Closes-Bug: #2062401
Change-Id: I2f2bdbc9e1c6ad6da4ac7098ddd36143123c3062
(cherry picked from commit 904fae2a)

Change-Id: I3c419b4b2a86eb059744b05e4be797ea7ace8816
(cherry picked from commit 31c7317a)

Since we switched to proxysql as default - we've been seeing healthy
mariadb nodes being marked as OFFLINE_HARD by proxysql due to timeouts.

This seems to fix that.

Change-Id: I969db1f2b6c7a7ea821ec51c73e9e9a519febc4c
(cherry picked from commit 31ab71ac)

This is a follow up to Ie7ae47e5a624da31cf6cc5cd3c9239450487b8ed, but
for cephadm jobs.

Change-Id: I194529c178efcff0cb2a0ee1e14273389cd8f87f
(cherry picked from commit 4e2af187)

Ansible 2.16.14 includes a fix for CVE-2024-11079 [1], which results
in the following error:

    jinja2.exceptions.UndefinedError:
        'ansible.vars.hostvars.HostVarsVars
         object' has no attribute 'ansible_host'

This issue occurs because Ansible now includes `localhost` in
`hostvars` without the `ansible_host` variable set, which
unexpectedly breaks our test inventory rendering.

This patch adds a filter to exclude `localhost`
during the rendering process, as it is irrelevant for our purposes.

[1] https://github.com/ansible/ansible/pull/84353

Change-Id: Ie7ae47e5a624da31cf6cc5cd3c9239450487b8ed
(cherry picked from commit 0298699a)

Change-Id: I14d3ee3f94b8138188f70b2b5106145d33943eaf
(cherry picked from commit 61045807)

Change-Id: I044affa85290caf9d71e8ba5bc7cf0f5be8941e7

Change-Id: I3434eb7f751553a3a29a718ac20975529736bc3b
(cherry picked from commit 1cec85d6)

I have no clue how it worked previously in CI, but now
it's using default path to the inventory - which does
not exist.

In addition to that, type=int in cliff will default to
None, so the check had to be rewritten - because we
always did cert expiry check instead of generating them.

Change-Id: I84d71558c2666ba2cfa47054f782d25ff0c1f691
(cherry picked from commit cd8ecfc8)

Moving the CLI to python allows for easier
maintenance and larger feature-set.

This patch introduces a few breaking changes!
The changes stem the nature of the cliff package.
- the order of parameters must be
  kolla-ansible <action> <arguments>
- mariadb_backup and mariadb_recovery now are
  mariadb-backup and mariadb-recovery

Closes-bug: #1589020
Signed-off-by: Roman Krček <roman.krcek@tietoevry.com>
Change-Id: I9749b320d4f5eeec601a055b597dfa7d8fb97ce2

Change-Id: I33a3ec11b0cdef94b08cd7551008284755824cb7

This commit adds TLS connection between ProxySQL and MariaDB.
Frontend TLS ( between services and ProxySQL) will be
added in another commit.

Parialy Implements: mariadb-ssl-support

Change-Id: I154cbb096469c5515c9d8156c2c1c5dd07b95849
Signed-off-by: Matus Jenca <matus.jenca@dnation.cloud>

We are not testing mariadb backup, let's test
it in mariadb scenario.

Change-Id: I81d6ee944b3ed0e75129772bb993ce7a6147c3e8

After switching to ProxySQL as default we see following logs:
CRITICAL neutron [None req-c214fdae-5da7-402d-92b0-0572c278a5b5 - - - - - -] Unhandled error: sqlalchemy.exc.OperationalError: (pymysql.err.OperationalError) (9001, 'Max connect timeout reached while reaching hostgroup 0 after 10150ms')

Mainly in upgrade jobs, which otherwise pass successfuly - just fail on this
check.

Change-Id: I4336ec62a0a2dfbe815842f1bacb02135bcf4c0e

It's been some time since ProxySQL has been
with us in Kolla. Let's switch the load balancer
for MariaDB connections from HAProxy to ProxySQL.

Depends-On: https://review.opendev.org/c/openstack/kolla/+/928956
Change-Id: I42ba4fb83b5bb31058e888f0d39d47c27b844de5

ubuntu-ceph is broken for now due to [1], also there are no
download.ceph.com packages for Noble - so we're using Ubuntu
provided ones from proposed - because current version
in regular repos is built from git sha instead of a release
and is not suitable for running outside of Ceph upstream CI.

[1]: https://tracker.ceph.com/issues/66389

Depends-On: https://review.opendev.org/c/openstack/kolla/+/907589

Change-Id: I384068572d8a1a495c60b401dc4144a0a80802f1

Change-Id: I85982366d97e40d0e6ba97ce4c4c1ddd2c10b749

Change-Id: I972608b81edb2f5c118cf9a2f734f83ea399a849

This patch removes the hardcoded `distro_python_version`
mapping and usage from the configuration and templates,
aligning with the dynamic Python version detection
introduced in the dependent patch below.

The changes simplify the kolla-ansible roles by using
general `python3` paths, ensuring compatibility across
distributions without requiring version-specific handling.

