`
男人50
  • 浏览: 237475 次
  • 性别: Icon_minigender_1
  • 来自: 珠海
社区版块
存档分类
最新评论

利用MyCAT实现MySQL的读写分离和主从切换

 
阅读更多
实验环境介绍

采用的是在同一台机器上部署MyCAT,MySQL二主一从,其中3306和3307互为主从,3308为3306的从库。具体部署情况如下:

MyCAT:192.168.237.15,Port:8066/9066

MySQL Master:192.168.237.15,Port:3306

MySQL Master(Standby):192.168.237.15,Port:3307

MySQL Slave:192.168.237.15,Port:3308


导入测试数据
真实数据库上创建testdb1、testdb2、testdb3库下的travelrecord表。

    mysql> show create table travelrecord \G
    *************************** 1. row ***************************
           Table: travelrecord
    Create Table: CREATE TABLE `travelrecord` (
      `id` int(11) NOT NULL AUTO_INCREMENT,
      `org_code` varchar(20) NOT NULL,
      `test_name` varchar(20) DEFAULT NULL,
      PRIMARY KEY (`id`)
    ) ENGINE=InnoDB DEFAULT CHARSET=utf8
    1 row in set (0.00 sec)




***读写分离(负载均衡)***
1、首先设定MyCAT实例中的逻辑库名(schema name)为DB1。
      并定义这个逻辑库下存在一个逻辑表(table name)为travelrecord,该逻辑表所属的数据分片有:dataNode(dn1)
2、定义dataNode所在的dataHost、以及dataNode对应的真实database(testdb)。
      定义dataHost拥有的writeHost和readHost(3306为write host,3307为备用write host,3308为read host)
据此,在MyCAT的schema.xml文件中配置如下:

    <?xml version="1.0"?>
    <!DOCTYPE mycat:schema SYSTEM "schema.dtd">
    <mycat:schema xmlns:mycat="http://org.opencloudb/" >
            <!--定义MyCAT实例的逻辑库和逻辑表-->
            <schema name="DB1" checkSQLschema="false" sqlMaxLimit="100">
                    <table name="travelrecord" primaryKey="id" dataNode="dn1,dn2,dn3" />
            </schema>
            <!--定义MyCAT逻辑表中提到的所有dataNode所在的真实Host,以及dataNode所属的真实数据库-->
            <dataNode name="dn1" dataHost="237_15" database="testdb1" />
            <dataNode name="dn2" dataHost="237_15" database="testdb2" />
            <dataNode name="dn3" dataHost="237_15" database="testdb3" />
    
    <mycat:schema xmlns:mycat="http://org.opencloudb/" >
                    <table name="travelrecord" primaryKey="id" dataNode="dn1,dn2,dn3" />
            </schema>
            <!--定义MyCAT逻辑表中提到的所有dataNode所在的真实Host,以及dataNode所属的真实数据库-->
            <dataNode name="dn1" dataHost="237_15" database="testdb1" />
            <dataNode name="dn2" dataHost="237_15" database="testdb2" />
            <dataNode name="dn3" dataHost="237_15" database="testdb3" />
    
            <!--定义dataNode提到的dataHost的连接限制数、负载均衡取向,以及真实的读写地址(writeHost和readHost)-->
            <dataHost name="237_15" maxCon="1000" minCon="3" balance="1"
                    writeType="0" dbType="mysql" dbDriver="native" switchType="1"  slaveThreshold="100">
                    <heartbeat>select user()</heartbeat>
                    <writeHost host="MySQL_M1" url="192.168.237.15:3306" user="root" password="123456">
                    <readHost host="MySQL_S1" url="192.168.237.15:3308" user="root" password="123456" />
                    </writeHost>
                    <writeHost host="MySQL_M2" url="192.168.237.15:3307" user="root" password="123456">
                    <!--M2无对应的Slave库,此处留白 -->
                    </writeHost>
            </dataHost>
    </mycat:schema>


参数解读

schema.xml中的balance的取值决定了负载均衡对非事务内的读操作的处理:

balance="0":读请求仅发送到writeHost上(不开启读写分离)。

balance="1":读请求随机分发到当前writeHost对应的readHost和standby的writeHost上。

balance="2":读请求随机分发到当前dataHost内所有的writeHost和readHost上。

balance="3":读请求随机分发到当前writeHost对应的readHost上。


对于事务内的SQL默认走写节点;

