For some reason this configuration was omitted or incomplete in these
jobs. Without this configuration, Kayobe development configuration adds
eth1 to the breth1 bridge if it exists, or leaves the bridge without a
port otherwise. This can lead to network connectivity issues, if the IP
address of eth1 becomes inaccessible when it is added to breth1.
This change fixes the issue by ensuring that the dummy1 interface
exists, and configuring the aio network to add it to breth1 as a port,
instead of eth1.
Change-Id: I47115e5d8d1fb448cb9e46dc43b56c64987e428b
Set packages argument explicitly as a list to avoid "No package matching
'<generator object select_or_reject at 0x7ff25f341750>' is available" error.
Story: 2008378
Task: 41298
Change-Id: Ibe44583cb93b7dca8f5091c893386a15288af915
Add file to the reno documentation build to show release notes for
stable/victoria.
Use pbr instruction to increment the minor version number
automatically so that master versions are higher than the versions on
stable/victoria.
Change-Id: I79b0d63a283b56a12b5acd3aaa740a6be3c3cb6f
Sem-Ver: feature
Switch all dependencies to use the OpenStack Victoria release.
This commit should be reverted on the master branch once the
Kayobe stable/victoria branch has been cut and RC1 released.
Change-Id: I7f1113577fd48aa0715b644cc66bfca197771ca6
This change adds a new 'scenarios' section to the configuration
documentation, with an initial 'all in one' scenario.
Change-Id: Ibe9cbbb59e2f72b18fdeb493feb085735edbbf8c
Story: 2004360
Task: 27960
Variables to customise hardware inspection timeouts were introduced in
change I2e45876f89121d66cf03b50824bd8258517b88cb. Unfortunately, the
wait_inspected_timeout variable is turned into a string despite
inspector_inspection_timeout being originally defined as an integer,
causing the following error:
Unexpected templating type error occurred on ({{ wait_inspected_timeout // wait_inspected_interval }}): unsupported operand type(s) for //: 'str' and 'int'
Change-Id: Ie40a3eed7146e2e2d8b1c08860da395b35ed99d6
Story: 2007844
Task: 40135
In some environments the can_talk_to_docker key is missing from the
result of docker_host_info when the call succeeds. Assume we can talk to
Docker if this key is not present, instead of failing due to a missing
attribute error.
Change-Id: I7670410394b0bb5fc96e4166b6a20714f847250b
Currently we filter out networks that are set to None. However, we
should also filter out networks set to an empty string. This affects the
overcloud ironic networks that were disabled in
I300d0ef136224126f25d2c70a80a42afeea5f586 when ironic is disabled.
Story: 2008207
Task: 41076
Change-Id: Iba16d6232e1a6ef069cd3ccb870e5bfa89cd913c
Without this setting, bifrost creates a bifrost firewalld zone only
allowing network traffic for Ironic services and assigns the
provisioning network interface to it, potentially causing loss of
connectivity.
Using the public zone is suggested as a workaround [1] but is not
sufficient: it allows SSH traffic, but blocks other services deployed on
the seed, such as Docker registry traffic.
[1] https://review.opendev.org/#/c/754406/
Change-Id: I80f9d95f02e11fda5916f9a9dd257b688a9db7e2
Story: 2008153
Task: 40899
To avoid switching existing deployments from devicemapper to overlay2,
we check the existing storage driver configuration directly with the
Docker daemon, or if unreachable by reading the /etc/docker/daemon.json
configuration file.
Co-Authored-By: Pierre Riteau <pierre@stackhpc.com>
Story: 2005667
Task: 30972
Change-Id: Iaf2ee8c9f302f4684ae039bb00b2e2e5969cf1fc
While we always test baremetal compute in CI, development environments
may not. Given that Ironic is now disabled by default, we should make
this work out of the box.
Story: 2008207
Task: 41003
Change-Id: Id3128380f5ff74d24265f6b2132c6d7992bf00ba
We do not need the workload provisioning, cleaning or inspection
networks when Ironic is disabled in the overcloud.
Change-Id: I300d0ef136224126f25d2c70a80a42afeea5f586
Story: 2008207
Task: 40992
Ironic is now disabled by default in the overcloud. This brings Kayobe's
default set of services into line with Kolla Ansible. For environments
using Ironic in the overcloud, set kolla_enable_ironic to true in
kolla.yml.
Story: 2008207
Task: 40991
Change-Id: I33eb4fa534847e199a599350b525d4762a2beaac
The flag dnf_install_epel was being used to predicate installation
of the EPEL RPM repo package, but not preventing the installation
of configuration for access to EPEL. This small patch completes
the job.
Updates the overcloud host configure CI job to install EPEL mirrors,
since it was previously relying on this buggy behaviour.
Change-Id: Ib417837d0772338b16ea9f7f2540549f277d5de8
When generating host variable files for Kolla Ansible, we have some
heavy usage of set_fact to set variables for network interfaces, in a
play targeted at all hosts. There are also tasks using the fail action
plugin to perform verification.
At scale this has a significant impact, due to the number of tasks
executed against all hosts. These tasks are executed at the beginning of
many commands, so the scope is broad. There is also some tricky logic
involved, which is difficult to express in Ansible/Jinja.
This change replaces the use of set_fact with a custom Ansible action
plugin. The plugin executes locally on the Ansible controller, and
returns a dict of Ansible facts to set for each host. The plugin is
executed once for each overcloud host, and returns all relevant facts.
The plugin also performs verification. Extraction into a Python module
allows for unit testing.
This has been shown to have a significant improvement on execution time,
particularly as the number of hosts reaches 100 or more.
Story: 2007993
Task: 40641
Change-Id: I443da1ae05fcca2d7d8dff7db563eeda37e9f502
Ironic has an issue where nodes will be switched to use the direct
deploy interface during upgrade, if it is enabled. This should be fixed
via https://review.opendev.org/755295, after which this patch should be
reverted.
Change-Id: I55af5d979bdbee9dd963845204c2f082fa7b27d7
When molecule runs, it installs ansible>2.8, which is version 2.10.0
since today. This now also installs the ansible-base package, currently
at version 2.10.1. Then it installs the version of ansible required by
kayobe, i.e. ansible>=2.8.0,<2.10.0 (currently 2.9.13).
We end up with a tox molecule venv containing:
ansible==2.9.13
ansible-base==2.10.1
This causes the following issue:
ERROR! Unexpected Exception, this is probably a bug: cannot import name 'AnsibleCollectionLoader' from 'ansible.utils.collection_loader' (/home/zuul/src/opendev.org/openstack/kayobe/.tox/molecule/lib/python3.8/site-packages/ansible/utils/collection_loader/__init__.py)
This commit adds requirements.txt to the molecule tox deps, which
ensures the correct version of ansible gets installed first.
Change-Id: I7ef890a2ae6fd67f53e1b8bca244a1c0d44fea53
Kayobe generates a host_vars file for each host in the Kolla Ansible
inventory. These contain network interfaces and other host-specific
things. Currently this is done by iterating over all hosts, which does
not scale well with a large number of hosts.
This change extracts the host vars generation into a separate role, and
executes it in a play targeted at all hosts, with delegate_to:
localhost. This ensures that host variable files are generated in
parallel.
Story: 2007993
Task: 40629
Change-Id: Iae75e17024adee9c2874c14d3ed36f4c87ba48d7
hacking 3.0.1 fix the pinning of flake8 to avoid bringing in a new
version with new checks.
bumping the min version for hacking so that any older hacking versions
which auto adopt the new checks are not used.
Change-Id: I3f4fca9a0df2f45241288fd876064433f4e2f737
