1223
|
|
Initial simple streams support
This branch adds a simple streams package to allow image metadata to be retrieved from json encoded metadata. The process starts by reading an index file from a known URL. The URL will be a static URL for EC2 (http://cloud-images.ubuntu.com/releases), and for openstack can be published in the keystone catalog or provided from an env config. This latter work to plug in the URLs from the providers will be done separately. This branch is just the core capability.
The data model used is a single implementation for all providers (so far EC2 and Openstack). Various providers use many different metadata attributes but these are ignored - only the common ones like id, region, arch, vtype etc are required to be read and used for filtering, and only the id is ultimately required by the calling business logic.
The tests are written to use example data to check the various processing scenarios, but also can be run live against canonistack or EC2. The live tests are a subset of the local tests and check that basic parsing and image lookup works with real data from two providers with significant differences in the exact data attributes used. The canonistack metadata has been published in a known public bucket.
R=rog, smoser, gz, jameinel, fwereade CC= https://codereview.appspot.com/9138044
|
Ian Booth |
11 years ago
|
|
|