以 /*balance*/ 开头,可以指定SQL使用特定负载均衡方案。例如在大环境开启读写分离的情况下,特定强一致性的SQL查询需求;

slaveThreshold:近似的主从延迟时间(秒)Seconds_Behind_Master < slaveThreshold ,读请求才会分发到该Slave,确保读到的数据相对较新。


schema.xml中的writeType的取值决定了负载均衡对写操作的处理:

writeType="0":所有的写操作都发送到配置文件中的第一个write host。(第一个write host故障切换到第二个后,即使之后修复了仍然维持第二个为写库)。推荐取0值,不建议修改.


分类演示

事先我们要打开MyCAT的debug模式。# vim log4j.xml,将info修改为debug

      &lt;root&gt;
        &lt;level value="<strong>debug</strong>" /&gt;
        &lt;appender-ref ref="FILE" /&gt;
         &lt;!--&lt;appender-ref ref="FILE" /&gt;--&gt;
      &lt;/root&gt;


下面通过变换配置文件中的balance值,执行写操作:select * from travelrecord来演示各种读写分离的情况。

(1)balance=“0”,查看debug日志,走MySQL_M1

    08/17 15:37:49.712 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:82) -execute mutinode query select * from travelrecord
    08/17 15:37:49.712 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:97) -has data merge logic
    08/17 15:37:49.713 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M1 for dataHost:237_15</strong>
    08/17 15:37:49.714 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M1 for dataHost:237_15</strong>
    08/17 15:37:49.715 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M1 for dataHost:237_15</strong>
    08/17 15:37:49.718 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (DataMergeService.java:102) -field metadata inf:[TEST_NAME=ColMeta [colIndex=2, colType=253], ORG_CODE=ColMeta [colIndex=1, colType=253], ID=ColMeta [colIndex=0, colType=3]]
    08/17 15:37:49.719 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:250) -on row end reseponse MySQLConnection [id=1, lastTime=1502955469708, user=root, schema=testdb2, old shema=testdb2, borrowed=true, fromSlaveDB=false, threadId=31, charset=utf8, txIsolation=3, autocommit=true, attachment=dn2{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5e5f215f, host=192.168.237.15, <strong>port=3306</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]


(2)balance=“1”,查看debug日志,走MySQL_M2或者MySQL_S1

    08/17 15:56:10.023  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:82) -execute mutinode query select * from travelrecord
    08/17 15:56:10.023  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:97) -has data merge logic
    08/17 15:56:10.023  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M2 for dataHost:237_15</strong>
    08/17 15:56:10.027  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M2 for dataHost:237_15</strong>
    08/17 15:56:10.028  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M2 for dataHost:237_15</strong>
    08/17 15:56:10.030  DEBUG [$_NIOREACTOR-0-RW] (DataMergeService.java:102) -field metadata inf:[TEST_NAME=ColMeta [colIndex=2, colType=253], ORG_CODE=ColMeta [colIndex=1, colType=253], ID=ColMeta [colIndex=0, colType=3]]
    08/17 15:56:10.031  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:250) -on row end reseponse MySQLConnection [id=8, lastTime=1502956570027, user=root, schema=testdb3, old shema=testdb3, borrowed=true, fromSlaveDB=false, threadId=17, charset=utf8, txIsolation=3, autocommit=true, attachment=dn3{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5209c6a4, host=192.168.237.15, <strong>port=3307</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]


