The valid types for nova_console_type defined in os_nova defaults/main.yml
do not include 'serial', but instead 'serialconsole'.
Change-Id: I700b699625ff134adb31e0f47dc5b7ad6ea0e1ed
With switching from LXB to OVN we need to have neutron_plugin_base
defined explitily in variables, otherwise ovn one will be used as
default.
Change-Id: I4da481d182ce8e207ff847c2adfebbbb17af358b
At the moment there is no handy functionality to override ENV variables
and store them in git near OSA config. As while we take into account
that ENV vars can already exist and we should use them, these VARs must
be defined in /etc/environement or bashrc files.
This patch aims to look for user.rc file inside OSA_CONFIG_DIR.
If it's present - it will be sourced with all it's content. This way
deployers can store environemnt variables overrides in git and they will
be loaded during on openstack-ansible startup.
Change-Id: Ie24ada54a0e0dc064be028929b416d983fdb5b49
On ubuntu 22.04 specifying vers and nfsvers raising error during mount:
mount.nfs: multiple version options not permitted
To make NFS jobs passing for 22.04 we drop out vers.
Needed-By: https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/865518/
Change-Id: I925a148febce2bbdc09e60020828947589905dcf
Zookeper is being added as a coordination backend which is
required for some service features to work properly.
Change-Id: I2d464ae339f057ea4ba08bd343b6a51c029a74f3
With migration to PKI role for Octavia in Yoga, we added upgrade
playbook to define certificates for users that are upgrading deployment.
With SLURP releases, we either need to drop this upgrade step now or
carry on until Antelope.
Change-Id: I2c72bf6324b443c6b01a7f3142ebc90768c1c9c8
We started seing tempest failing on keystone quite frequently since
moved keystone behind Apache instead of nginx. At the same time we see
no issues. Also there are no issues logged which points towards too
small amount of threads/connection available for normal work and even
serial tempest tests. We try to minorly increase amount of threads for
Apache to see if that helps.
Change-Id: I6ce27125d2fa670bb0f240195a59f8ea242369d9
With release of new setuptools running setup.py --version started
returining extra line. Eventually, running setup.py as a script is not
considered as good practice and we should not rely on it's output much.
Instead, we can leverage importlib.metadata as it should be more
reliable in long run.
Closes-Bug: #1997365
Change-Id: Ife227db4ad25851456e1acc6b4daf198f39d989a
Haproxy config check(/usr/sbin/haproxy -c -f /etc/haproxy/haproxy.cfg)
shows a warning:
'http-request' rules ignored for backend 'galera-back' as they require
HTTP mode.
It's caused by the way we configure stick tables in haproxy.
We don't need stick tables for galera as it has only one active backend.
Change-Id: Ie41e26c1dcc671c30d39b34d3f436e225e7580a5
With current "Ceph production example" the difference between ceph's
public and storage network is not clear.
We assign Storage Network to compute nodes, but it's not used there.
We also asign Storage Network to ceph monitors, but it's not used there
as well.
Same problems apply to AIO environment.
As Dmitriy suggested in [1], ceph should not use mgmt network for
storage traffic.
This change makes ceph use storage network for:
- OSD<>OSD communication
- client<>OSD communication
- client<>MON communication
I think it's the most common scenario where all ceph-related traffic
uses dedicated(storage) network and do not depend on mgmt network.
This change affects both "Ceph production example" docs and AIO
environments.
[1] https://review.opendev.org/c/openstack/openstack-ansible/+/856566
Change-Id: I74387a2e961e2b8355ea6a0c889b2f5674233ebf
Upgrade path is not tested for ceph-ansible and thus we can not
recommend it's integrated usage with OSA. These nits should be
documented properly to explain the risks.
This patch is alternative implementation to more extensive one [1]
[1] https://review.opendev.org/c/openstack/openstack-ansible/+/862508
Change-Id: I1031cff1b1e13ca3e0d8c97c9cd28f25bfd5c6f1
We are moving Ubuntu 22.04 support out of experimental state
which means that Jammy should be voting now. With that we are also
replacing focal jobs with jammy in gates to save CI resources.
Change-Id: I91c5e130348f4080cf947cb1097f78da6a34bfb4
Ubuntu Jammy repository is not available on download.ceph.com but
we still want/need to test out ceph deployment one way or another.
So instead of failing installation we install distro-provided ceph.
Change-Id: I3c1151449afe545780f98c01afa9fc6948f685a4
According to my tests, ``glance_show_multiple_locations`` is not
needed neither for ceph nor horizon's direct image upload.
For safety reasons [1] it's better to keep this option disabled.
[1] https://wiki.openstack.org/wiki/OSSN/OSSN-0090
Change-Id: I8c36196cd4b35de98cc3c1cabd8c8c2e03ecdae7
We've switched all services to store logs to journald by default and
rsyslog roles are not used except really small amount of usecases that
also hardly valid as of today. With that we deprecate repos and remove
their usega to reduce maintenance load.
Change-Id: Iefd4143f83f4df44b917180000a1aa57161b2811
We build LXC images with rpm/debootstrap and don't download external
ones for quite a while, thus we never actually use this override in our
roles and this overrided can be safely cleaned up.
Change-Id: If5d7777f80b67beb1e83d8f933552d0431d0f4a5
It is possible to install nova-serialconsole-proxy alongside
nova-compute to allow IPMI serial-over-lan consoles to be made
available through the openstack cli or through horizon.
The os_nova role needs to have logic to decide which ironic host to
install the nova serial proxy to, so we define an extra ansible group
for this purpose.
Change-Id: I936aa89129aeda56af9d9a1407073fccf00b7170
