Summary

In this chapter, we enabled a critical architectural safeguard design in our OpenStack environment by covering the HA pillar across the control and data planes. We now have numerous ways to construct a highly available OpenStack environment, depending on which HA strategy is preferred. In this chapter, we empowered the control plane services using HAProxy and Keepalived. Other deployments could employ Corosync, Pacemaker, and a selection of vendor solutions for load balancing. As demonstrated in this chapter, more than a single design pattern can be applied for common infrastructure services, such as databases and message queues, to achieve failover at best. Networking HA in OpenStack has achieved another level of maturity that enables cloud operators to choose between two options to minimize any potential risks of connectivity loss for the tenant networks: routing with VRRP and DVR.

This chapter also explored an attractive capability for cloud users to achieve failover...