    08/17 16:03:50.225  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:82) -execute mutinode query select * from travelrecord
    08/17 16:03:50.225  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:97) -has data merge logic
    08/17 16:03:50.225  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M2 for dataHost:237_15</strong>
    08/17 16:03:50.226  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_S1 for dataHost:237_15</strong>
    08/17 16:03:50.226  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M2 for dataHost:237_15</strong>
    08/17 16:03:50.227  DEBUG [$_NIOREACTOR-0-RW] (DataMergeService.java:102) -field metadata inf:[TEST_NAME=ColMeta [colIndex=2, colType=253], ORG_CODE=ColMeta [colIndex=1, colType=253], ID=ColMeta [colIndex=0, colType=3]]
    08/17 16:03:50.227  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:250) -on row end reseponse MySQLConnection [id=6, lastTime=1502957030209, user=root, schema=testdb2, old shema=testdb2, borrowed=true, fromSlaveDB=true, threadId=18, charset=utf8, txIsolation=3, autocommit=true, attachment=dn2{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5cd051b6, host=192.168.237.15, <strong>port=3308</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:03:50.228  DEBUG [$_NIOREACTOR-0-RW] (NonBlockingSession.java:229) -release connection MySQLConnection [id=6, lastTime=1502957030209, user=root, schema=testdb2, old shema=testdb2, borrowed=true, fromSlaveDB=true, threadId=18, charset=utf8, txIsolation=3, autocommit=true, attachment=dn2{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5cd051b6, host=192.168.237.15, <strong>port=3308</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:03:50.228  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDatasource.java:403) -release channel MySQLConnection [id=6, lastTime=1502957030209, user=root, schema=testdb2, old shema=testdb2, borrowed=true, fromSlaveDB=true, threadId=18, charset=utf8, txIsolation=3, autocommit=true, attachment=null, respHandler=null, host=192.168.237.15, port=3308, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:03:50.231  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:250) -on row end reseponse MySQLConnection [id=8, lastTime=1502957030209, user=root, schema=testdb3, old shema=testdb3, borrowed=true, fromSlaveDB=false, threadId=17, charset=utf8, txIsolation=3, autocommit=true, attachment=dn3{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5cd051b6, host=192.168.237.15, <strong>port=3307</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:03:50.232  DEBUG [$_NIOREACTOR-0-RW] (NonBlockingSession.java:229) -release connection MySQLConnection [id=8, lastTime=1502957030209, user=root, schema=testdb3, old shema=testdb3, borrowed=true, fromSlaveDB=false, threadId=17, charset=utf8, txIsolation=3, autocommit=true, attachment=dn3{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5cd051b6, host=192.168.237.15, <strong>port=3307</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:03:50.232  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDatasource.java:403) -release channel MySQLConnection [id=8, lastTime=1502957030209, user=root, schema=testdb3, old shema=testdb3, borrowed=true, fromSlaveDB=false, threadId=17, charset=utf8, txIsolation=3, autocommit=true, attachment=null, respHandler=null, host=192.168.237.15, port=3307, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:03:50.239  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:250) -on row end reseponse MySQLConnection [id=5, lastTime=1502957030209, user=root, schema=testdb1, old shema=testdb1, borrowed=true, fromSlaveDB=false, threadId=15, charset=utf8, txIsolation=3, autocommit=true, attachment=dn1{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5cd051b6, host=192.168.237.15, <strong>port=3307</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:03:50.239  DEBUG [$_NIOREACTOR-0-RW] (NonBlockingSession.java:229) -release connection MySQLConnection [id=5, lastTime=1502957030209, user=root, schema=testdb1, old shema=testdb1, borrowed=true, fromSlaveDB=false, threadId=15, charset=utf8, txIsolation=3, autocommit=true, attachment=dn1{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5cd051b6, host=192.168.237.15, <strong>port=3307</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]


(3)balance=“2”,查看debug日志,走MySQL_M1或者MySQL_S1

    08/17 16:10:48.867  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:82) -execute mutinode query select * from travelrecord
    08/17 16:10:48.867  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:97) -has data merge logic
    08/17 16:10:48.868  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_M1 for dataHost:237_15</strong>
    08/17 16:10:48.869  DEBUG [$_NIOREACTOR-0-RW] (MySQLConnection.java:459) -con need syn ,total syn cmd 1 commands SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;schema change:false con:MySQLConnection [id=1, lastTime=1502957448869, user=root, schema=testdb1, old shema=testdb1, borrowed=true, fromSlaveDB=false, threadId=38, charset=utf8, txIsolation=0, autocommit=true, attachment=dn1{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@17005763, host=192.168.237.15, <strong>port=3306</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    08/17 16:10:48.876  DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_S1 for dataHost:237_15</strong>
    08/17 16:10:48.876  DEBUG [$_NIOREACTOR-0-RW] (MySQLConnection.java:459) -con need syn ,total syn cmd 2 commands SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;schema change:true con:MySQLConnection [id=4, lastTime=1502957448876, user=root, schema=testdb2, old shema=testdb3, borrowed=true, fromSlaveDB=true, threadId=20, charset=utf8, txIsolation=0, autocommit=true, attachment=dn2{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@17005763, host=192.168.237.15, <strong>port=3308</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    此处之后日志省略...