Template files for Horizon, Ironic, Skyline, and others
have been updated to reflect this,
improving maintainability and reducing complexity.

Depends-On: https://review.opendev.org/c/openstack/kolla/+/926744
Change-Id: I85431b058b4184d96600cf17aaf8de871a018d61

From version 2.1, ProxySQL has a built-in ProxySQL
Prometheus exporter. This patch adds an option to
easily enable this exporter [1].

[1] https://proxysql.com/documentation/prometheus-exporter

Change-Id: I8776cdc0a6ec9e4e35a2424dd0984488514a711f

this is now more in line with other retry mechanisms:
use a local variable, use 10 attempts, sleep 10 seconds.

reason: there are still occasional resize failures in CI.

Signed-off-by: Sven Kieske <kieske@osism.tech>
Change-Id: I2ebb17459043ce677e72528bc7490b5728af396a

prometheus-msteams group got removed with I93d28ef138b4e784465f3a7eaa11101ea5877050
and is needed in upgrade jobs

Change-Id: Ief26303993128e327228ee438fe4fec5ff5c46e7

When cert generation during upgrade was fixed
it was omitted to also redirect stderr to the log file.

This commit fixes that.

Introduced-By: Ib257f8342c392b78e02ae31588ce5c6f8943f5b8
Closes-Bug: #2078364

Signed-off-by: Sven Kieske <kieske@osism.tech>
Change-Id: I43ca7d782bdcc53dcbc59dcf82ac1daa48e85ad5

There is a bug in current `tests/upgrade.sh`. If a
developer adds a new type of certificate i.e. for
RabbitMQ, it is not generated on upgrade because the
script does not run the `certificates` role.
This fix adds the `certificates` role to the script.

Change-Id: Ib257f8342c392b78e02ae31588ce5c6f8943f5b8

The prometheus-msteams project is no longer maintained [1]. As a result
support for deploying prometheus-msteams via kolla-ansible has been
dropped. Users are encouraged to migrate to the native Prometheus
Alertmanager integration with Microsoft Teams [2].

[1] https://github.com/prometheus-msteams/prometheus-msteams/issues/343
[2] https://prometheus.io/docs/alerting/latest/configuration/#msteams_config

Change-Id: I93d28ef138b4e784465f3a7eaa11101ea5877050

Library "distutils" is deprecated in Python 3.10:
https://peps.python.org/pep-0632/

The versions previously referenced using StrictVersion should be old
enough that they will not be used in a Dalmatian deployment:

- Ansible 2.11
- Docker API 1.42, included since Docker engine 23.0.0

Change-Id: Ie315004715a1cb5a91dd54bc64b0a8fd0af650ec

Change-Id: Ic87fb3e4c014d3090869d5631e02982829df6312

RabbitMQ uses quorum queues by default in both Dalmatian and Caracal, so
we don't need to run the migrations.

Change-Id: I1244f453af8f339cc3db7300951e366196e5291a

Change-Id: I2546d714d1e6a1b648a4c6242ece18a2108b9183

Closes-Bug: #2075966

Change-Id: I6ed82ecc7ca82c09aacb2a65254f794c72365188

Change-Id: I651f8e3fe9d0a247fa505e327847557ed426f6cf

Interface names with dashes can cause problems in Ansible since dashes
are replaced with underscores when referencing facts.

Change-Id: Ib15f81120d353ebcf8965dc0aaba90f3a9b0ec44
Related-Bug: #1927357

As our CI is ipv6 only and using cirros for tests.
Therefore we need to set force_config_drive = true.

Change-Id: I07058a4b30d0de1aaf8e82656024c5f6a99b290b

Closes-Bug: #2070051

Depends-On: https://review.opendev.org/c/openstack/ansible-collection-kolla/+/923544

Change-Id: I221eb136e77c61aef39e8646b48b927352d1419d

There was a bug where setting the test command
for the health check to 'NONE' would throw an error
in podman_worker. This was problematic since K-A
uses 'NONE' as an indicator that the health check is not enabled.

Closes-Bug: #2071912
Change-Id: I3140bb79eace58b23f579be3da569c502c52c38c
Signed-off-by: Ivan Halomi <ivan.halomi@tietoevry.com>

ansible-core 2.16 and later requires python 3.10+ (see [1])

[1]: https://docs.ansible.com/ansible/latest/reference_appendices/release_and_maintenance.html#ansible-core-support-matrix

Change-Id: Id5e10872de413e7b476c5343360d73c109b9667a

When using short notations like `1g` or `512m` to define the
container dimensions, we are always getting the container to
being restarted in each kolla-ansible run, even with no real
changes in the container configs.

Change-Id: Ic8e2dd42b95a8f5c2141a820c55642a3ed7beabd
Closes-Bug: #2070494

Change-Id: I9a71da5a00a4580b491a9c10ca34094b6654b438

it is already in init-runonce in the previous release

Change-Id: I7edc1216d3e74e995ed2051e9fd3d93196014836

Change-Id: I517d9f078a4dbb34178c851c1bc8b51b738882da