The values of these dictionaries are not used when SDK is in use,
which should soon account for all use cases. Eventually we should
probably look for plugins to return a proper class or typeddict but
that's a job for another day.
This began as a fix for in openstackclient/object/client.py which
referenced a non-existent class and quickly snowballed.
Change-Id: I7b807ec3a97124b35828ffdecbb36f6fde11e7b5
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Quotas details returned from the Neutron service are in different format
then quota details returned from Nova and Cinder services. This patch
fixes helper function to convert data from Neutron to the same
format as data from Nova and Cinder is given.
Closes-Bug: #2102513
Change-Id: I18649f6c2ee179b64b7e605f4ea07d4b0c7a1635
The fix ensures that if a user wants to set a default project,
they must also provide the project domain. If it's missing,
an explicit error message is shown, making it clear that the
project domain is required.
Also adding some unit tests by modifying respective calls.
Change-Id: Ia6e921a53da55ab1bce85a42c8160872a9d47d64
Closes-Bug: #2102146
We may need to remove this soon enough, given the new Docker rate limits
that we keep bumping into.
Change-Id: Id4a9d8df770d107986b20e4a98835ee4e0b6117d
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Add file to the reno documentation build to show release notes for
stable/2025.1.
Use pbr instruction to increment the minor version number
automatically so that master versions are higher than the versions on
stable/2025.1.
Sem-Ver: feature
Change-Id: I2789e8605f5e9bae63382ca9e822bbc3e2241f36
the flag should store False to 'allow_failure', not True.
Also, make the --allow-failure and --disallow-failure flags
mutually exclusive.
Change-Id: I03699e14d4d69d9f08caab647293732fc211dbad
This change fixes missing space in help messages to make those
consistent and pretty rendering when calling help.
Change-Id: I947374821a4dbb5e68651c0e72fc5fd2f938e6a1
openstacksdk's Credential expects user_id and project_id, not user and
project. Previously, we would send payloads like
{'type': 'ec2', 'blob': '{"access": "s3-user1", "secret": "s3-secret1"}'}
which Keystone would reject with
'user_id' is a required property
Change-Id: I0544bef7df9247395f0726ea075112d6ac992252
In [1], the "port list --long" command received a new column, showing
the trunk subports related to a parent port. The problem of this patch
is that the ``_formatters`` definition, that is shared with the "port
show" command too, is changed. The "trunk_details" information presented
in both commands differ:
* In the "port list" command, only the subports are presented, in order
to print a list of dictionaries without showing the ``trunk_id``.
* In the "port show" command, it is presented all the trunk information,
including the ``trunk_id``.
This patch includes functional tests to validate the fix.
[1]https://review.opendev.org/c/openstack/python-openstackclient/+/926611
Closes-Bug: #2098950
Change-Id: Ib1107fb3dbb025b39a7c55f90f5fe51ae433a72f
The '--flavor' option appears in the usage and arglist but is not
actually parsed. The '--flavor-id' option is what is silently parsed.
Since the goal is to allow the name or the id, this adds the '--flavor'
option to being parsed.
Closes-Bug: 2091731
Change-Id: Id83facd3825f472e7d864427699bd072d1c08779
Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
We need such temporary workaround until [1] in SDK will be merged, as
without that change here py{39,312}-tips jobs are failing on that SDK
patch.
[1] https://review.opendev.org/c/openstack/openstacksdk/+/939703
Related-bug: #1915151
Change-Id: Id39a6482de54fe78e26fa33c9252253886cf1f3d
Neutron API is accepting 'security_groups' field in
order to return the list of security_groups attached
to a port, but openstackclient is parsing the output
over a Openstack Port object that has security_group_ids
to map. This patch sends to the Neutron API the expected
field value and replace the output key to allow the
mapping just in case '--long' argument is passed.
Closes-Bug: #2095414
Change-Id: I188edc3c620ce29d7b16497ca24fd7d972a06618
creating the image is a 2step process, first an 'empty' image is created
and then the data is uploaded.
Currently the output of the 'image create' command is that 'empty'
image, in `queued` status etc.
A more user friendly approach would be to make a second refresh call
to show the user image as it is after data was uploaded.
Change-Id: I2f78b113dcc3c941f8cf8dd9b63262971a780a39
