← Back to team overview

openstack team mailing list archive

Deployment Blueprint Discussion: What is our Reference Deployment?

 

All,

This is the follow-up for our "Deployment Fixture" blueprint brainstorm at the conference (https://blueprints.launchpad.net/openstack-common/+spec/deployer-api ).

As an Essex release object, I'd like to be automatically deploy OpenStack from a community standard configuration description file (aka "Deployment Fixture").

To get there, the first step is for us to agree upon the target deployment layout.  I believe it is critical that this deployment is a production class deployment with reasonable scale, redundancy and separation of concerns.  This email is intended to be the start of a lively discussion about what that deployment should look like.  Once we've got a workable deployment layout then we can work on the description API.

Here's my starting proposal as a reference deployment  - PLEASE EXPAND & ADD DETAIL

OpenStack suite:

*        Nova Controller

o   2 nodes with redundant components for SQL, Message bus, and load balanced API

o   Backed by Glance

o   Integrated with Keystone

o   Fronted with Dashboard

o   3+ compute nodes

*        Dashboard

o   2 nodes with redundant components

*        Glance

o   backed by Swift

o   integrated by Keystone

*        Keystone

o   2 nodes with redundant components

o   SQL data store

*        Swift

o   Proxy on 2 nodes with redundant components

o   At least 5 nodes with 4+ drives per node

*        Networking

o   Use of VLANs to logical segmentation

OpenStack Networking?  Suggestions

Rob
______________________________
Rob Hirschfeld
Principal Cloud Solution Architect
Dell | Cloud Edge, Data Center Solutions
cell +1 512 909-7219 blog robhirschfeld.com, twitter @zehicle