openstack/openstack-ansible-os_tempest - openstack-ansible-os_tempest - OpenDev: Free Software Needs Free Tools

Since ansible-core 2.10 it is recommended to use modules via FQCN
In order to align with recommendation, we perform migration
by applying suggestions made by `ansible-lint --fix=fqcn`
Change-Id: I19536716c7ac6bb75ae264db07ad9bc7d51d1ae6

In order to reduce divergance with ansible-lint rules, we apply
auto-fixing of violations.
In current patch we replace all kind of truthy variables with
`true` or `false` values to align with recommendations along with
alignment of used quotes.
Change-Id: Id65c3c3a2a36c28fc9fb9a90b151570e9fa694bd

In some cases it might be required to have shared network for tempest
test. A good example might be trove test that assume having that.
Change-Id: I359c7caafc515ff2db0989e04fb0bc33ab843ba8

In some jobs where we run tempest we do not have all required services
Good example of that are infra jobs where only keystone is deployed.
During conversion to usage of openstack_resources role, a condition for
glance images creation task was missed, that avoided image uploads
when glance is not present.
Change-Id: Ibcf16abab20eadf984ad1c8307ad14b43a889793

With efforts to create a resources in same, unified way,
we convert tempest role to use openstack_resources
for creating and managing openstack resources, like projects, flavors,
networks, images, etc. This should reduce maintenance costs
in case of futher collection updates and unify approach.
Depends-On: https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/878794
Change-Id: I8d7609fac96935ed51e92ebf58515e8eb0c44d5c

Due to typo in condition, we ended up in removing excludelist file right
after it being provisioned in certain circumstances.
Change-Id: I455f4e796fb8dcac29fe7947fc192baee9367187

Combine all variables starting with tempest_tempest_conf_overrides
to give a complete set of configuration to tempest when multiple
different test scenarios are enabled at the same time.
Change-Id: Iba5e061d1682d16c0516249f628f948a20580df8

If the role variable overrides are updated to remove a set of
existing tempest test exclusions, the previous exclusions file
remains on the disk even though it is not use, and not passed
to the tempest command. This is confusing and the file should
not be present unless there are active exclusions in place.
Change-Id: I5d69bf6258a00ade825cd3c746d1443dc1a35120

Previously the os_tempest role had a single variable for
defining the include and exclude lists in order to select the
tempest tests to run.
This works for simple scenarios, where a single service
is deployed and the tests for that service are enabled through
the necessary variables set in user_variables. This approach is
used in openstack-ansible CI / AIO.
More complicated scenarios such as magnum+barbican+octavia, would
create several user_variables files each with conflicting settings
for the test settings. It is possible in this scenario for there to
be no valid tempest tests to run and tempest to fail immediately.
This patch adds the possibility to have many variables defining
the include/exclude lists which have names using a common prefix.
Any variable names matching the prefix are gathered and combined
with the original role default to make extending the test lists
easy to do in an incremental/distibuted way in the ansible variables
instead of having to maintain a single point defining all necessary
tests.
Change-Id: Ie3a9a7be849171af042567ba8a152e5df5d2cb53

These variable names collide with the best name prefix to use when
gathering many variable names into a combined include/exclude list.
Codesearch shows no use of these as overrides and the number of
cases when an override would be needed is small.
Change-Id: Ic64165e41d24ae8dc75061589de84fa57998f03d

More suitable replacement variables have been available for
several releases now so we can remove the deprecated vars.
Tripleo still defines these deprecated variables in several
places but seems to have made no attempt to move to the new
vars, and tripleo is itself now deprecated so we should move
ahead with removal of these vars from os_tempest.
Change-Id: I5a4a90bc963acc8b44caf7eb060b763e0f90a50f

With update of ansible-lint to version >=6.0.0 a lot of new
linters were added, that enabled by default. In order to comply
with linter rules we're applying changes to the role.
With that we also update metdata to reflect current state.
Change-Id: Ifcb6ebfa971e324e447509e50cc7294bddd6a4a0

Ubuntu distro installation installs tempest plugins from source.
For this scenario we're ensuring that wheels won't be built as
repo container that should handle wheels build is simply absent in this
scenario.
For the rest usecases we're sticking with our new default to build
wheels.
Change-Id: Id643946edf4b4e21f8420d048ce8293fc40a61fc

Temporary code handling both cases until support for the 2.x
collection can be merged.
Change-Id: I200bf8d09d489be1433c9a269cc6cd6a1378a6cc

This line was introduced by I0230f1a93d16243fc50df79b2d0780e744706c7b
and should already be covered by the distribution_major_version line
above.
Change-Id: Ibffaadb721796d6f657972633d19bdd36e411c52

Currently when ``tempest_router_create`` is enabled, playbook requires
``tempest_public_net_create`` to be set.
It's an incorrect behavior. Router creation should only require public
network to exist. It does not matter if that public network was created
by tempest or operator.
Change-Id: Icaa223ed03837a68e9e84a89560cc2df10e3ed3e

