- Add jobs removed in [1] to the local configuration.
- Fix SSH key distribution in multinode job. It was broken on a
recent Zuul update, it seems.
[1] - https://review.opendev.org/730813
Change-Id: I2572d92cbaabc1f02fc0790896ccaa3edfc5e6c2
(cherry picked from commit 76b1559d3a161b0d2326ada77bc69d089639075f)
(cherry picked from commit 100c950bc7)
Also pinning gettext as version 3.3.0 is not compatible with the ruby
version in CentOS7.
Change-Id: I8d226aa2218eafc07119f88eca5be24c3514b717
(cherry picked from commit 3be5458f4b)
We were lacking the "sync_db => true" parameter when calling aodh::api,
so the database was not initialized, and API calls were failing.
Change-Id: Idbc4f5a3310c0776ba4973b45dbaf554d8372b8e
Closes-Bug: #1855443
(cherry picked from commit 1f64071f5a)
Previously we were storing the whole contents of /var/tmp/packstack,
which includes all Puppet modules and is unnecessarily filling up
the log server.
Instead, let's store /var/tmp/packstack/latest, where the last
Packstack execution logs are kept. That is really what we need.
Also added 'ip netns' command to logs to help troubleshooting issues
with dhcp agent.
Change-Id: Ia138c13c0797d2b83c9e3b4037c4648abf8f7295
(cherry picked from commit cfdbeb5d02)
Currently, we are using cached image for UEC (tempest) but not for demo
provisioning.
Change-Id: I637edcdcc129dc8ed86eb4f3fdd4e216dd3f48a7
(cherry picked from commit f9cc0729ae)
Use https for cirros downloads
http is being much slower and anyway using https is a good practice.
Change-Id: Ib9b4194c88cd9065babb81c788ed52b2ee5b8d2f
(cherry picked from commit 97e5a40dbb)
The replication to github is no longer automatically managed by opendev.
Change-Id: I76eb105f6cd2624b6c9edfc0dcedd2f2c56aa5af
(cherry picked from commit 51fb71534a)
It's used in copy-logs.sh. Also copy all content of /var/tmp/packstack
instead of only latest directory.
Change-Id: I7c6d72876e881f73d2421206e004f514563ad803
(cherry picked from commit fcdaff0d25)
Some upstream infra is enabling nested virtualization but it's failing
randomly so let's force qemu virt_type.
We need it to debugs issues with jobs incorrectly detecting
virtualization time.
Change-Id: I1e53bbdcccb6402d343c98abf2696d452717d1a9
(cherry picked from commit c84e3b6e84)
Currently, u-c is not uses so we are not using the versions specified in
requirements project.
Change-Id: Id0fd9e98103ff804a43f2d4485f72ac48df261c8
(cherry picked from commit e68d358532)
Currently, OS_REGION_NAME is in keystone_admin rc file only when
keystone api is < 3. This is incorrect as region name is also meaningful
in newer versions and needed in some cases as shown in [1].
[1] https://bugzilla.redhat.com/show_bug.cgi?id=1716434
Change-Id: I45dc6511ee8874b9719be6816062777f2ce371f0
The 2.0.0 release introduces a requirement for using
python3.5 but we are running this in python 2.7
Also pining puppet-rabbitmq as commit [1] has broken compatibility with
puppet < 5 because of the usage of data in modules. I'm pining to the
same commit used in RDO for Rocky.
[1] cc78a33d98
Change-Id: I055ab99e48e018d041a2aef7c8998f72ebd4ff5e
(cherry picked from commit d2b1c67d57)
This is a mechanically generated change to replace openstack.org
git:// URLs with https:// equivalents.
This is in aid of a planned future move of the git hosting
infrastructure to a self-hosted instance of gitea (https://gitea.io),
which does not support the git wire protocol at this stage.
This update should result in no functional change.
For more information see the thread at
http://lists.openstack.org/pipermail/openstack-discuss/2019-March/003825.html
Change-Id: I8aeb0c5d40ea95fb14115e3d8aa22bdb3efb1fad
Currently, it points to master which is incorrect for CI jobs on
stable/rocky.
