Currently sometimes the check for ContainerCreating status is run too early,
Helm based services are not even spawning Pods yet.
Also improve wait time for ContainerCreating
Change-Id: Ifd08eccdbc15446e3f4e883712c5035df3b9e370
Although we didn't change policy of magnum service,
but we should add this to complete test scope of job
`magnum_tempest_plugin.tests.api.v1.rbac`.
Change-Id: I5dfbad43d5a515312ef285aff208e1c9e36cf684
Cluster creation fails on the CI due to an issue similar to what we've
seen before for another test (test_create_list_sign_delete_clusters).
Disable it for now.
story: 2007272
task: 38867
story: 2007851
task: 41443
Change-Id: Ie91fdfe71058b45c5fd1217920b8b5b993e47734
This is kind of a race condition issue, Tempest is doing a wait and check
when it is trying to delete a cluster. However, at the very end of the cluster
deletion process, a cluster GET request may get a HTTP 400 error which
will cause the Tempest case failed. This HTTP 400 exception reported
by a cluster GET request should be caught and take cluster as still existing.
Task: 40120
Story: 2007833
Change-Id: Ic97abf8d4d584f35b18bff9eab5d3e5ac80c0f02
Update docs building, cleanup a bit:
* Switch to sphinx-build
* Update requirements for Sphinx and openstackdocstheme for
python 3, create doc/requirements.txt
* Remove unneeded doc and translation sections from setup.cfg
* Remove install_command, it's unneeded, the default is fine.
* Move constraints into deps, use TOX_CONSTRAINTS instead of
obsolete UPPER_CONSTRAINTS
* Update to hacking 3.0, fix problems found
Change-Id: I76fe5602928bede439133993b0d85669228f25ed
In case the cluster could be deleted immediately after the creation
failure, there is no chance to see the reason for debugging purpose.
Change-Id: I627f5b8bef3bc8fef694837cd7f037dc01885b53
- This fix is primarily focused on using existing cluster template
use case.
- Rename 'keypair_id' to 'keypair_name' in the magnum tempest config.
- Add 'cluster_template_id' and 'cluster_creation_timeout' config
options.
- Move the cluster template and keypair to the class-scoped resources.
Change-Id: Ib5260757e373ce545ca6900acf8871b6d96dcfb8
The plugin defines a 'magnum' option in the '[service_available]' group
but does not respect it, so if the plugin is installed but
[service_available]/magnum=false, the tests will still run unless
blacklisted. This patch corrects the issue by defining the tempest
skip_checks class method. This also requires using the correct tempest
base class as recommended in the tempest guide[1].
[1] https://docs.openstack.org/tempest/latest/write_tests.html
Change-Id: I6b9a8f802129fa800d9f4a3eadd8943962c951c0
This patch is going to fix the k8s test gate for Magnum:
1. Given the test case test_create_list_sign_delete_clusters is the slowest
test case, we better add the Tempest 'slow' tag for it.
2. Raise the timeout from 1800s to 3600s
3. Fix a wrong method calling
Change-Id: I074434b3100178cba19b1f2f20da792871bb8530
* It provides proper directory structure and cleanup of
magnum tempest plugin tests.
* Added tools/pretty_tox.sh to trigger functional api tests
Change-Id: Ibdddd26da9cfb0d08c2977660320b2c052d7261b