Update the code so that "openstack server list --all"
also displays the names of private flavors.
Change-Id: I4804fcd905eaf67b1ad9b461084eaf0caa820d2f
Closes-Bug: #1742453
The base job has tox_install_siblings: false - which we want. But
that means we need tox_install_siblings: true on the tips job.
While we're at it - add fetch-tox-output so that we have tox log files
in the fetched build output for easier verification of what wound up
installed.
Don't look for ResourceNotFound string in test
The string ResourceNotFound is not in the error string anymore. Look for
the text that is.
Depends-On: https://review.openstack.org/541033
Change-Id: Id6de1485bcafb41f238f3e74277094ce64a6acf4
Zuul no longer requires the project-name for in-repo configuration.
Omitting it makes forking or renaming projects easier.
Change-Id: I92ea13e64dcdc24a8dd1fd6d17df9f771b2fc1e0
The find_ip from openstacksdk started being usable by OSC back in 0.9.15
but the local method never got replaced.
Change-Id: I18a334280e5f384f8bb96198cdad79c612a02290
Change I7e8c47dead1e019e8705db3ff7559dd39b1d90d9 broke
the ability to create a new release note from the venv
tox target because the reno requirement was moved from
test-requirements.txt. This fixes it.
Change-Id: Ifac83121388728fa445e0ed4433c9d981c057737
We store the created conn on the instance, but we never pull it back off
if there is already one present.
Change-Id: I2d890dd206d4ddf67fa42d798e6fd2c652799785
network.client.make_client() has always put a copy of it's SDK Connection
directly into ClientManager, the new-style Connection create will move
into osc-lib ClientManager, do it here too until then.
Change-Id: I1edfd19c9e73320768fb9640931fafe857c980b4
a piece of code in image client has some business logic behind
assert, which can be lost when running python in optimized mode (-O).
Change-Id: I2179970df495e1215d691915c51cebe5cb4541a7
The zuul job figures out whether to run "setup.py build_sphinx" or
"build-sphinx" based on whether the project is relying on pbr's
ability to auto-generate API reference docs. Because we are relying on
that, we want local builds to use "setup.py build_sphinx".
This reverts commit 42e0037cd055d81d9452f046c347f8e469d27350.
Change-Id: Ia01188110fa9c3ccaf3d794fa3a511a6780e683e
We have consolidated the fetch output roles into one
fetch-subunit-output, replace useage of old roles with new one.
Depends-On: I0cdfc66ee8b046affeb0b071fef38c21cb7a4948
Change-Id: Iae2892d9b4cd870a11579434edc9ee66bd16798c
Now we can associate a qos policy to the floating IP, and
dissociate it. The commands are:
$ openstack floating ip create --qos-policy ...
$ openstack floating ip set --qos-policy ...
$ openstack floating ip set --no-qos-policy ...
$ openstack floating ip unset --qos-policy
These commands are based on the neutron change:
I4efe9e49d268dffeb3df4de4ea1780152218633b
Partially-Implements blueprint: floating-ip-rate-limit
Change-Id: I932b32f78cc5a2b53926feaec1a0b392cf7e8b57
Now if Glance v2 cannot find image data it returns an empty
response with 204 status code, instead of raising an error.
Glance client handles this situation and wraps the response
with a RequestIdProxy object, whose 'wrapped' attribute is
None.
But when openstack client tries to parse this object using
glanceclient's save_image util function, it fails with
"NoneType object is not iterable" message, for the
object doesn't contain any data.
This patch adds additional check to prevent such behaviour
and raises SystemExit exception if no data was returned from
the server.
Glance v1 is not affected, because it raises an error if can't
find an image data.
Change-Id: I016a60462ba586f9fa7585c2cfafffd7be38de7b
Closes-Bug: #1741223
Right now, if a neutron port is owned by a container powered by
Kuryr, there is no way to list and filter those ports because
OSC assumed a neutron port is owned by either a server or router.
This patch adds support for that by introducing an option '--device-id'
to the 'port list' command.
Change-Id: Ib1fd27e8d843a99fb02ccabd8a12a24ac27cec9c
--swap will add a additional storage device,
which not affect the original swap partition/device.
This patch will clarify this misleading description.
Change-Id: Ic079c069985d39cc969b97876901007a81883f57
Signed-off-by: Chen Hanxiao <chenhx@certusnet.com.cn>
This commit makes to use stestr instead of ostestr
directly. ostestr>1.0.0 has started to use stestr instead of
testrepository. So there is no reason to use ostestr anymore.
Change-Id: I6327d50c9f6dd19f1de24b9b51532104fb3e916e
Our functional tests depend on OS_CLOUD (or other OS_* envvars)
to retrieve authentication information.
Functional test failure is caused by the lack of OS_CLOUD envvar.
This commit updates the job playbook to pass OS_CLOUD.
Change-Id: I903ce599082cc923f02e26a2058bbfa7eb9bb2d6
Per API reference, only 'changes-since' is accepted and the variant
with underscore is ignored, making the CLI functionality broken.
[dtroyer] added release note and fixed unit tests.
Change-Id: I0c596531a8af03da17d5ce39d75b12e941403aa5
Closes-Bug: 1732216
Do a dummy import to determine which SDK is installed (Pre/post merge).
This solves the DevStack error "Cloud defaults was not found" in -tips jobs.
Depends-On: Ia111f127fbdceac2afe20fd9d1fe032145cdd72c
Change-Id: I60c2d418dd5a393eee2cc2a5c2fdebfffdabf2d3