Also updating pins for puppet-redis and stdlib to match the versions we
have in RDO Trunk Rocky.
Change-Id: Ica13295b5464ed6477c9be8f33ee03181132df3e
"nova_metadata_ip" option is deprecated since pike in
favor of "nova_metadata_host", puppet-neutron accepts
both parameter since pike, so this patch set's the non
deprecated parameter for neutron metadata.
Closes-Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1649304
Change-Id: I1c045b46cb128a45219f90b4c71ae86936e8d959
(cherry picked from commit 952b0550f0)
The file ownership and group was set to root for
'/etc/nova/migration/identity' because of which
user gets Permission denied error while trying to
do resize or live-migration.
Changing file ownership and group to nova allows
to do live-migration and resize successfully.
Change-Id: Id46ea6b0f953e6e23cd95bae0889df906ae933b2
(cherry picked from commit 2750aab287)
r10k version 3.0.0 and later require a newer version of the cri gem,
which only work with ruby >= 2.3. This does not work on CentOS 7,
where we have an earlier version.
Change-Id: I99958dd82bb95b8f3d7a6a2b0c86fd2c1f21c802
(cherry picked from commit 4f8e95eccf)
Previously, we had a cron job to flush removed keystone tokens. Since
[1] this is not required anymore, but we need to add a cron job to
rotate Fernet keys.
[1] - https://review.openstack.org/544547
Change-Id: I331788ea08322a6f982c87eb195a619bab1c4d2e
Currently they are set using nova_config in vncproxy.pp but now
they have been properly implemented in puppet-nova [1].
[1] https://review.openstack.org/#/c/582881/
Change-Id: Ifec1dd09eb9ae1acb80001cb0ede1c5efb118d17
When setting service-workers we are setting that value for
workers and threads. This is different to what we do for other
services as aodh or keystone.
Change-Id: Ic6185f9781d164d33df6b26eea4ebf052a37e00d
Resolves: rhbz#1598043
[1] removed some deprecated parameters in puppet-neutron. Some of
them are still being used by Packstack, so let's remove them.
[1] - https://review.openstack.org/575976
Change-Id: I4b4beac2f4061689056c399e05d9dd5c6689e82f
service-workers default to $::processorcount which is
8 in nodes used in CI jobs, this leads to high memory
usage, setting it lower would help in reducing random
failures in scenario000 used by other projects in there
gate.
Change-Id: I10cb927db7648b384ce41ef4ae1050a5d4ed8e9e
puppet-gnocchi removed the deprecated params in [1],
this patch cleans up in packstack.
[1] https://review.openstack.org/#/c/569532
Change-Id: I53ccc22c04c9aa430ee3e037dd8f7b46d1366295
According to Openstack summit session [1] stestr is
maintained project to which all Openstack projects
should migrate.
Let's switch it then.
[1] https://etherpad.openstack.org/p/YVR-python-pti
Change-Id: I2fdbbafb9ec36c7bcbab6038387aaa31f41316dd
Signed-off-by: Charles Short <zulcss@gmail.com>
rpc_backend setting is taken care by default_transport_url.
puppet-* project are removing rabbit_params in [1], let's
remove the usage from packstack.
[1] https://review.openstack.org/#/q/topic:remove_rabbit_params
Change-Id: Ic08cb1aa5284c127d290868bba72ebdc2c08fe4c
puppet-nova has removed the following deprecated options
from nova::api class in [1]:-
- default_floating_pool
- aliases
Switch to ::nova::network::neutron and ::nova::pci to
set them.
Default for default_floating_pool will be taken care by
puppet-nova.
[1] https://review.openstack.org/#/c/569540
Change-Id: Ifc425070bff473b388bd20cb5cea446cf4862c59
Until https://bugs.launchpad.net/packstack/+bug/1771500
is fixed, temporary switch to ovs in scenario 001.
