This matches a recent change in kolla-ansible [1].

[1] https://review.opendev.org/c/openstack/kolla-ansible/+/913724

Change-Id: I8a7a84025e93c4dcb1285bc28f3bd481bbdfff32

This should have been added when proxysql was added to kolla [1] and
kolla-ansible [2].

ProxySQL is disabled by default in this change to allow backports. It
will be enabled in a follow-up patch only for the master branch.

[1] https://review.opendev.org/c/openstack/kolla/+/769385
[2] https://review.opendev.org/c/openstack/kolla-ansible/+/770215

Change-Id: I683b685bf4e2f625144e4656bcd69c03eda7aa51

Depends-On: https://review.opendev.org/c/openstack/kolla-ansible/+/925581
Change-Id: Id9d90f424a5b0f7eedd1a11512bb8b72abba8959

Support for NCLU was added in Networking Generic Switch in the Zed
release[1].

Support for NVUE was added in Networking Generic Switch in the
2023.2 release[2].

[1] https://review.opendev.org/c/openstack/networking-generic-switch/+/743269
[2] https://review.opendev.org/c/openstack/networking-generic-switch/+/907551

Closes-Bug: #2083840

Change-Id: I973c6484ccd2a36990cf39df102e65be0376f08c

Change-Id: Id3e4e7ca285a94e7270eb4850c66e9094782f841

This updates the group and mode set on the /var/log/journal directory to
match default ownership and permissions used by systemd-journald.

Closes-Bug: #2083494
Change-Id: I15323b95199b9ef19504c7408dd89720f9ba9c32

Depends-On: https://review.opendev.org/c/openstack/kolla-ansible/+/923946
Change-Id: Icd6988668646e0ecb89822d46dafa33f9dd618e1

Use correct system scope - all - while getting the list of ironic
nodes in ipa-images role.

Closes-Bug: #2083014
Change-Id: I4da3a3739d57817ffbd094caea0fc80c3ad8eff8

When kolla_base_arch and ansible_facts.architecture differs
we need to run multiarch/qemu-user-static image - see [1].

[1]: https://github.com/multiarch/qemu-user-static



Co-Authored-By: Bartosz Bezak <bartosz@stackhpc.com>

Change-Id: If149418f509c687c3e5d1072cc39a80af33dec5a

Some network devices may use SSH key exchange algorithms that are no
longer supported by the Ansible control host. This will cause
ssh-keyscan to fail, preventing Kayobe from configuring the devices.
This change makes it possible to work around the issue by setting
switch_skip_keyscan to true for the affected devices. The SSH known
hosts file on the Ansible control host will need to be populated
manually.

Change-Id: I4e3394cff1fd86eb5c1a4be55d6fd7fd080b2944

eos_config does not support provider, nor connection: local from at
least 2022 [1] - since ansible>=7 (ansible-core 2.14).

[1]: https://github.com/ansible-collections/arista.eos/commit/588d94b2ed22817b63f996187e7d757d3ebad725

Closes-Bug: #2058365
Change-Id: Ief37b7bee758e34dddb881a9e68728013861c0ec

During the ``kayobe overcloud inventory discover`` command, Kayobe
queries the Bifrost ironic node list and updates the Kayobe inventory.
If the nodes have a driver_info.ipmi_address field set, an ipmi_address
variable will be defined in the Kayobe inventory for the host.

This does not work for Redfish based hosts, which use
driver_info.redfish_address in the Ironic node. This change adds a new
``redfish_address`` variable and extends the command to discover the
address from the Bifrost node inventory.

Change-Id: Ieb1e61ddadb542732739d1f478b1e5a176e0ac88

Previously Kolla Ansible hard-coded Neutron physical networks starting
at physnet1 up to physnetN, matching the number of interfaces in
neutron_external_interface and bridges in neutron_bridge_name.

Sometimes we may want to customise the physical network names used.
This may be to allow for not all hosts having access to all physical
networks, or to use more descriptive names.

For example, in an environment with a separate physical network for
Ironic provisioning, controllers might have access to two physical
networks, while compute nodes have access to one.