(4)balance=“3”,查看debug日志,走MySQL_M1的从库:MySQL_S1

    08/17 16:15:54.267 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:82) -execute mutinode query select * from travelrecord
    08/17 16:15:54.267 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:97) -has data merge logic
    08/17 16:15:54.268 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_S1 for dataHost:237_15</strong>
    08/17 16:15:54.270 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_S1 for dataHost:237_15</strong>
    08/17 16:15:54.271 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (PhysicalDBPool.java:452) -<strong>select read source MySQL_S1 for dataHost:237_15</strong>
    08/17 16:15:54.274 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (DataMergeService.java:102) -field metadata inf:[TEST_NAME=ColMeta [colIndex=2, colType=253], ORG_CODE=ColMeta [colIndex=1, colType=253], ID=ColMeta [colIndex=0, colType=3]]
    08/17 16:15:54.276 &nbsp;DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:250) -on row end reseponse MySQLConnection [id=4, lastTime=1502957754253, user=root, schema=testdb1, old shema=testdb1, borrowed=true, fromSlaveDB=true, threadId=21, charset=utf8, txIsolation=3, autocommit=true, attachment=dn1{SELECT *
    FROM travelrecord
    LIMIT 100}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@5615b9cb, host=192.168.237.15, <strong>port=3308</strong>, statusSync=null, writeQueue=0, modifiedSQLExecuted=false]
    此处之后日志省略...



MyCAT插入测试数据:mysql> insert into travelrecord values(1,"china86","mike");

查看debug日志,写操作走MySQL_M1

    08/17 16:38:49.995  DEBUG [$_NIOREACTOR-0-RW] (ServerQueryHandler.java:56) -ServerConnection [id=1, schema=DB1, host=127.0.0.1, user=test,txIsolation=3, autocommit=true, schema=DB1]insert into travelrecord values(1,"china86","mike")
    08/17 16:38:50.002  DEBUG [$_NIOREACTOR-0-RW] (NonBlockingSession.java:113) -ServerConnection [id=1, schema=DB1, host=127.0.0.1, user=test,txIsolation=3, autocommit=true, schema=DB1]insert into travelrecord values(1,"china86","mike"), route={
       1 -&gt; dn1{insert into travelrecord values(1,"china86","mike")}
       2 -&gt; dn2{insert into travelrecord values(1,"china86","mike")}
       3 -&gt; dn3{insert into travelrecord values(1,"china86","mike")}
    } rrs
    08/17 16:38:50.003  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:82) -execute mutinode query insert into travelrecord values(1,"china86","mike")
    08/17 16:38:50.003  DEBUG [$_NIOREACTOR-0-RW] (MySQLConnection.java:459) -con need syn ,total syn cmd 1 commands SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;schema change:false con:MySQLConnection [id=2, lastTime=1502959130003, user=root, schema=testdb1, old shema=testdb1, borrowed=true, fromSlaveDB=false, threadId=41, charset=utf8, txIsolation=0, autocommit=true, attachment=dn1{insert into travelrecord values(1,"china86","mike")}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@6b4de96c, host=192.168.237.15, port=3306, statusSync=null, writeQueue=0, modifiedSQLExecuted=true]
    08/17 16:38:50.006  DEBUG [$_NIOREACTOR-0-RW] (MySQLConnection.java:459) -con need syn ,total syn cmd 1 commands SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;schema change:false con:MySQLConnection [id=3, lastTime=1502959130006, user=root, schema=testdb2, old shema=testdb2, borrowed=true, fromSlaveDB=false, threadId=40, charset=utf8, txIsolation=0, autocommit=true, attachment=dn2{insert into travelrecord values(1,"china86","mike")}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@6b4de96c, host=192.168.237.15, port=3306, statusSync=null, writeQueue=0, modifiedSQLExecuted=true]



