Docs adapted to match.
Removed the unsupported-for-quay option to set up
a pull-through cache.

Closes-Bug: #1942134
Change-Id: If5a26b1ba4bf35bc29306c24f608396dbf5e3371

Continuing fixing CI after [1], this patch fixes the other branch
that [2] has not previously included.

[1] https://review.opendev.org/c/openstack/kolla-ansible/+/805449
[2] 02e07a08

Change-Id: I44014a93b92b5a8782e34cf394881dec74cdeea1

As a result of https://review.opendev.org/c/openstack/kolla-ansible/+/805449
CI is failing, because we don't have a TLS certificate on our registry.

This workaround will get our CI to be green while a proper patch
(TLS certs for registry) can be worked out.

Change-Id: Ia45c8a764a1f87d1c44717c4da3b9a3f94cdc967

Related-Bug: #1941982
Change-Id: I0e03db1177931ee6d17b21f614573575c3493eef

Currently only operations done with default kolla_toolbox user are logged
to /var/log/kolla/ansible.log.

In order to fix logging, permissions to ansible.log must allow writing
for other users in kolla group - and then a separate patch will follow
to make custom ansible.cfg file usable by other toolbox users.

Partial-Bug: #1942846
Change-Id: I1be60ac7647b1a838e97f05f15ba5f0e39e8ae3c

This is required for libvirtd with cgroupsv2 (Debian Bullseye and
soon others).
Otherwise, device attachments simply fail.
The warning message suggests filtering will be disabled but it
actually just fails the action entirely.

Change-Id: Id1fbd49a31a6e6e51b667f646278b93897c05b21
Closes-Bug: #1941940

corrected nits from:
https://review.opendev.org/c/openstack/kolla-ansible/+/800068
https://review.opendev.org/c/openstack/kolla-ansible/+/803644

Change-Id: Ia30afd795067a36b132a8c75c72dd7c65d624a83

Just like I added Cinder volume upgrade testing before, let's
also test similarly for Nova and Neutron. :-)

More robust debugging and refactor included.

Related-Bug: #1941706
Depends-On: https://review.opendev.org/c/openstack/kolla-ansible/+/806476
Change-Id: Id79df44254603f9b37ce7da2bfc796fc0b1ac91f

It was removed in [1] as part of cgroupsv2 cleanup.
However, the testing did not catch the fact that the legacy
cgroups behaviour was actually still breaking despite latest
Docker and setting to use host's cgroups namespace.

[1] 286a03ba

Closes-Bug: #1941706
Change-Id: I629bb9e70a3fd6bd1e26b2ca22ffcff5e9e8c731

Change-Id: Iebcac0827c6f715c6b804223cdcf2cc2e425120b

to behave like it is most commonly expected - query Nova in the
same region.

Closes-Bug: #1939291
Change-Id: I584a83d352c747a799b5dab1d3b8159ba3805454

To follow best security practices and help fellow operators.

More details inline and in the linked bug report.

Closes-Bug: #1940547
Change-Id: Ide9e9009a6e272f20a43319f27d257efdf315f68

For now role haproxy is maintaining haproxy
and keepalived. In follow-up changes there is also
proxysql added.

This patch is *only* renaming/moving stuff to more
prominent role loadbalancer, and moving also specific
templates to subdirectory.

This was done only to better diff in follow-up
changes.

Change-Id: I1d39d5bcaefc4016983bf267a2736b742cc3a555

Sometimes, the registries may intermittently fail to deliver the
images. This is often seen in the CI, though it also happens with
production deployments, even those with internal registries and/or
registry mirrors - due to sheer load when trying to pull the
images from many hosts.

This patchs adds two new vars to control retry behaviour.
The default has been set to make users happier by default. :-)

Change-Id: I81ad7d8642654f8474f11084c6934aab40243d35

It seems to have been mistakenly introduced by
de00bf49
"Simplify handler conditionals"

Change-Id: I65b6e322fa11a870f32099bbfd62150cbea4feb5