Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I have 5 physical servers with low end memory & cpu resources. I want to create 1 cluster using all these servers and want to run mysql db on the same such that mysql db would utilize 5 server's CPU power to execute db queries & same for memory. Could you please help me understanding how to achieve this?

Regards,

share|improve this question

2 Answers 2

Depends on what you are doing. If you need availability set up your hosts in master-slave combinations to allow for maximum availability using MySQL's built-in master-slave replication. If you need high performance I would recommend MySQL cluster. Depending on your application, scalability should also be a factor when deciding which method to choose.

Also another great blog to watch put on by the guy's who wrote "High Performance MySQL". Really smart crew and great readings:

http://www.mysqlperformanceblog.com/2010/01/16/how-expensive-is-mysql-replication-for-the-master/

share|improve this answer
    
Thanks for the response. I am looking for performance than availability. So I need to use all 5 servers simultaneously. I read some documents for "eucalyptus" but didn't get exact clues. –  Vijay Gharge May 5 '10 at 9:48

You won't be able to simply & efficiently use all 5 servers CPUs and memory, but it is possible to cluster MySQL.

See here for a guide: http://www.howtoforge.com/loadbalanced_mysql_cluster_debian

share|improve this answer
    
Hi Chris, Thanks for feedback. Mysql is just an example. I wanted to efficiently use all 5 servers. Similar way google use their low end infrastructure. Please suggest –  Vijay Gharge May 5 '10 at 11:22
    
Google run an entire custom stack specifically tailored to the functions they perform. There's no 'one size fits all' solution for clustering. Just specific methods for specific usage scenarios. –  Chris Thorpe May 5 '10 at 21:01

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.