本次實踐基於ubuntu系統;
mycat:1.6.5;
採用docker
拉起3個mysql
容器,端口分別位於33061,33062,33063。
java
sudo apt-get install openjdk-8-jdk-headless
mycat
將mycat
安裝包解壓到/usr/local
下:java
sudo chown -R $USER /usr/local/mycat
mycat
/usr/local/mycat/bin/mycat start
查看logs/wrapper.log
監控啓動狀態node
mycat
server.xml
中配置可訪問用戶:<user name="mycat"> <property name="password">mycat</property> <property name="schemas">db1</property> </user>
這裏的db1
必須是schema.xml
中配置的,不然報錯。mysql
schema.xml
中配置參數:<?xml version="1.0"?> <!DOCTYPE mycat:schema SYSTEM "schema.dtd"> <mycat:schema xmlns:mycat="http://io.mycat/"> // schema標籤的name對應server.xml中的schema <schema name="db1" checkSQLschema="true" sqlMaxLimit="1000000"> <table name="user" primaryKey="id" autoIncrement="true" dataNode="dn$1-3" rule="mod-long" /> </schema> <dataNode name="dn1" dataHost="node1" database="node" /> <dataNode name="dn4" dataHost="node1" database="node" /> <dataNode name="dn7" dataHost="node1" database="node" /> <dataNode name="dn11" dataHost="node1" database="node" /> <dataNode name="dn2" dataHost="node2" database="node" /> <dataNode name="dn5" dataHost="node2" database="node" /> <dataNode name="dn8" dataHost="node2" database="node" /> <dataNode name="dn11" dataHost="node2" database="node" /> <dataNode name="dn3" dataHost="node3" database="node" /> <dataNode name="dn6" dataHost="node3" database="node" /> <dataNode name="dn9" dataHost="node3" database="node" /> <dataNode name="dn12" dataHost="node3" database="node" /> // dataHost的name對應dataNode中的dataHost <dataHost name="node1" maxCon="1000" minCon="10" balance="1" writeType="0" dbType="mysql" dbDriver="native" switchType="1" slaveThreshold="100"> <heartbeat>select user()</heartbeat> // 在這裏配置docker拉起來的3個容器 <writeHost host="hostM1" url="192.168.1.5:33061" user="root" password="root"> </writeHost> </dataHost> <dataHost name="node2" maxCon="1000" minCon="10" balance="1" writeType="0" dbType="mysql" dbDriver="native" switchType="1" slaveThreshold="100"> <heartbeat>select user()</heartbeat> <writeHost host="hostM2" url="192.168.1.5:33062" user="root" password="root"> </writeHost> </dataHost> <dataHost name="node3" maxCon="1000" minCon="10" balance="1" writeType="0" dbType="mysql" dbDriver="native" switchType="1" slaveThreshold="100"> <heartbeat>select user()</heartbeat> <writeHost host="hostM3" url="192.168.1.5:33063" user="root" password="root"> </writeHost> </dataHost> </mycat:schema>
root@063b64a0619f:/# mysql -u mycat -p -P 8066 -h HOST
此處HOST
爲mycat
安裝所在的ip。sql
mysql> show databases; +----------+ | DATABASE | +----------+ | db1 | +----------+ 1 row in set (0.01 sec)
這裏的db1是schema標籤中對應的名稱,這是一個虛擬庫。docker
mysql> use db1; Reading table information for completion of table and column names You can turn off this feature to get a quicker startup with -A Database changed mysql> show tables; +----------------------------+ | Tables in datacache | +----------------------------+ | user | +----------------------------+ 5 rows in set (0.01 sec)
這裏的user表實際上目前也是個虛擬表,只有當在子節點中建立表以後,這個表纔有意義。數據庫
mysql> CREATE TABLE IF NOT EXISTS `user` ( -> `id` varchar(20) NOT NULL DEFAULT '0' COMMENT 'ID', -> `name` varchar(20) NOT NULL DEFAULT '' COMMENT '', -> `created_at` datetime NOT NULL DEFAULT now() COMMENT '', -> PRIMARY KEY (`id`) -> ) ENGINE=MyISAM DEFAULT CHARSET=utf8; Query OK, 0 rows affected, 1 warning (0.01 sec)
1.這裏的id
不採用int類型,是由於我採用的全局序列是默認的本地時間戳方式,int長度不夠;
2.這裏建立表結構,必須在對用的全部dataNode上建立相同的表,若是隻是在mycat庫中建立,只會在第一個dataNode中建立成功,並不能在全部dataNode中一塊兒建立,我想這是mycat仍然會進行優化的地方吧。ubuntu
mysql> INSERT INTO `user` (`name`,`date`) VALUES ('mycat','2017-10-10'); Query OK, 1 row affected (0.01 sec)
mysql> select * from user; +----+-------+---------------------+ | id | name | created_at | +----+-------+---------------------+ | 1 | mycat | 2018-02-01 07:12:26 | +----+-------+---------------------+ 1 row in set (0.01 sec)
1.當插入多條數據時,數據會根據mod-long
的分片方式分散到不一樣節點上;
2.若是以時間維度做爲篩選條件,會遍歷全部節點,因此,根據個人業務需求,我將分片方式改成sharding-by-month
。bash
本來是分佈在3個節點上的相同database上,可是發現當我插入一條數據,就會產生3條數據;
解決方式是分佈到不一樣的database上。
<schema name="db1" checkSQLschema="true" sqlMaxLimit="1000000"> <table name="user" primaryKey="id" autoIncrement="true" dataNode="dn$1-12" rule="sharding-by-month" /> </schema> <dataNode name="dn1" dataHost="node1" database="node1" /> <dataNode name="dn4" dataHost="node1" database="node2" /> <dataNode name="dn7" dataHost="node1" database="node3" /> <dataNode name="dn11" dataHost="node1" database="node4" /> <dataNode name="dn2" dataHost="node2" database="node1" /> <dataNode name="dn5" dataHost="node2" database="node2" /> <dataNode name="dn8" dataHost="node2" database="node3" /> <dataNode name="dn11" dataHost="node2" database="node4" /> <dataNode name="dn3" dataHost="node3" database="node1" /> <dataNode name="dn6" dataHost="node3" database="node2" /> <dataNode name="dn9" dataHost="node3" database="node3" /> <dataNode name="dn12" dataHost="node3" database="node4" />
這樣雖然須要建立12個database,數據準確性問題獲得解決。app