Depending on the way the testing script is called, it may sense
stdin being available and not being a tty [1]. The logic in the
Glance part of the client is then broken [2] as it detects both
volume and stdin as sources of the image and fails.
The proposed approach uses the script module for these tests.
This fixes the cephadm upgrade jobs where this script was called
using the shell module which causes this behaviour. The script
module is not affected. FWIW, the command module is affected
the same as shell because shell extends the command module.

[1] https://opendev.org/openstack/python-openstackclient/src/commit/6810414e45a32dd44263dff47fec161989508ef0/openstackclient/image/v2/image.py#L114-L120
[2] https://opendev.org/openstack/python-openstackclient/src/commit/6810414e45a32dd44263dff47fec161989508ef0/openstackclient/image/v2/image.py#L414

Change-Id: I3ce4059063e3a51eb1c5b42bda79893b56bd85f0
Story: 2006294

Follow-up to [1].
This unbreaks CI which was broken by a recent change to
master upper-constraints. [2]

[1] I82c4ff3b246c7db65ac8660b096d4a9a15dda8d7
[2] https://review.opendev.org/c/openstack/requirements/+/843722/3/upper-constraints.txt#b244

Change-Id: Idc7a893f53c6975cc4a3f6090020d973fa2288bb

Change-Id: Id8573ad6727c62e355923b787c51766bc27ed848

Per Clark Boylan's feedback [1], retries cause a retry not only
for pre playbook failures but also for cases where Ansible detects
network connectivity issues and they are caused by disks getting
filled to their fullest. This is an issue we experience that
sometimes results in a POST_FAILURE but certain FAILUREs are
retried which wastes CI resources.
The problematic jobs are ceph jobs. They are to be looked into.
Backport to all branches.
We can adjust retries for the core jobs that do not exhibit the
nasty behaviour but first we can try running without retries
to measure the troublesomeness.

[1] https://review.opendev.org/c/openstack/kolla-ansible/+/843536

Change-Id: I32fc296083b4881e8f457f4235a32f94ed819d9f

Docs and reno included.

Change-Id: I5099b08953789b280c915a6b7a22bdd4e3404076

Change-Id: I44c8a059918918f5bdc1c0190110a05d8d7f8f2d

Closes-Bug: #1975598
Change-Id: If4c85f8e960141d08a89accdc11a3271f31974c1

From:

(kolla) 13:11 (s) marcin@puchatek:kolla-ansible$ kolla-genpwd
Traceback (most recent call last):
  File "/home/marcin/.virtualenvs/kolla/bin/kolla-genpwd", line 8, in <module>
    sys.exit(main())
  File "/home/marcin/.virtualenvs/kolla/lib/python3.10/site-packages/kolla_ansible/cmd/genpwd.py", line 135, in main
    genpwd(passwords_file, length, uuid_keys, ssh_keys, blank_keys,
  File "/home/marcin/.virtualenvs/kolla/lib/python3.10/site-packages/kolla_ansible/cmd/genpwd.py", line 59, in genpwd
    with open(passwords_file, 'r') as f:
FileNotFoundError: [Errno 2] No such file or directory: '/etc/kolla/passwords.yml'

To:

(kolla) 13:17 (s) marcin@puchatek:kolla-ansible$ kolla-genpwd
ERROR: Passwords file "/etc/kolla/passwords.yml" is missing

Change-Id: I18a9559daeb3d124a03dcb735ebb01a2cf24f617

Change-Id: Ic55a7384c298ef517aa4b7f0eaa67d53b5d70df0

Change-Id: Ib4b15ed4feac82d8492b1c0f0238a752eac668e6

Since enabling libvirt SASL authentication, the masakari instance
monitor fails to connect to libvirt. We see the following error in logs:

    libvirt.libvirtError: authentication failed: Failed to start SASL
    negotiation: -4 (SASL(-4): no mechanism available: No worthy mechs
    found)

This change adds support for SASL authentication in Masakari instance
monitor.

Depends-On: https://review.opendev.org/c/openstack/kolla/+/834456
Closes-Bug: #1965754
Change-Id: I974046662b383a12ac6281b725523760a96657bd

It was removed as part of improvements in [1]. But it seems it's
a different type/class of errors and should still be filtered
out here.
This time for all the services to avoid extra copies.

[1] Ie6a95c3ca84df82ca8463cf76054e125cb1ffff1

Change-Id: If7c5b938dc7417831a90f360d6e3655d1ad4084e

The projects started dropping support for Python 3.6 and thus
CentOS Stream 8 which K-A tests against, resulting in broken CI.
This change ensures that we keep upper-contraints also on Yoga as
we do for the images we deploy.

Change-Id: I82c4ff3b246c7db65ac8660b096d4a9a15dda8d7

According to OVN SRIOV docs [1] - neutron-ovn-metadata-agent needs
to be deployed on network nodes (OVN chassis).

[1]: https://docs.openstack.org/neutron/latest/admin/ovn/sriov.html

Change-Id: I111df79f08eb76a735f33c6edf03b5a302fb6cc3

Per [1].

[1] http://lists.zuul-ci.org/pipermail/zuul-discuss/2022-May/001801.html

Change-Id: I2197b5436a4b4d4220c916d2f3129f49f13b70ec

In the last PTG it was decided to drop the keystone_token_provider variable, because there is no other option anymore.

Signed-off-by: Ramona Rautenberg <rautenberg@osism.tech>
Change-Id: I1ee2c3f9b7dbbbf4633c5874cdbb3c4f8c09e277

Redis Sentinel modifies its config.
Since we now check config in check-containers,
the redis_sentintel container is restarted each time.

This patch together with kolla patch fixes this issue.

Closes-Bug: #1967664
Depends-On: https://review.opendev.org/c/openstack/kolla/+/840954
Change-Id: Ia76329cb851d266cd888f51b7f70a4cdc40e3ac3

"Smoke tests" for barbican, cinder, glance and keystone have been removed as discussed in PTG April 2022.

Signed-off-by: Tim Beermann <beermann@osism.tech>
Change-Id: I613287a31e0ea6aede070e7e9c519ab2f5f182bd

Add an enable_cinder_backend_pure_iscsi and
enable_cinder_backend_pure_fc options to etc/kolla/globals.yml
to enable use of the FlashArray backend.
Update the documentation to include a section on configuring
Cinder with the FlashArray.

Implements: blueprint pure-cinder-driver
Change-Id: I464733f1322237321ed1ffff8636cf30bd1cbb38