Managing instance failover

Providing HA to cloud resources is a critical topic that OpenStackers have had to handle since the early releases of OpenStack. Workload users seek different ways to increase the availability of their virtual machines through manual tooling and scripting, which can be overlooked if not efficiently managed. With the introduction of the Masakari project in OpenStack, cloud operators can offer workload users an automated service that ensures HA for KVM-based instances, reducing the need for manual scripting and seamlessly integrating within the OpenStack ecosystem.

Masakari also uses Corosync and Pacemaker. As a native HA load balancing stack solution for the Linux platform, Pacemaker is a cluster resource manager that depends on Corosync to control and organize HA across the hosts in OpenStack. Corosync ensures that cluster communication is based on the messaging layer and manages the VIP assignment to one of the nodes. Once a cluster of workload instances...