Bionic check jobs are now passing so move these to voting and
add gate/periodic jobs for bionic.
Introduce a release note detailing the major changes and indicating
a viable upgrade path from Xenial to Bionic.
Change-Id: I0897e4f22b3949ce07a2ea0f2ee9c4f45c344e04
The 'install_method' variable can be used to deploy OpenStack services
either from source or from distribution packages. Since this variable
is meant to be set during the initial deployment and can't be changed
afterwards, it is important to document it in order to let deployers
know about it.
Implements: blueprint openstack-distribution-packages
Change-Id: I59fe66a5cef45487afbfc090aca93cb567427f7a
Currently the default CentOS value is to set printk to
7 4 1 7, showing up all the DEBUG messages.
OpenSUSE Leap 42.3 is using 1 4 1 7, while Ubuntu is
using 4 4 1 7.
This patch changes the default console_loglevel to 4 in AIOs
to only print KERN_WARNING, KERN_ERR, KERN_CRIT, and
KERN_ALERT, as this seems the best balance of reducing logging IO
while still exposing relevant errors in tests.
This patch also ensures users are aware that centos have a higher level
of kernel logging by default, so they can take their own decisions
about it.
Change-Id: I4ceb5b7d436b258656b0193dd5e2ef3efb44f592
In Ubuntu 14.04, the kernel shipped was lightweight, and some
extra features like using vhost_net, required the package
linux-image-extra.
In Ubuntu 16.04 and above, this is not the case anymore,
as many kernel packages ship with extra features.
vhost_net is generally shipped with the kernel package nowadays
and was the only one documented for the addition of the
linux-image-extra into our roles.
Change-Id: I70fa446c2dd00e12702280388299d189e0f3e5c4
Depends-On: https://review.openstack.org/#/c/585376/
Closes-Bug: 1783342
The deployment host documentation gets out of date really quickly.
This patch makes it self maintaining. Note that due to the fact
that RDO/SUSE do not implement their repositories for a new
release until much, much later in the cycle, we use the previous
release repo for master. This will self-update during our RC
when the stable branch is cut.
Some formatting adjustments have made to prevent the interpretation
of URL's as links by sphinx and to remove the use of multi-line shell
commands.
Closes-Bug: #1759721
Change-Id: Id28540d335390a52e0eb65a734a1d594235265d6
We don't need that many variables for the deployment guide, there
is barely any substitutions done. Simplify the documentation to
a bare minimum.
On top of this, we adapt for the latest openstackdocstheme.
Change-Id: I3502b6a943bacc6dac1b757961348eb04771750d
The OpenStack Ansible Deployment manifesto [1] was lost in times in
pages long forgotten of the wiki.
Instead, it should be part of the documentation, in our reference
section. The reference is a great place because it's where we
explain WHY we are doing things.
To increase visibility of this page, the design decisions are
directly linked from the deploy guide appendix about
"Why OpenStack-Ansible".
Change-Id: Ifb3ef041c773ee6dbe6606e2e5efc0542776dfc5
Disks requirements aren't clear, and people can think it's
always required to have 1TB disk on cinder nodes. There is no
hard requirement there, so we should remove this requirement,
mark it as recommendation, and point its only for LVM backends.
Change-Id: Ic45a7799e5092653d70fce9f2b6b7ae3f03d8775
This brings consistency between target hosts configuration and
deploy host configuration, to be easier to read.
Change-Id: Ibaa01dfc6190f41ea0a3d1ca353296c3e8ec4f7f
This modifies the overview section to be one page instead of
4 long sections. We just refer to the architecture reference
for details about the architecture.
Change-Id: I1f5da9acaa677b2cd752cda30c33eba846a96eb4
- It's easy for a new deployer to miss the importance of the
configuration. This highlights the step.
- It makes more sense to be guided from the deploy-guide to
the Operations guide first, before the contributing.
- Reduced the commands to run for a first verification of
the cloud.
Change-Id: I7bc7dce2d4a9f0449a3f70cc2717f5e0a8345dc4
There are details about inventory configuration in the deploy
guide that could go to the reference.
Change-Id: I712470742250f37a410eb88087b14403ac777493
Moving the ceph example into the user guide.
This structure also allows more text content, with for example
a guide for integrating with a ceph cluster deployed outside
OSA, and other ceph integrations (Rados GW for example)
Change-Id: Ib3eb33e23cfc50966b41842a02feeaf21f836450
The deploy guide also have appendixes pointing to the reference,
which were pointing back to the deploy guide. This is very
hard to follow. The advanced configurations and inventory should
be merged into our reference guide, to avoid duplications.
Change-Id: I2d1e653036d380c3eae6a3a4906d1db890baecb2
The SSL was deeply hidden into the advanced configuration options,
while it's a common user story.
Change-Id: Ie09b94fb086f512ec1b233cb76754654ba032418
To be more like a walkthrough, the next step after the
verification of the environment would be to contribute and
operate OpenStack-Ansible.
This could also bring more visibility to contributions.
Change-Id: I991ad7963bcd777e279738bb9d4ca03992f46280
Do not show a deep ToC when landing on the page, instead of show
a more user-friendly access page.
Change-Id: I0a571db85d67ff6a89b11adfbc72b0e27374bf9f
Anyone wanting to quickly deploy an openstack-ansible cloud has no
way to know we have an AIO toolkit that could help.
This should fix it.
Change-Id: I7cabfac0bfb01b77e2a250e5879671c95cbb46c6
Bumping shas to head of stable/queens. For projects that don't
have that branch yet, a reminder was written on the same line as
the SHA.
Updates documentation to the appropriate branching.
On top of that:
- Placement API now only accepts GET, not HEAD
- Placement API now returns HTTP 200
- Designate returns the API versions under /, not under
/versions [1]
- Cloudwatch was disabled upstream [2], so all references are
removed.
[1]: https://developer.openstack.org/api-ref/dns/
[2]: https://bugs.launchpad.net/heat/+bug/1743707
Change-Id: I29e63008bf5c3cd8ed2deb8dcd13bc6a988891f5
Move the playbooks/inventory folder, group_vars, and host_vars to
inventory/ in the root of the OpenStack-Ansible repo. This helps better
organize the repo structure since playbooks/ will now only contain
playbooks, shared task files, and included repo package var files.
group_vars and host_vars are moved alongside the inventory since that's
the default place that Ansible expects those folders and to help better
prepare for Ansible 2.4 where multiple inventories can be loaded,
automatically including relative group and host var files.
Effected docs, scripts, and variables have been updated with the new
paths.
Change-Id: If50e2412c3fd6575d7041deb8ecc9480b04184cc
There are details that confuse people (like the usage of may), when
they are configuring a ceph environment.
This should clarify the situation.
Change-Id: Ic8b734eabcfa218c687db69290c8b3cef1def7c5