To start read the OpenStack Common Review Checklist
For any new or change to a test it has to be verified in the gate. This means that the first thing to check with any change is that a gate job actually runs it. Tests which aren't executed either because of configuration or skips should not be accepted.
If a new test is added that depends on a new config option (like a feature flag), the commit message must reference a change in DevStack or DevStack-Gate that enables the execution of this newly introduced test. This reference could either be a Cross-Repository Dependency or a simple link to a Gerrit review.
While checking in the job logs that a new test is actually executed, also pay attention to the execution time of that test. Keep in mind that each test is going to be executed hundreds of time each day, because Tempest tests run in many OpenStack projects. It's worth considering how important/critical the feature under test is with how costly the new test is.
For any change that adds new functionality to either common functionality or an
out-of-band tool unit tests are required. This is to ensure we don't introduce
future regressions and to test conditions which we may not hit in the gate runs.
API and scenario tests aren't required to have unit tests since they should
be self-verifying by running them in the gate. All service clients, on the
other hand, must have unit tests, as they belong to tempest/lib.
Tests should only be added for published stable APIs. If a patch contains tests for an API which hasn't been marked as stable or for an API which doesn't conform to the API stability guidelines then it should not be approved.
When creating new tests that are similar to existing tests it is tempting to simply copy the code and make a few modifications. This increases code size and the maintenance burden. Such changes should not be approved if it is easy to abstract the duplicated code into a function or method.
When a new test is being proposed, question whether this feature is not already tested with Tempest. Tempest has more than 1200 tests, spread amongst many directories, so it's easy to introduce test duplication. For example, testing volume attachment to a server could be a compute test or a volume test, depending on how you see it. So one must look carefully in the entire code base for possible overlap. As a rule of thumb, the older a feature is, the more likely it's already tested.
When tests are being added that depend on a configurable feature or extension, polling the API to discover that it is enabled should not be done. This will just result in bugs being masked because the test can be skipped automatically. Instead the config file should be used to determine whether a test should be skipped or not. Do not approve changes that depend on an API call to determine whether to skip or not.
With the introduction of the Tempest external test plugin interface we needed to provide a stable contract for Tempest's configuration options. This means we can no longer simply remove a configuration option when it's no longer used. Patches proposed that remove options without a deprecation cycle should not be approved. Similarly when changing default values with configuration we need to similarly be careful that we don't break existing functionality. Also, when adding options, just as before, we need to weigh the benefit of adding an additional option against the complexity and maintenance overhead having it costs.
When a new test is being added refer to the :ref:`TestDocumentation` section in hacking to see if the requirements are being met. With the exception of a class level docstring linking to the API ref doc in the API tests and a docstring for scenario tests this is up to the reviewers discretion whether a docstring is required or not.
Make sure that any test that is renamed, relocated (e.g. moved to another class), or removed does not belong to the interop testing suite -- which includes a select suite of Tempest tests for the purposes of validating that OpenStack vendor clouds are interoperable -- or a project's whitelist or blacklist files.
It is of critical importance that no interop, whitelist or blacklist test reference be broken by a patch set introduced to Tempest that renames, relocates or removes a referenced test.
Please check the existence of code which references Tempest tests with: http://codesearch.openstack.org/
Make sure that modifications to an interop test are backwards-compatible. This means that code modifications to tests should not undermine the quality of the validation currently performed by the test or significantly alter the behavior of the test.
Reference the :ref:`test-removal` guidelines for understanding best practices associated with test removal.
Release notes are how we indicate to users and other consumers of Tempest what has changed in a given release. Since Tempest 10.0.0 we've been using reno to manage and build the release notes. There are certain types of changes that require release notes and we should not approve them without including a release note. These include but aren't limited to, any addition, deprecation or removal from the lib interface, any change to configuration options (including deprecation), CLI additions or deprecations, major feature additions, and anything backwards incompatible or would require a user to take note or do something extra.
Sometimes we have some bugs in deprecated code. Basically, we leave it. Because we don't need to maintain it. However, if the bug is critical, we might need to fix it. When it will happen, we will deal with it on a case-by-case basis.
It's OK to hold off on an approval until a subject matter expert reviews it.
Every patch needs at least single +2's before being approved. A single Tempest core reviewer can approve patches but can always wait for another +2 in any case. Following cases where single +2 can be used without any issue:
If any trivial patch set fixes one of the items below:
- Documentation or code comment typo
- Documentation ref link
- Example: example
Note
Any other small documentation, CI job, or code change does not fall under this category.
If the patch unblocks a failing project gate, provided that:
- the project's PTL +1's the change
- the patch does not affect any other project's testing gates
- the patch does not cause any negative side effects
If fixing and removing the faulty plugin (which leads to fail voting
tempest-tox-plugin-sanity-checkjob) and unblock the tempest gate