Move to using local K-A fork that specifies 2023.1-eom tag for Kolla Ansible collection in its requiremens.yml

This bug would cause nova-compute services to freeze, which would result
in frequent monitoring alerts.

Switch to using the StackHPC fork - master branch - when building blazar

* [2023.1] Fix eswitchd and neutron_mlnx_agent not using our fork 

The neutron_mlnx_agent based containers are not using the neutron plugins, so we need to customize the pip package directly.

* Bump neutron containers

* Update kolla-image-tags.yml

* Update kolla-image-tags.yml

This fixes issues with creation and failover of TLS-terminated Octavia
load balancers following our bump of pyOpenSSL [1] because of the
removal of load_pkcs12 [2].

[1] https://github.com/stackhpc/requirements/pull/20
[2] https://bugs.launchpad.net/octavia/+bug/2042787

* Add sed for multiline baseurl in yum repos

* use new Rabbitmq Cloudsmith repos

* use rabbitmq built from Cloudsmith repo

* Add sed for multiline baseurl in yum repos

* use new Rabbitmq Cloudsmith repos

* use rabbitmq built from Cloudsmith repo

* (2024.1) Reintroduce networking-mlnx

* Adds built images

* Reintroduce networking-mlnx

* Use stackhpc fork

* Use images built from StackHPC branch

* Pin ironic neutron agent

This wasn't rebuilt

* Update kolla-image-tags.yml

* Remove WIP references

* Adds release note

* Update kolla-image-tags.yml

* Simplify

* Update etc/kayobe/kolla-image-tags.yml

---------

Co-authored-by: Tom Clark <tom@stackhpc.com>

More info: https://bugs.launchpad.net/kolla-ansible/+bug/2062401

* Remove stackhpc/neutron source for 2024.1

* Add downstream upper-constraints ENV variable

Bifrost and Kolla-toolbox needs override as their build
process is different to other Kolla images.

* Update pulp repo version for Caracal

* Update Openstack RPM repo to Caracal

* Set koll_build_args as empty dict to avoid build error

* Update StackHPC Openstack rpm repo target to Caracal

* Re-add omitted rhel9 influxdb repo

* Update centos9 caracal repo version

* define openstack-base from kolla_sources

* Use upstream requirements for testing

* bring back stackhpc requirements

* Remove antelope prometheus version override

* Bump Ceph to Reef

* Remove horizon stackhpc fork source

* Fix broken package repo versions

* Update version of RL9 third-party repos

Update threasuredata from 4 to 5
Update mariadb from 10.6 to 10.11

* Explicitly set version of prometheus version

* Add missing requirements option

* Update rpm repos to match Caracal

* Fix RabbitMQ repo naming

* Fixup Caracal Kolla config

* Bump pulp repo versions

* Build OVN Octavia provider from upstream

* Revert "Explicitly set version of prometheus version"

This reverts commit 901f67f7a45a6008fb545eea13b4a7e7fe73ce85.

* Pulp repo versions eof newline

---------

Co-authored-by: Seunghun Lee <seunghun@stackhpc.com>

The bug fix we were using in our fork was backported upstream [1].

[1] https://review.opendev.org/c/openstack/ovn-octavia-provider/+/926715

This resolves a bug breaking OVN IPv4 load balancers with health
monitors [1].

[1] https://bugs.launchpad.net/neutron/+bug/2072754

This is necessary to address OSSA-2024-002 [1] until patches are merged
upstream.

[1] https://security.openstack.org/ossa/OSSA-2024-002.html

This patch updates:
- OVN to 24.03 (latest LTS)
- OpenvSwitch to 3.3 (since 3.2 is officially EOL upstream)
on Rocky Linux 9

We are currently carrying patches for the CVE-2024-32498 QCOW vulnerability.

This reverts commit 1e7527e3.

We are currently carrying patches for the CVE-2024-32498 QCOW vulnerability.

With the upstream CS8 repos gone, an issue with the extras &
extras-common yum repo setup in our container images was shown.
We were relying on the upstream extras repo, rather than the one in
release train.