Exploring HA strategies

A robust OpenStack cloud platform includes fault tolerance at every level of its architecture. This can be very successful if it is planned in advance. Starting with a small cluster is easy and achievable, but growing it is a challenge. The hallmark of the basic OpenStack component is that it can run on commodity hardware. OpenStack is designed to scale massively and provide HA by leveraging more advanced HA techniques at each level of the infrastructure. This can include automatic failover and geo-redundancy. Chapter 4, OpenStack Compute – Compute Capacity and Flavors, introduced the concepts of cells, regions, and availability zones, which provide more robust and advanced fault tolerance and availability capabilities for massive OpenStack deployment at scale.

Measuring HA

Service availability should be measured and defined by standard metrics. This can be summarized using the following formula:

Availability = MTTF/ (MTTF + MTTR)

In the preceding...