Codership will have booth E21 at OpenStack Summit 3-5 November in Paris. Come and meet Galera expert and talk your OpenStack Galera plans.
Galera presentation Wednesday 5th, 11.30 on Demotheater stage:
Designing for Scale: Data Storage in OpenStack and Galera Cluster
By Jay Pipes, Mirantis
If you’re going to build a house, you don’t start by finding a plot of quicksand to build on. A good house requires a good foundation. Same goes for your OpenStack deployments. The foundation of your OpenStack deployment is composed of the data stores you use for the various types of data that flow through the OpenStack componentry.
Before you deploy OpenStack, spend some time thinking about how to build on a good foundation. In this session, we will analyze the numerous OpenStack components with an eye for how the data each component produces and consumes is stored in memory and on disk. We will teach you how to think about your data — its size and shape, how frequently it is written or read, how aggressively it is cached, how the data naturally partitioned or grouped — and how to structure your data storage foundation to most effectively group like data together.
We will discuss traditional databases like MySQL and PostgreSQL, as well the synchronously-replicated MySQL Galera Cluster database. The tail end of the talk will focus specifically on optimizing MySQL Galera Cluster performance for varying workloads. Attendees should come away from the session with a great understanding of data architecture for OpenStack as well as a good set of best practices for tuning database storage for their OpenStack deployments.