When to use MySQL replication or DRBD for HA on Xen VM?

Posted by user62513 on Server Fault See other posts from Server Fault or by user62513
Published on 2011-01-06T22:44:41Z Indexed on 2011/01/06 22:55 UTC
Read the original article Hit count: 219

Filed under:

high-availability

|

mysql-replication

|

drbd

I'm setting up a database which needs to be needs to provide High Availabilty.

My primary concern is high performance and robustness (I don't want something that will fail fast and badly). The database is accessed by the application at an average of 300 qps. It's will run on Xen VMs and it has some InnoDB tables as well as MyISAM tables. The VMs are connected via ethernet 100Mbit/s ethernet cables.

Which of the two - MySQL replication or DRBD - would you recommend in such a situation?

Or should I use DRBD to make the master database Highly Available and use MySQL replication on the slaves?

I'm a developer so these things are all not so easy for me to make a sound judgement.

© Server Fault or respective owner

Related posts about high-availability

Berkeley DB Java Edition High Availability/Replication Whitepaper

as seen on Oracle Blogs - Search for 'Oracle Blogs'
My colleague Sam Haradhvala has written a great whitepaper on BDB JE HA/Replication. It takes a high level look at what JE/HA is, when you might use it, and the key architectural tradeoffs you might consider. >>> More
High Availability - "Keeping Active"

as seen on Oracle Technology Network - Search for 'Oracle Technology Network'
Use Oracle Active Data Guard to set up a physical standby database for reporting and recovery. >>> More
Stratus: Where High-Availability Leads to Profitability

as seen on Internet.com - Search for 'Internet.com'
Server Snapshot: Stratus' fault-tolerant servers are designed to fill a clearly defined niche. Its positioning as the "availability company" has set it apart from the OEM Rat Pack and resulted in steady growth. >>> More
Stratus: Where High-Availability Leads to Profitability

as seen on Internet.com - Search for 'Internet.com'
Server Snapshot: Stratus' fault-tolerant servers are designed to fill a clearly defined niche. Its positioning as the "availability company" has set it apart from the OEM Rat Pack and resulted in steady growth. >>> More
Google Apps security and high availability support EZasset success

as seen on Google Code - Search for 'Google Code'
Editor's Note: This post was written by Paul Marsolan from StrongTech , an asset management company, and makers of EZasset. We invited StrongTech to share their experiences building... >>> More

Related posts about mysql-replication

MySQL replication Slave_IO_Running: No

as seen on Server Fault - Search for 'Server Fault'
Hi all, I have two servers that I am trying to get replication of one database between. I found a setup guide on sourceforge that I followed and I have tried various other settings since then, but no matter what I do, when I start the slave, the 'Slave_IO_Running' setting is always No.... I have… >>> More
MySql Replication with a star topology

as seen on Server Fault - Search for 'Server Fault'
My company currently operates in 3 separate locations connected by slow vpn links. Each site hosts a dedicated MySql server. I need to aggregate the data from all three of them onto a single server for corporate reporting. The powers that be have stated I cannot use circular replication or federated… >>> More
Fixing MySQL Replication

as seen on Internet.com - Search for 'Internet.com'
Continuing last months article on "MySQL Replication Pitfalls," Sean Hull discusses what to do to make your replication setup more resilient. >>> More
Advanced MySQL Replication - Improving Performance

as seen on Internet.com - Search for 'Internet.com'
MySQL Replication can be made quite reliable and robust if the right tools are used to keep it running smoothly--but what if enormous loads on the primary server are overloading the slave server. Are there ways to speed up performance, so the slave can keep up? >>> More
MySQL Replication Error

as seen on Server Fault - Search for 'Server Fault'
I recently updated my master server to 5.1.41 and noticed that the slave was no longer replicating. It was returning this erorr: 091208 12:53:31 [ERROR] Slave I/O: error connecting to master '[email protected]:3306' - retry-time: 10 retries: 86400, Error_code: 2026 091208 12:53:41 [ERROR] Slave I/O:… >>> More