This change extends the 'physical_network' network attribute to make it
possible to customise the Neutron physical network names used for the
OVS, OVN, Linux bridge and OVS DPDK plugins. The default behaviour is
unchanged.

Depends-On: https://review.opendev.org/c/openstack/kolla-ansible/+/922320
Change-Id: I214444c60653f484fcda6275cc725879d14f9e7a

Added initial support for systemd-networkd link configuration, now
you can configure and rename the name of a network interface if you
know the MAC address of the interface.

Also added unit tests and fixed issues in the test_overcloud_host_configure.py:
 * Added unit test for networkd.
 * Fixed pep8 issues.
 * Removed unused import.
 * Fixed 'not in' issue in assert.

Change-Id: I8321183dbc747ef521aa0d2660ebeef8b0342c6a
Signed-off-by: Maksim Malchuk <maksim.malchuk@gmail.com>

Change-Id: I7d71689ef9b5a8a092fe78da4ee3d2a645a1316e

Bump stackhpc.libvirt-vm to v1.16.1.

Change-Id: I88f20112ab4740fe114a6484f88f4fbfb66863e5

Follows the same pattern as apt config.yml

Change-Id: Ifb347372032120ec676747bad6201e0d82725976

Follow up to [1] where I wasn't sure if this package was necessary.

[1] https://review.opendev.org/c/openstack/kayobe/+/923556/11

TrivialFix

Change-Id: Id4b3d26b869ba3999a4c8a647eaccdf19539bfad

This change bumps up the versions of Ansible installed in the
kolla-ansible virtualenv to a maximum of 10.x (ansible-core 2.17.x) and
minimum to 9.x. This is to align with the Kolla version requirements
currently in master.

Depends-On: https://review.opendev.org/c/openstack/kolla-ansible/+/921743
Change-Id: I4147a27ed04d28e832f11662e45dbae2851f63a8

There's no good reason to provide our own defaults for Ironic driver and
interface configuration. It easily gets out of sync with changes in
Ironic. Let's remove our defaults, and let Ironic provide them.

Note that this changes several of the defaults, which are listed in the
release note. Users are still able to override them using the existing
variables in ironic.yml.

Change-Id: Ie27adca0029e41439720eaac49b24cf13fc8068d

When using overcloud Ironic, IPA images for Ironic inspector are
downloaded to the Ansible control host to a cache directory, by default
/opt/kayobe/images/ipa/. They are later copied into the local Kolla
Ansible configuration under
etc/kolla/config/ironic/ironic-agent.{initramfs,kernel}.

The use of a shared cache directory results in problems when multiple
users share a single Ansible control host, since the cache is created as
writeable only for the user that created it. Other users sharing the
same Ansible control host will be unable to write to the cache.

We may also see issues if multiple Kayobe environments using different
IPA images are deployed from one Ansible control host.

The cache is not strictly necessary, since we can download the images
directly to the kayobe-config repo. This change avoids the use of the
cache. The performance impact should be minimal, only requiring an
additional download when a fresh kayobe-config is used.

Change-Id: I022c53afc0f64ccc79eeff4a220ade4c9216edfc
Closes-Bug: #2069845

Currently Kayobe sets ipv4_interface_mac to pxe interface
MAC address.

In cases where provisioning network interface and admin
network interface are different - this feature can be used
to get MAC address from Ironic introspection data.

Change-Id: Ie3c9248f0b3e47e3645e1886c0492265d52969c9

This works around a bug in the ansible packaged version of
community.docker when using requests>=2.32. It is also more flexible if
we can control the versions.

Also fixes an issue where kolla collections would not be installed if
the same role existed in the kayobe ansible collections path.

This relies on the support added in kolla-ansible to use ansible-core,
see: https://review.opendev.org/c/openstack/kolla-ansible/+/896406.

Closes-Bug: #2072980
Change-Id: Id9b19f10bc8ac38927914fb9782892b7daf1be82

Useful if you're configuring your own private repos, and don't/can't
sign them.

Change-Id: I2d592c5479530b2fe9a60c14ee59e817b8402490

Change-Id: I8f75fb1237183fbfbb790cd3089a2346a7e49277

Follow-up on I43a9c2a57fcfe2c9d84f39903aac7c258f9a897f. As discussed
before we set slightly more accurate name for the manage containers
role since using 'deploy-containers' role with action 'destroy' can
be a little confusing.