***主从切换(双主failover)***

如果我们细心观察schem.xml文件的话,会发现之前有一个参数:switchType尚未提及。

<dataHost name="237_15" maxCon="1000" minCon="10" balance="1" writeType="0" dbType="mysql" dbDriver="native"switchType="1"  slaveThreshold="100">


参数解读

switchType="-1":不自动切换

switchType="1":自动切换

switchType="2":基于MySQL主从复制的状态来决定是否切换。需修改heartbeat语句:show slave status

switchType="3":基于Galera(集群多节点复制)的切换机制。需修改heartbeat语句:show status like 'wsrep%'


switchType="2"的切换演示

先修改heartbeat语句

            &lt;dataHost name="237_15" maxCon="1000" minCon="3" balance="1"
                    writeType="0" dbType="mysql" dbDriver="native" switchType="2"  slaveThreshold="100"&gt;
                    &lt;heartbeat&gt;show slave status&lt;/heartbeat&gt;


为了区分是主库还是从库插入的数据,M1、M2设置不同自增步长,通过id的奇偶性来判断,避免通过查debug日志的方式来区分。

MySQL_M1:auto_increment_increment=2,auto_increment_offset=1

MySQL_M2:auto_increment_increment=2,auto_increment_offset=2


(1)M1、M2两个节点均正常的情况下,设想情况:MyCAT选择M1插入数据

插入两组测试数据:

    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","tom");
    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","jack");
    mysql&gt; select *from travelrecord;
    +----+----------+-----------+
    | id | org_code | test_name |
    +----+----------+-----------+
    |  1 | china86  | tom       |
    |  3 | china86  | jack      |
    +----+----------+-----------+

id为奇数,故写入的数据库是M1,符合预期

(2)M1宕机,设想情况:MyCAT切换至M2作为写入节点
此时,我们模拟故障,手动停止3306上的MySQL实例
[root@237_15 conf]# /etc/init.d/mysql3306 stop
Shutting down MySQL............ SUCCESS!

再往MyCAT内插入两组数据:

    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","marry");
    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","ellen");
    mysql&gt; select *from travelrecord;
    +----+----------+-----------+
    | id | org_code | test_name |
    +----+----------+-----------+
    |  1 | china86  | tom       |
    |  3 | china86  | jack      |
    |  4 | china86  | marry     |
    |  6 | china86  | ellen     |
    +----+----------+-----------+

id为偶数,故写入的是M2,符合预期

再截取一段debug日志,确认写入的实例端口为3307(MySQL_M2)

08/17 22:10:40.330  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:171) -received ok response ,executeResponse:true from MySQLConnection [id=7, lastTime=1502979040315, user=root, schema=testdb2, old shema=testdb2, borrowed=true, fromSlaveDB=false, threadId=5, charset=utf8, txIsolation=3, autocommit=true, attachment=dn2{insert into travelrecord(org_code,test_name) values("china86","ellen")}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@7450171c, host=192.168.237.15, port=3307, statusSync=null, writeQueue=0, modifiedSQLExecuted=true]


(3)M1恢复正常,设想情况:MyCAT继续使用之前的M2作为写入节点
重启3306上的实例(MySQL_M1),再往MyCAT内插入两条数据:

    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","tiger");
    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","rabbit");
    mysql&gt; select *from travelrecord;
    +----+----------+-----------+
    | id | org_code | test_name |
    +----+----------+-----------+
    |  1 | china86  | tom       |
    |  3 | china86  | jack      |
    |  4 | china86  | marry     |
    |  6 | china86  | ellen     |
    |  8 | china86  | tiger     |
    | 10 | china86  | rabbit    |
    +----+----------+-----------+

id为偶数,故虽然M1已重启,但写入的数据库仍是M2,符合预期
再截取一段debug日志进行确认,写入的实例端口仍为3307(MySQL_M2)

