Disk image elements for deployment images of OpenStack
Go to file
Steve Baker 2ce67c3dbb Install modified udev rule to fix lvextend unmount
This fix has been proposed to lvm2 upstream[1]. This change can be
reverted once the fix is packaged. It is proposed here to unblock
upstream and downstream delivery pipelines.

[1] https://github.com/lvmteam/lvm2/pull/105

Change-Id: If187a2b1ec61ec47738b99b40919d4cc65fa9505
Closes-Bug: #2000226
Related: rhbz#2158628
2023-02-02 12:55:24 +00:00
doc remove unicode literal from code 2022-02-23 02:53:21 +00:00
elements Install modified udev rule to fix lvextend unmount 2023-02-02 12:55:24 +00:00
releasenotes Add release note for thin provisioning 2022-10-21 10:21:37 +13:00
tests Add doc/requirements 2021-01-06 16:22:12 +02:00
tools tox: Keeping going with docs 2019-11-03 02:46:54 +08:00
zuul.d Use python job template for Zed 2022-05-20 08:02:59 +00:00
.gitignore Migrate from testr to stestr 2021-07-12 20:34:26 +05:30
.gitreview OpenDev Migration Patch 2019-04-19 19:33:10 +00:00
.stestr.conf Migrate from testr to stestr 2021-07-12 20:34:26 +05:30
HACKING.rst Optimizing the safety of the http link site in HACKING.rst 2018-11-16 09:50:05 +08:00
LICENSE Fix copyrights for HP work. 2012-11-15 16:20:32 +13:00
MANIFEST.in Package with pbr 2013-08-06 12:50:39 +12:00
README.rst Replace git.openstack.org URLs with opendev.org URLs 2019-05-28 14:40:51 +08:00
babel.cfg Make it possible for openstack-CI to run tests 2013-02-04 22:26:17 -08:00
requirements.txt Add doc/requirements 2021-01-06 16:22:12 +02:00
run-flake8 PEP8 on the doc and elements files 2014-08-19 10:02:00 +02:00
setup.cfg setup.cfg: Replace dashes with underscores 2021-05-14 16:52:22 +08:00
setup.py Disable auto discovery 2022-03-28 15:23:54 +05:30
test-requirements.txt Migrate from testr to stestr 2021-07-12 20:34:26 +05:30
tox.ini Fix issues with tox 4.2.4 2023-02-02 13:27:43 +01:00


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: