Book Image

Production Ready OpenStack - Recipes for Successful Environments

By : Arthur Berezin
Book Image

Production Ready OpenStack - Recipes for Successful Environments

By: Arthur Berezin

Overview of this book

OpenStack is the most popular open source cloud platform used by organizations building internal private clouds and by public cloud providers. OpenStack is designed in a fully distributed architecture to provide Infrastructure as a Service, allowing us to maintain a massively scalable cloud infrastructure. OpenStack is developed by a vibrant community of open source developers who come from the largest software companies in the world. The book provides a comprehensive and practical guide to the multiple uses cases and configurations that OpenStack supports. This book simplifies the learning process by guiding you through how to install OpenStack in a single controller configuration. The book goes deeper into deploying OpenStack in a highly available configuration. You'll then configure Keystone Identity Services using LDAP, Active Directory, or the MySQL identity provider and configure a caching layer and SSL. After that, you will configure storage back-end providers for Glance and Cinder, which will include Ceph, NFS, Swift, and local storage. Then you will configure the Neutron networking service with provider network VLANs, and tenant network VXLAN and GRE. Also, you will configure Nova's Hypervisor with KVM, and QEMU emulation, and you will configure Nova's scheduler filters and weights. Finally, you will configure Horizon to use Apache HTTPD and SSL, and you will customize the dashboard's appearance.
Table of Contents (16 chapters)
Production Ready OpenStack - Recipes for Successful Environments
Credits
About the Author
About the Reviewers
www.PacktPub.com
Preface
Index

Configuring Cinder scheduler filters and weighers


Cinder supports configuring multiple storage backends with identical volume_backend_name. In this configuration, Cinder-scheduler automatically schedules the placement of new volumes based on scheduler, which filter out storage backends that don't fit the filter policy, when multiple backends fit the filter policy, the backend storage is chosen, based on scheduler weighers.

CapacityFilter and CapacityWeigher are the default filter and weigher, respectively. When they are enabled, cinder-scheduler assigns the highest weight to the backend with the most available capacity.

VolumeNumberWeigher spreads volumes evenly across all storage backends that passed the filters, so for example, if two storage providers share the same volume_backend_name, and 10 volumes are created, each provider backend would host five volumes. This weigher provides another means of balancing workloads across multiple backend providers.

DriverFilter and GoodnessWeigher...