    08/17 22:17:20.549  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:171) -received ok response ,executeResponse:true from MySQLConnection [id=5, lastTime=1502979440533, user=root, schema=testdb1, old shema=testdb1, borrowed=true, fromSlaveDB=false, threadId=4, charset=utf8, txIsolation=3, autocommit=true, attachment=dn1{insert into travelrecord(org_code,test_name) values("china86","tiger")}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@7cacca01, host=192.168.237.15, port=3307, statusSync=null, writeQueue=0, modifiedSQLExecuted=true]
    ...
    08/17 22:17:31.698  DEBUG [$_NIOREACTOR-0-RW] (MultiNodeQueryHandler.java:171) -received ok response ,executeResponse:true from MySQLConnection [id=5, lastTime=1502979451682, user=root, schema=testdb1, old shema=testdb1, borrowed=true, fromSlaveDB=false, threadId=4, charset=utf8, txIsolation=3, autocommit=true, attachment=dn1{insert into travelrecord(org_code,test_name) values("china86","rabbit")}, respHandler=org.opencloudb.mysql.nio.handler.MultiNodeQueryHandler@73118f1d, host=192.168.237.15, port=3307, statusSync=null, writeQueue=0, modifiedSQLExecuted=true]


(4)M1从故障中恢复,M2可用,重启MyCAT,设想情况:MyCAT还是会继续沿用M2作为写入节点

重启MyCAT,再插入两条数据,写入的数据库还是M2,符合预期

    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","orange");
    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","apple");
    mysql&gt; select *from travelrecord;
    +----+----------+-----------+
    | id | org_code | test_name |
    +----+----------+-----------+
    |  1 | china86  | tom       |
    |  3 | china86  | jack      |
    |  4 | china86  | marry     |
    |  6 | china86  | ellen     |
    |  8 | china86  | tiger     |
    | 10 | china86  | rabbit    |
    | 12 | china86  | orange    |
    | 14 | china86  | apple     |
    +----+----------+-----------+


(5)修改writeHost的index信息,指定写入节点为M1,设想情况:MyCAT将写入节点切换至M1
打开conf目录下的dnindex.properties文件

    #update
    #Thu Aug 17 22:08:16 CST 2017
    237_15=1

可以看到当前所用的writeHost的index为1(0表示使用的是配置文件schema.xml中第一个writeHost,1代表了第二个,以此类推)

我们尝试修改index为0,并重启MyCAT

再插入两组数据,发现id已经变为奇数,说明写入的数据库重新划给MySQL_M1了,符合预期

    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","black");
    mysql&gt; insert into travelrecord(org_code,test_name) values("china86","white");
    mysql&gt; select *from travelrecord;
    +----+----------+-----------+
    | id | org_code | test_name |
    +----+----------+-----------+
    |  1 | china86  | tom       |
    |  3 | china86  | jack      |
    |  4 | china86  | marry     |
    |  6 | china86  | ellen     |
    |  8 | china86  | tiger     |
    | 10 | china86  | rabbit    |
    | 12 | china86  | orange    |
    | 14 | china86  | apple     |
    | 15 | china86  | black     |
    | 17 | china86  | white     |
    +----+----------+-----------+


注:可以通过管理端口连接MyCAT,执行show @@heartbeat,若 RS_CODE=-1 表示该节点心跳检测出错。



https://blog.csdn.net/leonpenn/article/details/77278360
分享到:
评论

相关推荐

    基于Mycat实现Mysql读写分离以及分库分表.doc

    基于Mycat实现Mysql读写分离以及分库分表详解 本文档详细介绍了基于Mycat实现Mysql读写分离以及分库分表的技术,涵盖了Mycat安装、配置、读写分离、分库分表等多方面的知识点。 一、读写分离 Mycat读写分离是指将...

    MyCat实现mysql读写分离配置说明

    使用MyCat实现mysql读写分离配置说明

    mysql与mycat的读写分离

    Mycat是一款开源的数据库中间件,它可以实现对多个MySQL实例的连接管理和负载均衡,支持SQL路由等功能,进而实现读写分离。 ##### 启用Mycat读写分离 完成MySQL主从配置后,下一步是启用Mycat的读写分离机制。这...

