This patch aims at fixing the broken gate jobs because of cinder and glance patches [1], [2], [3] and [4]. * Remove parameter `--source-replicated` to drop volume replication v1 support * Address some timing issues with volume transfer requests * Only run Image v1 tests when the test cloud has v1 available * Get tolerant of unexpected additional attributes being returned in Image data [1].https://review.openstack.org/#/c/586293/ [2].https://review.openstack.org/#/c/532503/ [3].https://review.openstack.org/#/c/533564/ [4].https://review.openstack.org/#/c/578755/ Co-Authored-By: Dean Troyer <dtroyer@gmail.com> Co-Authored-By: Monty Taylor <mordred@inaugust.com> Depends-on: https://review.openstack.org/588664 Change-Id: I2a785750e92155185d3344e6116c7f5c6fdd3cbe Signed-off-by: Fan Zhang <zh.f@outlook.com> (cherry picked from commit def83a0e94de2f98e3bd68ee412c0f0a2c316f32)
When creating a volume from a snapshot, the size parameter is required and type is checked. Since we have to pass something and it needs to be a valid data type (None is not acceptable) grab the size from the snapshot object and pass it. Change-Id: Ie23e3d23828919234e40336b5c65b22e140d337c
This patch implements an endpoint lookup when showing limits. This addresses the issue when showing limits without both Nova and Cinder and will display limits if one is missing. Change-Id: I2214b281e0206f8fe117aae52de2bf4c4e2c6525 Closes-bug: #1707960
In reviews we usually check import grouping but it is boring. By using flake8-import-order plugin, we can avoid this. It enforces loose checking so it sounds good to use it. This flake8 plugin is already used in tempest. Note that flake8-import-order version is pinned to avoid unexpected breakage of pep8 job. Setup for unit tests of hacking rules is tweaked to disable flake8-import-order checks. This extension assumes an actual file exists and causes hacking rule unit tests. Change-Id: I12b596820727aeeb379bee16c2bc993dee9eb637
[breakage related to os-client-config 1.28.0] os-client-config 1.28.0 add a check if filebased and envvars are both used. This check causes OSC unit test failure. OSC now instantiates OpenStackConfig twice as a workaround. The unit test mocks _load_config_file() and it returns a config dict, but os-client-config OpenStackConfig.__init__ updates the dict returned. As a result, when OpenStackConfig is instantiated second time, the mock of _load_config_file returns a modified version of the config dict. This hits the new check in os-client-config 1.28.0. This commit changes the mock to use side_effect rather than return_value to ensure the original dict is used. [breakage related to osc-lib 1.7.0] The change in osc-lib 1.7.0 added "if" logic to avoid calling get() twice. In tests.unit.volume.test_find_resource, kwargs is empty dict in find_resource(), so the second call to get() is NOT called now. Removing the second elements of side_effect addresses the unit failure. Co-Authored-By: Rui Chen <chenrui.momo@gmail.com> Change-Id: Ib9d14661b2755bbd6619e15c0d9023fbc9d27d70 Closes-Bug: #1703782 Closes-Bug: #1703783
Current "volume list --name" command use "display_name" as search_opts to send to cinder API, and show the result table with "Display Name" column title in osc, cinder list API support "name" as search opts too, and there is "name" attribute in volume response body, so we can replace all "Display Name" by "Name" in order to keep "volume list" command consistent with other commands, like: server list, network list and so on, only use "Name" attribute for all objects. Support a mapping for volume list -c "Display Name" (Volume v1 and v2) and volume create/show -c "display_name" (Volume v1) for minimal backward compatibility until R release. Change-Id: I120be0118e7bb30093b4237c5eeb69a9eedef077 Closes-Bug: #1657956 Depends-On: I1fb62219b092346ea380099811cbd082cae5bafe
This has been sporadically failing in functional tests due to the way the volume qos spec list command calls get_associations() for each spec. When tests run in parallel occasionally a spec from another test is present in the list returned and is deleted before the get_associations() call is made, causing a NotFound exception. We should just keep going when this occurs. * make v1 match v2 * add tests to ensure the exception is being caught and handled Closes-Bug: #1687083 Change-Id: If2d17c1deb53d293fc2c7f0c527a4e4ef6f69976
In volume snapshot command, <volume> is the same as <snapshot-name> when --volume is not specified, but <volume> cannot be None, so when <snapshot-name> is not specified (<snapshot-name> is None), a NoneType error appears. So make <snapshot-name> no longer optional, it should be always present. Change-Id: I3d9f10753a8ef601e70816421c160598e2cc811f Closes-bug: #1659894
* Fix volume transfer request accept to actually not crash when trying to call Volume API. * Fix volume transfer request accept syntax to have only one positional argument, which is the ID of the resource in the command * Change the output column order in volume transfer request list to have ID followed by Name then the remaining columns. Closes-bug: 1633582 Change-Id: I5cc005f039d171cc70859f60e7fe649b09ead229