The endpoint discovery mechanism in compute, volume, and share use
no_port_cut_url to cut the URLs after the version (including the
version) if port is not present. However, when the port is present, the
no_port_cut_url method sets the top_level to True when calling do_get
method. This causes the do_get method also cut the top_level_path of the
URLs because by default the top_level_path is set to "". This behavior
is not desired because different clouds maybe have endpoints with
different top_level_path other than "". Instead of using
no_port_cut_url, we can use utils.get_base_url to properly cut the URLs
after and including the version.
Story: #2010968
Task: #49109
Change-Id: I221b6267afce90b8b7e22d468f0824c9365f4a91
We didn't take into account that the admin role and user might be
called on with uppercase first letter - Admin - on some systems.
The patch edits the logic so that we accept this admin name or
role.
Story: #2010880
Task: #48601
Change-Id: Ie027de58f77a58f44b31274e6d75dc830d3b717d
During cloud discovery resource creation, the domain_name and
user_domain_name information used to be hard coded to 'Default' which
might not be the case for all credentials, admin or non-admin.
This patch adds support for domain_name, project_domain_name and
user_domain_name information by defaulting all of them to 'Default'
however in the case they are available in the cloud credentials, using
them during python-tempestconf runtime as well as rendering them
correctly in the final tempest configuration.
Story: 2010811
Task: 48330
Change-Id: I4355837f4d8d514524039df6f3989dbd69f2c700
tripleo-ci-centos-9-content-provider has been failing for some
time and the other tripleo-ci-centos-9 jobs depend on it.
Based on the recent builds of the jobs outside of
python-tempestconf project it seems they aren't executed on master
anymore.
Due to boh of those reasons this commit drops those jobs.
Change-Id: Ibc7e3a0dd6b661d9c615e087921cc9dd426382a9
The tests require a specific dns setting in tempest.conf
pointing to the dns server:
dns.namservers = <ip>
This is complicated to set in our jobs, therefore the patch
excludes the tests.
Moreover, this patch [1] made everything install in global
venvs, however, centos-9 isn't yet compatible with that
and the jobs are failing now.
This patch disables installing in global_venv on the
centos 9 job.
[1] https://review.opendev.org/c/openstack/devstack/+/558930
Change-Id: I99c8cbc20be60b2f9475950667ac72114dc298a9
Added task to install proper version
of tox(<4) so the job won't fail and
set the job as voting again. Set
ensure_tox_version to <4 for zed admin
job and admin plugins job (since it failed,
on tox>4 related issue too.
Story: 2010573
Task: 47319
Depends-On: https://review.opendev.org/c/openstack/tempest/+/884584
Change-Id: I0e9ad3632d16a5aa8c0e98d3e0d10bee79cead99
We query the volume pools to check if the user has permissions to list
the backend storage, but we do not handle when the volume backend is not
available, only when the user is not allowed. This patch adds the
ServiceError exception to handle it and set the backend properly.
Change-Id: I0fd61813dbadf5ef3a0db55791e3b261d4c9d885
The commits marks several jobs as n-v to unblock the CI. The jobs
already have fixes proposed and are in the process of review.
Right now we wait on a change in Tempest which may take serveral
days to get merged.
The patch also fixes an issue with pep8 - too long line - seems
like the sensitivity of pep8 was increased because previously
the line was ok with pep8.
Change-Id: Icea222faadfe22ab029d0321fd68c3e6f1bbc0b4
A new tag has been released recently. This commit tags all
releasenotes written until the 3.4.0 tag and groups them
together.
Change-Id: Ic4ae5afa91b0942fe8d821177a8ab260129b6473
Live migration supposed to work in all qemu/kvm deployment.
It is expected switch to block magrate in case of no shared storage,
tempest expected to skip migration tests when there is no multiple
compute nodes and it would lead to test failure.
Also update the documentation regarding the options set by default.
Change-Id: Id9a25ad50df85f16f27a228a0075c2e073670937
This commit changes the default value of the following option
from
CONF.auth.tempest_roles _member_
to
CONF.auth.tempest_roles member
Change-Id: Ic8263d0bfd9b093aab7f1779d24b19ca2a9b30bd
We can remove test_manually_trigger_update_secondary_zone_negative
test from exclude_regex as bug 968717 was fixed by removing the
test from the plugin [1].
[1] https://bugs.launchpad.net/designate/+bug/1968717
Change-Id: Ie8d0cd289302b789d48c925c1ab0b9940a6e09ac
With the upgrade of tox to a newer version we started to encounter
failures in the opendev-tox-docs job. It looks like the interpretation
of the variable skipdist changed.
A similar problem was solved in tempest repository by removing the
skipsdist option from tox [1]. It looks like a similar solution
works for python-tempestconf as well.