    实战案例:利用Mycat实现MySQL的读写分离.md

    实战案例:利用Mycat实现MySQL的读写分离.md

    基于mycat的mysql高可用读写分离

    基于mycat的mysql高可用读写分离,适合mysql进阶。比较实用的技术。

    springboot实现mysql的读写分离的案例源码

    一般来说,读写分离有两种实现方式。第一种是依靠中间件MyCat,也就是说应用程序连接到中间件,中间件帮我们做SQL分离,去选择指定的数据源;第二种是应用程序自己去做分离。这里我用程序自己来做,主要是利用Spring...

    mycat+mysql读写分离集群离线环境搭建

    本篇文章将详细讲解如何在离线环境下,利用Mycat和MySQL搭建一个读写分离的集群。Mycat是一款开源的分布式数据库中间件,它能够实现数据库的水平扩展,而MySQL则是广泛使用的开源关系型数据库。 首先,我们来了解...

    第三版_基于Mycat的MySQL主从读写分离配置详解与示例

    三、Mycat实现读写分离 1. 安装Mycat:首先,你需要在服务器上下载并安装Mycat,配置相关环境变量,确保Mycat能够正常启动。 2. 配置Mycat Server.xml:这是Mycat的核心配置文件,包括服务器端口、数据源、路由策略...

    基于Mycat的MySQL主从读写分离配置详解与示例

    本文将详细介绍如何利用Mycat这一开源数据库中间件实现MySQL数据库的主从读写分离,并通过具体的配置步骤及示例进行说明。 #### 二、Mycat简介 Mycat是一款开源的数据库连接池产品,它位于Java应用程序和数据库...

    mycat实现读写分离

    本文将深入探讨MyCat如何实现读写分离,并结合提供的“mycat实现读写分离.pdf”文件,详细解析这一过程。 一、MyCat简介 MyCat(原名Mycat-S)是一个基于MySQL协议的数据库中间件,它将多个数据库节点透明地连接...

    【27】基于MyCat的MySQL高可用读写分离集群实战课程下载【No215】基于MyCat的MySQL高可用读写分离集群实战课程下载 .txt

    此外,通过MyCat还能实现读写分离、数据分片等高级功能,从而进一步提升系统的可扩展性和可用性。 #### 二、MySQL高可用性与读写分离 1. **高可用性**:在IT领域中,高可用性通常指系统能够持续提供服务的能力。...

    基于数据库中间件Mycat的MySQL主从与读写分离配置详解与示例

    通过以上分析,我们可以看到,利用Mycat实现MySQL的主从复制和读写分离是一项涉及多方面技术细节的任务。正确理解和处理如【心跳问题heartbeat bug #393】和【bug407:修复主从状态监控和读写分离】这类问题,对于...

    linux Mysql mycat主从复制读写分离部署完成

    linux Mysql mycat主从复制读写分离部署完成 技术:mycat mysql集群 linux mycat读写分离 说明包含: mysql数据库服务安装包 mysql+mycat主从复制读写分离部署帮助文档 mysql+mycat主从复制读写分离使用帮助...

    oracle读写分离(中间键Mycat)

    在本配置文档中,我们将关注如何利用Mycat中间键实现Oracle 11g的读写分离,并在Linux环境下进行相关组件的安装。 Mycat是一个开源的数据库中间件,它具有分库分表、读写分离、数据切片等功能,能够帮助我们构建大...

    MYSQL+MYCAT 读写分离

    在本文中,我们将详细介绍如何使用MYSQL和MYCAT实现读写分离架构。 MYSQL主从复制 MYSQL主从复制是指将一个MYSQL数据库服务器作为主服务器,将其它服务器作为从服务器,主服务器上的所有操作都会被复制到从服务器...

    Mycat实现mysql主从读写分离的配置文件

    这是Mycat实现mysql主从读写分离时用到的的配置文件。 内容包括:schema.xml和server.xml。 请根据实际情况替换里面的数据库地址、用户、密码。 相关操作教程:...

    基于mycat中间件mysql5.7读写分离

    本文将详细介绍如何利用mycat中间件实现MySQL5.7的读写分离。 一、读写分离与mycat简介 1.1 读写分离 读写分离是数据库架构设计中的一个重要策略,它将读取操作和写入操作分开,通常主数据库负责写操作,而从...

Global site tag (gtag.js) - Google Analytics