The newest openstacksdk is set to rename some of the field names of
return values. This also affects the openstack.cloud collection. This
patch checks for both field names to make sure the os_tempest role
doesn't break when the collections release.
Change-Id: I4b30650a6c78e76982e22746933be7c132490a43

Tempest currently has two different internal methods for providing
authentication to tests: dynamic credentials and pre-provisioned credentials.
Depending on which one is in use the configuration of Tempest is slightly
different.
We should provide a support for both of them.
https: //docs.openstack.org/tempest/latest/configuration.html#credential-provider-mechanisms
Change-Id: I26d69caa3f96a530bc0a4a21365404b1a84e489a

In most cases, many of tempest resources are not needed.
From the other hand, some of our contributors still need them so we
can't just drop this functionality from this role.
The best solution I see is to add more flexibility and let users
choose which resources should be created by tempest role.
- public network:
 new variable: bool tempest_public_net_create (default: true)
 if false: require tempest_neutron_public_network_id to be set
 if true: require all other tempest_public_net_* variables to be set
 explanation: it'd be useful to choose if user wants to use already
 existing public net or create a new one/
- private network:
 new variable: bool tempest_private_net_create (default: false)
 if true: require all tempest_private_* variables to be set
 explanation: by default tempest has use_dynamic_credentials &
 create_isolated_networks enabled, so it spawns an usable network,
 subnet, and router when needed for each project it creates, so in
 most cases it doesn't make sense to create special 'private' net.
- router:
 new variable: bool tempest_router_create (default: false)
 if true: both tempest_public_net_create and
 tempest_private_net_create should be enabled
 explanation: same case as for private network
- image:
 new variable: bool tempest_images_create (default: true)
 if false: require tempest_glance_image_id_1,
 tempest_glance_image_id_2 to be set
 explanation: tempest needs public images, so it'd be useful to just
 use already existing ones
- flavor:
 new variable: bool tempest_flavors_create (default: true)
 if false: require tempest_nova_flavor_id_1, tempest_nova_flavor_id_2
 to be set
 explanation: tempest needs public flavors, so it'd be useful to just
 use already existing ones
- projects:
 new variable: bool tempest_projects_create (default:
 "{{ tempest_public_net_create or tempest_private_net_create or
 tempest_public_router_create }}")
 explanation: by default tempest has use_dynamic_credentials &
 create_isolated_networks enabled, so we don't need to create any
 'static' projects, but network resources spawned by tempest
 playbook belong to this project, so it should exist if any of
 these resources are going to be created
I don't see any reason why users and tempest role should be created
so i dropped this functionality.
Depends-On: https://review.opendev.org/c/openstack/openstack-ansible/+/825166
Change-Id: Icb46f5cb9e2dda6511cc49e30587f541aa7e399f

Use a first_found lookup instead of a with_first_found loop so that
the 'paths' parameter can be used.
This ensures that only vars from the role are included, and not vars
from a parent calling role. This can happen when a parent role has
a higher priority vars file available for inclusion than the role
it calls.
Change-Id: I08dbacef3ef29548cce58cbaee79a3fa79086259

Ansible's combine() filter needs recursive=True parameter in order to recursively merge nested hashes.
https: //docs.ansible.com/ansible/latest/user_guide/playbooks_filters.html#combining-hashes-dictionaries
Change-Id: I804bf03f7e459a5e572fc55d1d40c3ba45d2d736

This add a aditional test in the tempest_test_includelist. This is
required right now as a workaround for tripleo jobs, while the
implementation of the includelist in openstack-tempest-skiplist tool.
Tripleo uses os_tempest to setup and run tempest, but it doesn't run all
tempest tests. Sometimes, the list of excluded tests matches the list of
included tests and so tempest fail with no tests to run. This
tempest_test_extra_test will ensure that it will execute at least one
keystone api test to avoid this.
Right now, the tempest_test_whitelist variable is parsed to ansible via
-e command line, and so it's hard to overwrite it on tripleo side, and
it's spreaded in several different jobs with several different set of
tests. So, this is the best/easy way to fix this problem, while we work
on a long term solution.
Change-Id: Ic5f5b44d81ed39b1fa622b5a1d76e482c2aa23aa

Add a file containing the list of failing tests is important to send
these to logstash and make it easy for people identify how often a
particular test is failing in a particular job.
This should not impact the time of the job or the size of the file, once
it's a small and very quickly command to run
Change-Id: Ib6c2dfc41ac48c3ca6822d6298e1cc55f5c28fbe

When we're in CI and not building wheels, we might appear missing
py3-dev inside utility container, which is required for netifaces. So we
add it as `tempest_devel_distro_packages` which will get package
installed on the required target.
Change-Id: Ie8e04a1635a32e368ec3906082f7773dcefd30d1

