同事 Struct Chen 的 Amoeba For MySQL 项目已经 出来不短时间了，可一直没有时间好好体验，最近一直在做一些软件和硬件的功能以及性能相关的测试工作，Amoeba For MySQL自然也将被列入计划之内。
对 Amoeba For MySQL 的测试到目前位置主要还是功能方面的测试，配置了一套四台机器的环境：
Server A： Amoeba Server
Server B： MySQL Master
Server C： MySQL Slave1
Server D： MySQL Slave2
软件环境：RHEL5.2 （X86_64），JAVA 1.5.0_16， Amoeba For MySQL 0.31
测试功能：
1、读写分离
2、数据垂直切分
3、数据水平切分
amoeba.xml配置如下：
< ?xml version="1.0" encoding="gbk"?>
< !DOCTYPE amoeba:configuration SYSTEM "amoeba.dtd">
<amoeba :configuration xmlns:amoeba="http://amoeba.meidusa.com/">
<server>
<property name="port">8066property>
<property name="ipAddress">10.0.38.21property>
<property name="readThreadPoolSize">500property>
<property name="clientSideThreadPoolSize">500property>
<property name="serverSideThreadPoolSize">500property>
<property name="netBufferSize">200property>
<property name="tcpNoDelay">trueproperty>
<property name="user">userproperty>
<property name="password">passwordproperty>
server>
<connectionmanagerlist>
<connectionmanager name="defaultManager">
<classname>com.meidusa.amoeba.net.AuthingableConnectionManagerclassname>
connectionmanager>
connectionmanagerlist>
<dbserverlist>
<dbserver name="master">
<factoryconfig>
<classname>com.meidusa.amoeba.mysql.net.MysqlServerConnectionFactoryclassname>
<property name="manager">defaultManagerproperty>
<property name="port">3306property>
<property name="ipAddress">10.0.38.37property>
<property name="schema">testproperty>
<property name="user">rootproperty>
<property name="password">passwordproperty>
factoryconfig>
<poolconfig>
<classname>com.meidusa.amoeba.net.poolable.PoolableObjectPoolclassname>
<property name="maxActive">500property>
<property name="maxIdle">200property>
<property name="minIdle">10property>
<property name="minEvictableIdleTimeMillis">600000property>
<property name="timeBetweenEvictionRunsMillis">600000property>
<property name="testOnBorrow">trueproperty>
<property name="testWhileIdle">trueproperty>
poolconfig>
dbserver>
<dbserver name="slave1">
<factoryconfig>
<classname>com.meidusa.amoeba.mysql.net.MysqlServerConnectionFactoryclassname>
<property name="manager">defaultManagerproperty>
<property name="port">3306property>
<property name="ipAddress">10.0.38.22property>
<property name="schema">testproperty>
<property name="user">rootproperty>
<property name="password">passwordproperty>
factoryconfig>
<poolconfig>
<classname>com.meidusa.amoeba.net.poolable.PoolableObjectPoolclassname>
<property name="maxActive">500property>
<property name="maxIdle">200property>
<property name="minIdle">10property>
<property name="minEvictableIdleTimeMillis">600000property>
<property name="timeBetweenEvictionRunsMillis">600000property>
<property name="testOnBorrow">trueproperty>
<property name="testWhileIdle">trueproperty>
poolconfig>
dbserver>
<dbserver name="slave2">
<factoryconfig>
... ...
factoryconfig>
<poolconfig>
... ...
poolconfig>
dbserver>
<dbserver name="virtualSlave" virtual="true">
<poolconfig>
<classname>com.meidusa.amoeba.server.MultipleServerPoolclassname>
<property name="loadbalance">1property>
<property name="poolNames">slave1,slave2property>
poolconfig>
dbserver>
dbserverlist>
<queryrouter>
<classname>com.meidusa.amoeba.mysql.parser.MysqlQueryRouterclassname>
<property name="ruleConfig">${amoeba.home}/conf/rule.xmlproperty>
<property name="functionConfig">${amoeba.home}/conf/functionMap.xmlproperty>
<property name="ruleFunctionConfig">${amoeba.home}/conf/ruleFunctionMap.xmlproperty>
<property name="LRUMapSize">1500property>
<property name="defaultPool">masterproperty>
<property name="writePool">masterproperty>
<property name="readPool">virtualSlaveproperty>
<property name="needParse">trueproperty>
queryrouter>
amoeba>
rule.xml配置：
< ?xml version="1.0" encoding="gbk"?>
< !DOCTYPE amoeba:rule SYSTEM "rule.dtd">
<amoeba :rule xmlns:amoeba="http://amoeba.meidusa.com/">
<tablerule name="test_horiz" schema="test" defaultPools="master">
<rule name="rule1">
<parameters>IDparameters>
<expression>< ![CDATA[ ID <= 100000000]]>expression>
<defaultpools>masterdefaultpools>
<readpools>slave1readpools>
<writepools>masterwritepools>
rule>
<rule name="rule2">
<parameters>IDparameters>
<expression>< ![CDATA[ ID between 100000001 and 200000000 ]]>expression>
<defaultpools>masterdefaultpools>
<writepools>master2writepools>
<readpools>slave2readpools>
rule>
<rule name="rule3">
<parameters>IDparameters>
<expression>< ![CDATA[ ID > 200000000 ]]>expression>
<defaultpools>masterdefaultpools>
rule>
tablerule>
<tablerule name="master" schema="test" defaultPools="master"/>
<tablerule name="slave1" schema="test" defaultPools="slave1"/>
<tablerule name="slave2" schema="test" defaultPools="slave2"/>
<tablerule name="master" schema="t" defaultPools="master"/>
amoeba>
由于第一步主要还只是测试读写分离与数据的垂直切分，所以水平切分的规则设置的很简单，仅仅是按照ID的范围设置简单的规则。而设置一些更为复杂的规则。
总体测试效果还是比较满意的，预期的功能都全部正常，当然所用于测试的SQL语句也大都比较简单，主要还是根据当前工作中遇到的一些SQL。
这次测试没有进行性能测试，等后面再找时间测试一下，不知道有没有哪位朋友已经做过了相应的性能测试没？性能测试之后就要开始规划应用到某些应用上面去了