Also, set inject_partition to the default value of -2 in nova.conf.
We are having problems in the gate in OVH nodes (see [1]), and
doing this prevents us from using the problematic libguestfs code.
[1] - https://bugs.launchpad.net/packstack/+bug/1701032
Change-Id: Ifed85df073104684cd415792785acd12e2f89aa4
Related-Bug: #1771500
Co-Authored-By: Javier Peña <jpena@redhat.com>
The baremetal scheduling options were deprecated
in Pike and the ironic_host_manager was deprecated
in Queens and has now been removed. Deployments
must use resource classes now for baremetal scheduling.
Depends-On: https://review.openstack.org/565805/
Change-Id: Iea17567402360a49fa9c89845916bc09a2b4e2ad
We can move to the current master repo for puppet-rabbitmq, since
this is what we are testing in RDO. This requires a few changes:
- Update the URL used in the Puppetfile, and set the branch to 'master'.
- Remove old workarounds in amqp.pp
- Set loopback_users properly, based on what is requested by [1].
[1] - 0ada399b33
Change-Id: I9262836eb05a47bfe78ea0ac6d1c04199f29f527
Previously, Packstack wasn't deploying metadata service for OVN as
it didn't exist before Queens.
This patch is adding support to this service on compute nodes and
disabling config_drive.
Depends-On: I278ad5521c9529e5bd71f712c44c6a69f485d1b6
Change-Id: I6e169dea7b4a9d0d64bd61a226641b591eaf1da5
Co-Authored-By: Miguel Angel Ajo <majopela@redhat.com>
Co-Authored-By: Alfredo Moralejo <amoralej@redhat.com>
Co-Authored-By: Javier Peña <jpena@redhat.com>
Currently, test scenarios 001-003 will only look for x86 cirros
images. This patch adds support by grabbing the system arch and
creating a string to match the downloaded cirros image in
run_tests.sh.
Currently, ppc64le and x86 will use different cirros versions,
as x86 experiences kernel panics with 0.4.0, which power needs.
Change-Id: Ibc9c2f4d9d8d34e416b3b57cfde0a51d0917e689
Currently run_tests.sh hardcodes what images are downloaded and used
for testing. This patch grabs the system's arch and uses it to build
the urls and filenames for the cirros images. Additionally, ppc64le
images are not provided until 4.0 so switching cirros version based
on arch is also added here.
Change-Id: I73b65c566153c027988b5781870e44c47c6e8b1f
Fix https://review.openstack.org/#/c/468252/ has been merged in cinder
and proper default configuration parameter is defined in puppet-cinder
so we don't need to set this parameter anymore.
This reverts commit f72ccac503.
Change-Id: I353b73ee351b5ef5c239d882cbc5477ae471391f
We now have puppet_forge 2.2.9, r10k 2.6.2 and semantic_puppet 1.0.1, which do not create any incompatibility.
This reverts commit edbf361c7c.
Change-Id: I54ccdd24ce3717feaf4600c4cb4df38bae8b9b06
Previously, the heat dashboard was part of Horizon, but this was changed
late in the Queens cycle, where a separate heat-dashboard repo
was created [1]. We need to install it if we want to keep the previous
functionality.
[1] - https://github.com/openstack/heat-dashboard
Change-Id: I253071e323c3e3a9cadabb8724cb686b6911322a
This commit will bump the major version number reported by pbr, so
the Rocky versions have a higher version.
Change-Id: I4698a01f49d092f02dd39a8a4ed1ed74ef480883
Sem-Ver: api-break
This has been removed by the Keystone project as of [1].
[1] - https://review.openstack.org/543060
Change-Id: I29c08c958ed83b4245f20e73f4fec4841635e85d
* https://review.openstack.org/#/c/504345/ moves the tempest run
cli to stestr. In order to avoid breakage switch to stestr from testr.
* Check for .testresposity and .stestr folder and then use
respective test runner. It will avoid the breakage in CI.
Change-Id: I1fdbe2f441a73f80d8d6e7a8263d300939a4fba3