO ][logstash.inputs.jdbc ](0.004000s) SELECT * FROM (select * from cxx_user_info where id > 0) AS `t1` LIMIT 500 OFFSET 0
[2017-07-19T15:09:00,731][INFO ][logstash.inputs.jdbc ](0.007000s) SELECT * FROM (select * from cxx_article_info where id > 0) AS `t1` LIMIT 500 OFFSET 0
[2017-07-19T15:10:00,173][INFO ][logstash.inputs.jdbc ](0.002000s) SELECT count(*) AS `count` FROM (select * from cxx_article_info where id > 3) AS `t1` LIMIT 1
[2017-07-19T15:10:00,174][INFO ][logstash.inputs.jdbc ](0.003000s) SELECT count(*) AS `count` FROM (select * from cxx_user_info where id > 2) AS `t1` LIMIT 1
[2017-07-19T15:11:00,225][INFO ][logstash.inputs.jdbc ](0.001000s) SELECT count(*) AS `count` FROM (select * from cxx_article_info where id > 3) AS `t1` LIMIT 1
[2017-07-19T15:11:00,225][INFO ][logstash.inputs.jdbc ](0.002000s) SELECT count(*) AS `count` FROM (select * from cxx_user_info where id > 2) AS `t1` LIMIT 1

4、扩展

1）多个表无非就是在input里面多加几个类型，在output中多加基础
类型判定。
举例：

if [type]=="cxx_user"  

2）input里的type和output if判定的type保持一致，该type对应ES中的type。

后记

死磕ES，有问题欢迎大家提问探讨！

作者：铭毅天下
转载请标明出处，原文地址：
http://blog.csdn.net/laoyang360/article/details/75452953