The plugin pytest-split-tests is used to split the tests in several
groups so that the tests can be executed in parallel is upstream CI.
The issue is thet pytest-split-tests last release was more that a
year ago, upstream developers have not been responsive, and there is
a bug when the number of tests to be executed by a group is zero.
The patch to fix this issue has been open for a year.
This patch deprecates the use of pytest-split-tests, changing the
plugin used to split the tests to pytest-split, which is actively
mainatined.
The idstart needs to be larger than UID_MAX or GID_MAX from /etc/login.defs.
This is "Require idstart to be larger than UID_MAX" for freeipa.
Fixes: #896 (Invalid RID/SID SSSD backtrace after deployment)
Some DNS tests have been disabled for Fedora latest, but not for Fedora
Rawhide. Therefore these tests are filin still in nighty:
- dnsforwardzone
- test_dnsconfig_forwarders_ports
The client part installation is checking for the client packages. These
packages are part of the server packages that have been installed with
the server role and therefore the task is not needed.
This is helping to reduce the deployment time of a server.
Before "short description" was used in most plugins, modules and also
in the new module templates.
ansible-doc was therefore not showing the short description. To fix the
issue the flag was renamed to short_description instead.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2121362
'ansible-doc' -l lists most idm modules as 'UNDOCUMENTED'
When enabling or disabling multiple tests, a comma separated list must
be used, but current code is using a colon to split the list.
This patch fixes this behavior.
Attributes 'first' and 'last' are required if user does not exist, but
current documentation doesn't make it clear. This patch adds a note on
both attributes to make clear the cases where the attribute is required
Since test configuration can vary in different scenarios (test images)
this patch adds a script to list the scenarios configuration, and a
step to the playbook test jobs to display the scenario configuration.
Currently, all tests are scheduled to execution, even those that are
not executed due to being absent from the list of enabled tests
configured in the IPA_ENABLED_* variables. The tests that are not
executed are marked 'skipped'.
This patch change this behavior by not scheduling tests that are not
configured to be executed. It means that tests not the IPA_DISABLED_*
lists are not skipped anymore, but not scheduled to be executed. If
any test is in IPA_ENABLED_* lists, only those tests are marked for
execution. A side effect is that there is no visual feedback on which
tests were not executed, as disabled tests are not evaluated anymore.
Also, when IPA_SERVER_HOST was not set, all tests were skipped, but
an error should raised in this case, as there are no hosts to run the
tests against.
This patch modifies this behavior to fail the test with an exception if
IPA_SERVER_HOST is not set.
Sometimes, mostly due do differences in package versions, there are
some tests that fail on a single distribution which cannot be fixed
by ansible-freeipa, requiring that the offending package is fixed.
To keep tests running succesfully we have options to disable the
failing tests, but this changes are globally applied, meaning that, by
disabling a test, it is disable in all tested distributions.
This patch allows tests to be enabled or disabled for a specific
distribution, by setting the configuration on the 'variable' template
for the specific testing scenario.
The NTP server chrony was always enabled and set up due to overwriting
the parameter ipaclient_no_ntp for the client part deployment.
For IPA deployments up to 4.6 no_ntp was always used for the client
part deployment in ipa-replica-install. But afterwards ntp was
configured in the replica deployment part if no_ntp was not set.
The ipareplica roles always relied on the client for setting up the NTP
server but overwrote the setting for the client deployment part. This
did not result in a failure to enable the chrony server in RHEL and Fedora
based distributions as NTP server was always required by the ipa-server
package.
Fixes: #871 (ipa-replica-install with no-ntp is ignored)
flake8 reports an issue in ipavault:
plugins/modules/ipavault.py:528:20: E275 missing whitespace after keyword
The missing whitespace has been added: "and not(" -> "and not ("
The internal parameter sid_generation_always is generated in
ipareplica_test to enable SID generation if ipareplica_setup_adtrust is
not enabled.
This parameter was not used for ipareplica_prepare though, therefore
adtrust.install_check was not executed and did not set the attribute
adtrust.netbios_name. As a result adtrust.netbios_name was None and the
try to use this as the new NetBIOS domain name failed with an
INVALID_SYNTAX error in adtrustinstance while executing
ipareplica_setup_adtrust.
This issue only occurs if SIDs are not enabled in the domain yet for
example with an old deployment.
The SID is always generated in the command line installers in newer IPA
versions. This also needs to be done in the ipaserver and ipareplica roles.
For the IPA versions that are supporting this, the adtrust setup is always
executed to generated the SIDs, but only configures AD trust if
ipaserver_setup_adtrust or ipareplica_setup_adtrust is also enabled. A
check has been added to ipaserver_test and ipareplica_test to only enable
the SID generation for the IPA versions supporting this.
This is related to https://pagure.io/freeipa/8995
Fixes:
- https://bugzilla.redhat.com/show_bug.cgi?id=2110478
- https://bugzilla.redhat.com/show_bug.cgi?id=2110491
Due to an issue with python-dns, FreeIPA is raising an expection when
setting a DNS forwarder with a custom port. Separating the test for
ipadnsconfig that use forwarders with custom allows the other tests
to be correctly executed.
When running tests/sanity/sanity.sh locally, podman might be available
instead of Docker. Due to current configuration, only Docker is used by
sanity.sh.
This patch searches for the availability of docker, which is kept as
the default container engine to use, and use podman only if docker is
not found.
This change also allows the execution of the script from a directory
other than the repository root.
api_check_ipa_version was using packaging.version. IPA is using
pkg_resources.parse_version in ipaplatform.tasks.parse_ipa_version.
With this change tasks.parse_ipa_version from ipaplatform is used to
have exactly the same version comparison that also IPA has.
Additionally tasks is added to __all__.
Checking if some output is present or absent from standard streams was
done by simple string searching. Due to recent changes in FreeIPA, this
search is not effective due to capitalization differences in boolean
values output. Changing the string searching to regular expression
searches fixes this behavior for current and previous versions of
FreeIPA.
This patch also adds more information on the assert tests in case of an
error, so that it is easier to understand why the test failed.
The fallback function used to compare IPA versions was spliting the
version string into a tuple of strings, and the comparison of the tuple
would fail if comparing a field with one digit aginst a two-digit one,
for example, '8' with '10', as the string comparison would put '10'
before the '8'.
This patch forces the version fields to be converted to integers, so
a numerical comparison will be performed. If a version string field
cannot be converted to a number, than the string comparison will still
be used.