'--whitelist-file' option is deprecated, use '--include-list' instead which is functionally equivalent.
Right now Tempest still supports this option for backward compatibility, however, it will be removed soon.
This change addresses the occurences of whitelist and blacklist in the os_tempest OSA role.
Change-Id: Id021d084867ef2891c381c5863b82421e90ee2ca

This will allow us to collect tempestconf.log by CI
Change-Id: Ieb050fe65947e3d2da1beb5062f509036e32a7da

Change-Id: Ide5368b9c26b81a2824fef350ee9d9dd692cdf18

Save the output of tempestconf to a file is good to debug what's
happening with tempestconf and debug the code
Change-Id: I34896c7ab3c0b75f2587a6bef3076366af67c592

All references to Gentoo, SUSE, Debian stretch and Centos-7 are removed.
Conditional tasks, ternary operators and variables are simplified where possible
OS specific variables files are generalised where possible
Change-Id: Ie84846f8590c3ae60afd671d973a4de3c644d526

See https://github.com/ansible/ansible/issues/73654
Change-Id: Ic75eac3e27fd542370a04f9bd946d907f0756061

os_ prefixed module names are deprecated for a while and will be removed
Current usage generates a warning for users.
Change-Id: I85b5666fabfd17d4731bfa085a2e89a17b184b3a

This is necessary to support the new pip resolver.
Depends-On: I9be6bbf4a29a4da2ddf96dc0336bc2a7d8ec9281
Depends-On: I49c75dd11d6c4e8d37fe013b7ffdfd56ff193fcd
Change-Id: If53aca6f16324d305b150466b94d3f18c8a1096a

ansible block w/ error handling was incorrect: the first block
failed when tempest executed, then the second one (which creates
stackviz) never got executed.
It also includes in the block assertion to check if tempest
failed (this is actually redundant check, not really needed).
Change-Id: Ie99060fce29f35ced77efb7b29b8f5f6e37abbcd
Closes-Bug: #1905460 

If a user sets tempest_cleanup variable to true, tempest cleanup will be
executed prior tempest tests (to generate saved_state.json) and after the
tests in order to delete leaked resources if any.
Change-Id: Ief00b43237cb634ee25a18264461bfda8464bbc7

This patch is part of a set that aims to break out tempest
resource creation from the role, and introduces the
tempest_default_role_resources var. When true, default role
resources are created. When false, deployer-specific resources
must be created prior to executing os-tempest-install.yml or
including the os_tempest role in other playbooks.
Change-Id: I42cb8ffbe8aa8cfdcdcfcbbccf28117a27c519b1

OVN requires MTU to be set no more then 1442, so we need ability to
define non-default MTU for networks.
Change-Id: I486981f95cdfa5498f5e31222de4e58d8aa7a516

The tempest init was checking if the workspace/etc directory exists, and
were creating a new workspace without pass a name, the default is
workspace, but passing a --name ensure it won't break if tempest decides
to change it.
Also, the rename command only change the name in the
$HOME/.tempest/workspace.yml, it doesn't change the path for the
particular workspace. The same happens with the move command, it just
update the path on the workspace.yml file.
With this patch, the tempest workspace is being moved properly to a
workspace_$CURDATE, all the files are copied to the new workspace,
delete the old workspace directory, and call tempest init properly.
Also adding tripleo upgrade job as nv for now, since this issue only
happen in upgrade jobs.
Change-Id: I8a3b79352819f5e980eaea7482cd6b1f1bfc47be
Closes-Bug: #1896126 

Do some role cleanup to path for not supported OS,
like gentoo or suse.
This includes bionic distro jobs, since this path is not supported
for bionic in Victoria
Fix ansible linter errors as a result of linter update.
Change-Id: I619dc6da251bd9d6f9d117e7b0cdedc2e59bf581

Depends-On: https://review.opendev.org/718431
Depends-On: https://review.opendev.org/718362
Depends-On: https://review.opendev.org/730083
Change-Id: I8648a3ec8955d74d813a0b5b1cb90ab3a5e6ed34

We need to set mode for copy operation otherwise ansible-2.9.12
can lead to too restrictive default permissions[1].
Setting mode for copy operations handles the warning.
[1.] https://github.com/ansible/ansible/issues/67794
Change-Id: Ieae36a1d85a7da84ee2cd982a20dcabe9e65511f
Signed-off-by: Chandan Kumar (raukadah) <chkumar@redhat.com>

Setting this condition will make possible to only install tempest
without creating weird resources. We still will install them,
when deployer explicitly set to run tempest with the role.
Change-Id: I37bc266eebd3a8455acac179055b87810143dfc4

This patch fixes the warning:
[DEPRECATION WARNING]: The 'os_project_facts' module has been renamed to
'os_project_info', and the renamed one no longer returns ansible_facts.
Change-Id: I7fe4bda991b968530453c620e1ea56a187a26924