Book Image

High Availability MySQL Cookbook

By : Alexander Davies
Book Image

High Availability MySQL Cookbook

By: Alexander Davies

Overview of this book

High Availability is something that all web sites hope to achieve, especially those that are linked to big companies.MySQL, an open source relational database management system (RDBMS), can be made highly available to protect from corruption, hardware failure, software crashes, and user error. Running a MySQL setup is quite simple. Things start getting complex when you start thinking about the best way to provide redundancy. There are a large number of techniques available to add 'redundancy' and 'high availability' to MySQL, but most are both poorly understood and documented.This book will provide you with recipes showing how to design, implement, and manage a MySQL Cluster and achieve high availability using MySQL replication, block level replication, shared storage, and the open source Global File System (GFS).This book covers all the major techniques available for increasing availability of your MySQL databases. It demonstrates how to design, implement, troubleshoot and manage a highly available MySQL setup using any one of several techniques, which are shown in different recipes. It is based on MySQL Cluster 7.0, MySQL (for non clustered recipes) 5.0.77, and CentOS / RedHat Enterprise Linux 5.3.The book starts by introducing MySQL Cluster as a technology and explaining how to set up a simple cluster. It will help you to master the options available for backing up and restoring a file in the MySQL Cluster. By following the practical examples in this book, you will learn how to manage the MySQL Cluster. Further, we will discuss some troubleshooting aspects of the MySQL Cluster.We also have a look at achieving high availability for MySQL databases with the techniques of MySQL Replication, block level replication, shared storage (a SAN or NAS), and DRBD.Finally, you will learn the principles of Performance tuning and tune MySQL database for optimal performance.
Table of Contents (17 chapters)
High Availability MySQL Cookbook
Credits
About the Author
About the Reviewers
Preface
Base Installation
Index

Multiple storage node failures


MySQL Clusters are designed to survive node failures, regardless of exactly how they occur. In the case of multiple node failures, working out what can happen can be a little more complicated. There are several options for multiple node failures as follows:

  • One node can fail in multiple nodegroups, but one node remains per nodegroup (cluster will remain working).

  • All the nodes in a single nodegroup can fail (in such a case, the cluster will be shut down).

  • One node per nodegroup can split into one group. Another node can split into another group, which are then partitioned from each other. This occurs most often when some nodes are connected to one network switch, the others are connected to a different switch, and the connections between the switches fail. This can cause a split-brain problem, requiring an arbitrator to shut down some nodes to ensure only one group is left alive.

Getting ready

In this recipe, we will cover how MySQL Cluster handles the failure...