Change-Id: I6d4a3479d492e0fe86ce3de7414fa47c584577ea
Signed-off-by: Maksim Malchuk <maksim.malchuk@gmail.com>

Co-Authored-By: Mark Goddard <mark@stackhpc.com>

Change-Id: I900e3000e492b4eb668815ee6789c9d6e3637c68

This reverts commit b4448d6b.

Change-Id: I76aa0fcf362b77fe22f738b9d1c60cbc363531a4
Signed-off-by: Maksim Malchuk <maksim.malchuk@gmail.com>

Closes-Bug: #2067288
Change-Id: I57fb4c1697a31d4f592d79a31a04507cf91e97c7

This can be useful when using a seed VM that is not deployed by kayobe,
and can therefore not use ``kayobe seed deprovision``, especially when
iterating on kayobe-config during the inital stages of a deployment, or
for development, where you want to re-run the playbooks from a clean-ish
state.

Change-Id: I43a9c2a57fcfe2c9d84f39903aac7c258f9a897f

This allows us to continue execution until a certain proportion of hosts
fail. This can be useful at scale, where failures are common, and
restarting a deployment is time-consuming.

The default max failure percentage is 100, keeping the default
behaviour. A global max failure percentage may be set via
kayobe_max_fail_percentage, and individual playbooks may define a max
failure percentage via <playbook>_max_fail_percentage.

Related Kolla Ansible patch:
https://review.opendev.org/c/openstack/kolla-ansible/+/805598

Change-Id: Ib81c72b63be5765cca664c38141ffc769640cf07

The kolla_enable_letsencrypt variable was undefined, which was causing
overcloud container image build to fail when the variable was not
defined in Kayobe configuration.

The bool filter was missing for kolla_enable_letsencrypt, which was
causing letsencrypt images to be built when kolla_enable_letsencrypt was
set to "false" or "no".

Change-Id: I234b7a0434ba3d533a66b766331fd489078fbcf8

We were still using stream8 images which cannot install packages
anymore, due to the end of life of CentOS Stream 8.

Change-Id: I8604c342e4dbc41dd5bee0f28a92c3c3cd37d631

This was missed from I2ea9b2ec58913722c4defffbeee88cc420dcbdab.

Change-Id: I84971f8daa4caf2227863810c7b12e59192c2515

This is useful to ensure logs are persisted across reboots.

Change-Id: Ie5ff7536b160196dc57448c3ed982d15826b2cbe
Closes-Bug: #2055299
Co-Authored-By: Piotr Parczewski <piotr@stackhpc.com>

Change-Id: I704b4e79ce5519a57830fbcaa51b8cf19f3f249a

It's inactive and hasn't produced a 2024.1 release [1]. There are some
efforts to restore Freezer, but let's remove it for now.

[1] https://governance.openstack.org/tc/reference/emerging-technology-and-inactive-projects.html#current-inactive-projects

Change-Id: I65eccf6bbe13534418c2f5b0aa5397ba170bd17a

Switch all dependencies to use the OpenStack 2024.1 release.

This commit should be reverted on the master branch once the Kayobe
stable/2024.1 branch has been cut and RC1 released.

Change-Id: I02e0f97f84f1369943c840f6159c705dc2436750

Change-Id: Ia2a513ec7b14c52e1a89add2f3ba0c5e8e0202ea

This change bumps up the maximum supported Ansible version to 9.x
(ansible-core 2.16.x) and minimum to 8.x. This synchronises Kayobe with
Kolla Ansible.

Notable changes
---------------

- Removed use of get_md5 when using stat module, See:
  https://github.com/ansible/ansible/commit/d955fb1590efaaf996d7a48746e15f71cc65f583

- Remove use of include (instead of import_tasks/include_tasks) which
  has now been removed. See:
  https://github.com/ansible/ansible/commit/8db9bd757444fba6dd23fcb067e261d877926a33

Change-Id: I2ea9b2ec58913722c4defffbeee88cc420dcbdab