Disk image elements for deployment images of OpenStack
You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Go to file
Cédric Jeanneret 7adf4508cf Fix issues with tox 4.2.4
tox now hard fails if there is mismatch in spec
attributes[1]. This works around the issue
temporarily. We probably have to drop py38 jobs
and specify basepython for py39 target in the

Removes ansible-galaxy --timeout for py* jobs to
workaround. We may hit the timeout with py* jobs.

[1] https://github.com/tox-dev/tox/pull/2824/

Change-Id: Ie9bacf18cf167139601eff80bba91f2b3454b5dd
4 months ago
doc remove unicode literal from code 1 year ago
elements Switch to LVM thin provisioning 9 months ago
releasenotes Add release note for thin provisioning 7 months ago
tests Add doc/requirements 2 years ago
tools tox: Keeping going with docs 4 years ago
zuul.d Use python job template for Zed 1 year ago
.gitignore Migrate from testr to stestr 2 years ago
.gitreview OpenDev Migration Patch 4 years ago
.stestr.conf Migrate from testr to stestr 2 years ago
HACKING.rst Optimizing the safety of the http link site in HACKING.rst 5 years ago
LICENSE Fix copyrights for HP work. 11 years ago
MANIFEST.in Package with pbr 10 years ago
README.rst Replace git.openstack.org URLs with opendev.org URLs 4 years ago
babel.cfg Make it possible for openstack-CI to run tests 10 years ago
requirements.txt Add doc/requirements 2 years ago
run-flake8 PEP8 on the doc and elements files 9 years ago
setup.cfg setup.cfg: Replace dashes with underscores 2 years ago
setup.py Disable auto discovery 1 year ago
test-requirements.txt Migrate from testr to stestr 2 years ago
tox.ini Fix issues with tox 4.2.4 4 months ago


Team and repository tags


Image building rules for OpenStack images

These elements are used to build disk images for deploying OpenStack via Heat. They are built as part of the TripleO umbrella project.


Checkout this source tree and also the diskimage builder, export an ELEMENTS_PATH to add elements from this tree, and build any disk images you need:

virtualenv .
source bin/activate
pip install dib-utils pyyaml
git clone https://opendev.org/openstack/diskimage-builder.git
git clone https://opendev.org/openstack/tripleo-image-elements.git
export ELEMENTS_PATH=tripleo-image-elements/elements
diskimage-builder/bin/disk-image-create -u base vm bootstrap local-config stackuser heat-cfntools -a i386 -o bootstrap

Common element combinations

Always include heat-cfntools in images that you intend to boot via heat : if that is not done, then the user ssh keys are not reliably pulled down from the metadata server due to interactions with cloud-init.


OpenStack images are intended to be deployed and maintained using Nova + Heat.

As such they should strive to be stateless, maintained entirely via automation.


In a running OpenStack there are several categories of config.

  • per user - e.g. ssh key registration with nova: we repeat this sort of config every time we add a user.
  • local node - e.g. nova.conf or ovs-vsctl add-br br-ex : settings that apply individually to machines
  • inter-node - e.g. credentials on rabbitmq for a given nova compute node
  • application state - e.g. 'neutron net-create ...' : settings that apply to the whole cluster not on a per-user / per-tenant basis
We have five places we can do configuration in TripleO:
  • image build time
  • in-instance heat-driven (ORC scripts)
  • from outside via APIs
  • orchestrated by Heat

Our current heuristic for deciding where to do any particular configuration step:

  • per user config should be done from the outside via APIs, even for users like 'admin' that we know we'll have. Note that service accounts are different - they are a form of inter-node configuration.
  • local node configuration should be done via ORC driven by Heat and/or configuration management system metadata.
  • inter-node configuration should be done by working through Heat. For instance, creating a rabbit account for a nova compute node is something that Heat should arrange, though the act of creating is probably done by a script on the rabbit server - triggered by Heat - and applying the config is done on the compute node by the local node script - again triggered by Heat.
  • application state changes should be done from outside via APIs

Copyright 2012,2013 Hewlett-Packard Development Company, L.P. Copyright (c) 2012 NTT DOCOMO, INC.

All Rights Reserved.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at


Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Release notes for the project can be found at: