OpenStack uses a database to store data for all OpenStack services such as compute nodes available, running instances, state and so on. The DB takes part in every user request, from listing all instances to creating new VMs, so it is central to the operation of OpenStack. In a high availability environment, the database should be clustered so as to avoid a single point of failure (SPOF).
In this post, we will setup up high availability clustering for the MySQL backend in OpenStack using Galera Cluster and ClusterControl. You will learn how to set up a multi-master MySQL setup, fronted by a redundant HAProxy load-balancer.
Our end goal is to achieve resilience to database failures, and allow OpenStack components to communicate with the database nodes via one Virtual IP address.