Amazon Cloud 上的 MySql 复制故障转移
我已经配置了 MySql 复制,其中 1 个主站、1 个备用站和 1 个从站。 Master 和 Standby 位于美国东部地区的不同可用区,Slave 位于美国西部地区。备用服务器和从属服务器都从主服务器复制。
应用层可以进行从Master到Standby的故障转移。用于在Master挂掉时将Slave的复制切换到Standby。我无法确保备用设备的数据比从设备的数据领先。如果备用设备的数据落后于从设备的数据,我需要扫描从设备的二进制和中继日志以找出丢失的内容并将其传输到备用设备。这就是理论。
有谁知道有什么工具可以处理这个问题吗?或者适合我的情况的替代解决方案? 最终目标是应用层的MySql Server 的可用性。我尝试过 MySql-Proxy 但不太满意。 Mysql Cluster 不是我们的选择。
提前致谢。
I have configured a MySql Replication with 1 Master 1 Standy and 1 Slave. Master and Standby are put on different availability zones on US East region and the Slave is on the US West region. Both Standby and Slave are replicating from the Master.
The Application layer can do the failover from Master to the Standby.For switching the replication of the Slave to the Standby when the Master dies. I have no way to assure that the Standby's data is ahead comparing to the Slave's data. If Standby's data is behind the Slave's data I need to scan the Slave's binary and relay logs to find out what is missing and transfer it the Standby. That's the THEORY.
Does anyone know any tool which can handle this ? or alternative solution for my context ?
The ultimate goal is the availability of MySql Server for application layer. I have tried MySql-Proxy but not quite happy with it. Mysql Cluster is not an option for us.
Thanks in advanced.
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。

绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
Kane,
看看 Schooner Information Technology(我的公司)的 SchoonerSQL。它旨在通过多线程同步复制(在当前版本中)在本地实现零从属滞后,并通过多线程异步 WAN 复制(将在不到 2 周内发布)跨区域自动故障转移。 http://www.schoonerinfotech.com/products/schoonersql
Kane,
Take a look at SchoonerSQL from Schooner Information Technology (my company). It is designed to have zero slave lag locally through multi-threaded synchronous replication (in current release) and automatic fail-over across regions through multi-threaded Asynchronous WAN replication (to be released in < 2 wks). http://www.schoonerinfotech.com/products/schoonersql