[1] https://review.opendev.org/c/openstack/tempest/+/868527
Depends-On: Ie5ea7f9305c7e7611f205d2e0703849bd6321a68
Change-Id: I9ee6570b72c10ea0466819008135ddebe891a445
- adds antelope zuul template
- adds stable/zed jobs
- adds openstack-tox-py39 jobs as its not part of antelope template
- removed all victioria, wallaby and xena jobs
- docs, yoga and tripleo-ci-centos-9-standalone jobs are temporarily
marked as n-v to unblock the gate
Depends-On: https://review.opendev.org/c/openinfra/ansible-role-refstack-client/+/865522
Change-Id: I31b23b077bd2dc91524fde16b08de7aed1ca4c29
TempestConf has a set of class variables that do not need to be
class variables because there are no multiple instances of
TempestConf class that could share the value.
The reason why we need to move the class variables to instance
variables is that some unit tests under certain conditions were
failing because multiple unit tests were changing the
TempestConf.priority_sectionkeys value.
Change-Id: Ic6f51be112aa9f93904fcab289d78a7e06f20f62
Currently tests from designate-tempest-plugin fail when the
[dns].nameservers value is missing in tempest.conf. The failure
is caused by this patch [1].
This patch removes the designate tests from the admin-plugins job
for now. Once we fix the issue (either by fixing it in
designate-tempest-plugin or in python-tempestconf) we can
add the designate tests back.
[1] https://review.opendev.org/c/openstack/designate-tempest-plugin/+/860116
Change-Id: I6a4ca43db2ffe602318141b0981f287d15cda3ac
When you have several controllers, if the first controller have
cinder-backup service disabled, the pacemaker return it as down, and so
the cinder-backup is set to False, even though it is running in a
different controller.
This patch do a loop to check if the cinder-backup is enabled in other
controllers, if so return it as True.
Change-Id: I3a16ec4f7b45267fcb8f9c6d01f28fe40faf7661
Since some of backends support thin_provisioning,
it's necessary to have a new parameter for skipping
the test in unsupported environments.
Change-Id: I9164a9f9d7d0dc913a1b91eb0efa741db8d523b8
packstack-centos8s job has been superseded by packstack-centos9
one. This commit updates the parent of our packstack jobs.
The commit also uncomments and puts a few jobs back to voting
after the CI crisis has been resolved.
Change-Id: I9ba1859637c14a86496afdc8eb0bf55a6794a495
Manila's devstack plugin can be told not to install
the manila tempest plugin hostwide. Installing the
plugin host-wide isn't reqired and would cause issues
with requirements for other packages.
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
Change-Id: I5820671240a21c3c795082791400463011a73b12
python 3.6 and 3.7 are dropped (upstream tooling as e.g.
Devstack doesn't support this those anymore) together with
train and ussuri jobs.
refstack-client jobs are non-voting now until a fix on
refstack-client's side has landed. Also victoria jobs are
marked as non-voting because they require a patch on devstack's
side which isn't ready to be merged yet [1].
manila-tempest-plugin needs to be pinned in Victoria and
Wallaby because the plugin requires newer tempest version
than is available there.
The patch also makes neutron version discovery with a not
top level URL, seems like something changed because
the discovery has been failing.
[1] https://review.opendev.org/c/openstack/devstack/+/844774
Change-Id: Ia3d78e84ad9eee515eb8ca017ea84df1fd160929
The job was failing because of a bug described in the launchpad [1].
As this bug is resolved and the job is no longer failing we can
start using this job again.
Before enabling the devstack-admin-plugins job we need to fix two
things:
- Put test_manually_trigger_update_secondary_zone_negative designate
plugin test to exclude_regex as it is currently broken.
- Allow git to read plugins directories. This is necessary because
of a patch that was recently added to git [2].
[1] https://bugs.launchpad.net/neutron/+bug/1959125
[2] 8959555cee
Change-Id: I5bf3826f0011a9392a38ceceb51c113d38231ad6
The test class had hardcoded DEFAULT_IMAGE_DIR value to /img/
which resulted in unit tests failures when this path existed
because the tests expect a call to makedirs which didn't happen
as the DEFAULT_IMAGE_DIR already existed.
The patch fixes that by passing exist_ok=True to makedirs call
which will result in always calling makedirs, whether img_dir
exists or not.
Change-Id: I47d15747536394006f6b502b3025220a5077e547
Using a deployer-input file that includes object-storage-feature-enabled
section without discoverability fails, although discoverability is an
optional parameter
Change-Id: I52838fc6b6f7e576604764e1b8b7938fe4b2ea0d
If there is no public network set, python-tempestconf doesn't
set network.floating_network_name option. In order to avoid
the following error when heat_plugin options are being set:
No option 'floating_network_name' in section: 'network'
the commit adds a try except block to catch this situation
and log it for the user.
Change-Id: Ib1f04584b620fc662fd415c4430a03ea92c35b76
Some changes have been merged to python-manilaclient recently
which aren't packaged yet and therefore cause problems in our
gates - the jobs are failing to create a share group type.
Installing the client from source will fix our gates.
Change-Id: Ia4d856118fa204f821c3ec0fdf39d210eaad34b1
Let's activate a few jobs which were commented out in order to
unblock the gates and merge necessary changes.
Change-Id: I07b923d48aab5908c48fcfdd3a3fb23207178be7