Re: How to modify the application to implement the separation of write/read

--__=_Part_Boundary_005_031188.027212
Content-Type: text/plain;
	charset="UTF-8"
Content-Transfer-Encoding: 7bit

Hi Mark,

I have two mysql servers, but I configure both of them are masters, and at the same both of them are also slaves, that is, I configure the dual-replication between them.

I want to know if I configure mysql servers like this and use the configuration "jdbc:mysql:replication" to access mysql, if one of the master is unavailable, can I get failover if I am trying to write to the master with the help of driver?

Do I have to modify the code of application to implement failover?

Thanks,
Peter




发件人: Mark Matthews <[email protected]>
发送时间: 2010-03-12 09:44
主 题: Re: How to modify the application to implement the separation of write/read
收件人: mqboss <[email protected]>
抄 送: java <[email protected]>






On Mar 11, 2010, at 7:30 PM, mqboss wrote: 

> Hi all,  
>  
> I want to test the mode of Master/Slave for mysql. 
> Also I want to separate the write and read of sql requests, that is, I want to make Slave server of mysql to handle read requests and make Master server to handle write requests. 
>  
> I want to know how to modify the code of my application. Some people suggest me to use mysql-proxy to access mysql server. 
> I checked the version of mysql-proxy, I found the latest version is just 0.8.0. 
> Does someone have met this problem? How to implement the separation of read and write?  
> Please give me some advice. 
>  
> I use java with the Connector/L jdbc driver, I checked the webpage of http://dev.mysql.com/doc/refman/5.1/en/connector-j-reference-replication-connection.html, it suggests to use "ReplicationDriver" to replace common driver. 
> I use Hibernate to access Mysql, the connection poll is c3p0-0.9.0, can I just need to modify the configuration of JDBC URL, like "jdbc:mysql:replication://master,slave1,slave2,slave3/test" to implement the separation of write/read?  
> After configuring JDBC URL like this, can I make the query SQL requests to access the Slave server, the update/insert/delete SQL requests to access the Master server, and if the Master server is shotdown, the driver will failover to the Slave server? 
> I am not sure whether do I need to modify something else? Like my application code. 

Peter, 

If you use "jdbc:mysql:replication", you still have to change your application to let the driver know whether it can use a slave or not. You do this by calling Connection.setReadOnly(true) to direct queries to the slave(s), and Connection.setReadOnly(false) to direct queries to the master. 

If the master is unavailable, with "jdbc:mysql:replication" you do not get failover if you are trying to write to the master, as it's a little bit more complicated than the driver can handle to promote one of the existing slaves to a master. 

    -Mark 
--  
Mark Matthews 
Principal Software Developer - Enterprise Tools 
Oracle 
http://www.mysql.com/products/enterprise/monitor.html 






--  
MySQL Java Mailing List 
For list archives: http://lists.mysql.com/java 
To unsubscribe:    http://lists.mysql.com/[email protected] 
--__=_Part_Boundary_005_031188.027212--