COA High Availability & Fault Tolerance Flashcards

What is the recommended Glance image storage backend for eliminating single points of failure in HA deployments?

Answer: Swift or Ceph RBD

Swift or Ceph RBD provide distributed, replicated storage for Glance images, eliminating single points of failure in image storage.

What does a Ceph replication factor of 3 mean for OpenStack volume storage?

Answer: Each data object is stored on three separate OSDs

A replication factor of 3 means Ceph writes each data object to three different OSDs, tolerating up to two simultaneous OSD failures without data loss.

Which Nova scheduler feature ensures that related instances are placed on different physical hosts to avoid correlated failures?

Answer: Anti-affinity server groups

Anti-affinity server groups instruct Nova's scheduler to place group members on different physical compute hosts, preventing all instances from failing together.

What is the key difference between `nova evacuate` and `nova host-evacuate`?

Answer: host-evacuate moves all instances from a specified host at once

`nova host-evacuate` bulk-evacuates all instances from a failed host simultaneously, while `nova evacuate` handles a single instance at a time.

Which library provides distributed coordination and lock management for OpenStack services to prevent split-brain?

Answer: Tooz with ZooKeeper or etcd backend

Tooz provides a distributed coordination library that OpenStack services use with backends like ZooKeeper or etcd for distributed locking and leader election.

What is the primary function of VRRP in Neutron's L3 HA router implementation?

Answer: Providing automatic failover from an active router to a standby router

VRRP in Neutron L3 HA mode allows a standby router to take over the virtual IP when the active router fails, maintaining tenant network connectivity.