While releasing new version in pypi, it perform the validation on the package and give error if there is any formatting issue. warning: check: Duplicate implicit target name: "storyboard". error: Please correct your package. - http://logs.openstack.org/68/645868/1/check/openstack-tox-validate/9d9219e/job-output.txt.gz#_2019-03-22_20_15_40_332673 This commit fixes the format error to release the patrole. Change-Id: Ie840743b61bab5c833f554bb40b6a162e790a29f
11 KiB
Team and repository tags
Patrole - RBAC Integration Tempest Plugin
Patrole is a set of integration tests to be run against a live OpenStack cluster. It has a battery of tests dedicated to validating the correctness and integrity of the cloud's RBAC implementation.
More importantly, Patrole is a security validation tool for verifying that Role-Based Access Control is correctly configured and enforced in an OpenStack cloud. It runs Tempest-based API tests using specified RBAC roles, thus allowing deployments to verify that only intended roles have access to those APIs.
Patrole is currently undergoing heavy development. As more projects move toward policy in code, Patrole will align its testing with the appropriate documentation.
- Free software: Apache license
- Documentation: https://docs.openstack.org/patrole/latest
- Source: https://git.openstack.org/cgit/openstack/patrole
- Bugs: https://storyboard.openstack.org/#!/project/openstack/patrole
- Release notes: https://docs.openstack.org/releasenotes/patrole/
Design Principles
As a Tempest plugin, Patrole borrows some design principles from Tempest design principles, but not all, as its testing scope is confined to policies.
Stability. Patrole uses OpenStack public interfaces. Tests in Patrole should only touch public OpenStack APIs.
Atomicity. Patrole tests should be atomic: they should test policies in isolation. Unlike Tempest, a Patrole test strives to only call a single endpoint at a time. This is because it is important to validate each policy is authorized correctly and the best way to do that is to validate each policy alone, to avoid test contamination.
Complete coverage. Patrole should validate all policy in code defaults. For testing, Patrole uses the API-to-policy mapping contained in each project's policy in code documentation where applicable.
For example, Nova's policy in code documentation is located in the Nova repository under
nova/policies
. Likewise, Keystone's policy in code documentation is located in the Keystone repository underkeystone/common/policies
. The other OpenStack services follow the same directory layout pattern with respect to policy in code.Note
Realistically this is not always possible because some services have not yet moved to policy in code.
Customizable. Patrole should be able to validate custom policy overrides to ensure that those overrides enhance rather than undermine the cloud's RBAC configuration. In addition, Patrole should be able to validate any role.
Self-cleaning. Patrole should attempt to clean up after itself; whenever possible we should tear down resources when done.
Note
Patrole modifies roles dynamically in the background, which affects pre-provisioned credentials. Work is currently underway to clean up modifications made to pre-provisioned credentials.
Self-testing. Patrole should be self-testing.
Features
- Validation of default policy definitions located in policy.json files.
- Validation of in-code policy definitions.
- Validation of custom policy file definitions that override default policy definitions.
- Built-in positive and negative testing. Positive and negative testing are performed using the same tests and role-switching.
- Valdation of custom roles as well as default OpenStack roles.
Note
Patrole does not yet support policy.yaml files, the new file format for policy files in OpenStack.
How It Works
Patrole leverages oslo.policy
(OpenStack's policy
enforcement engine) to determine whether a given role is allowed to
perform a policy action, given a specific role and OpenStack service.
The output from oslo.policy
(the expected result) and the
actual result from test execution are compared to each other: if both
results match, then the test passes; else it fails.
Terminology
- Expected Result - The expected result of a given test.
- Actual Result - The actual result of a given test.
- Final Result - A match between both expected and actual results. A
mismatch in the expected result and the actual result will result in a
test failure.
- Expected: Pass | Actual: Pass - Test Case Success
- Expected: Pass | Actual: Fail - Test Case Under-Permission Failure
- Expected: Fail | Actual: Pass - Test Case Over-Permission Failure
- Expected: Fail | Actual: Fail (Expected exception) - Test Case Success
- Expected: Fail | Actual: Fail (Unexpected exception) - Test Case Failure
Quickstart
To run Patrole, you must first have Tempest installed and configured properly. Please reference Tempest_quickstart guide to do so. Follow all the steps outlined therein. Afterward, proceed with the steps below.
You first need to install Patrole. This is done with pip after you check out the Patrole repo:
$ git clone https://git.openstack.org/openstack/patrole $ pip install patrole/
This can be done within a venv.
Note
You may also install Patrole from source code by running:
pip install -e patrole/
Next you must properly configure Patrole, which is relatively straightforward. For details on configuring Patrole refer to the Patrole Configuration.
Once the configuration is done you're now ready to run Patrole. This can be done using the tempest_run command. This can be done by running:
$ tempest run --regex '^patrole_tempest_plugin\.tests\.api'
There is also the option to use testr directly, or any testr based test runner, like ostestr. For example, from the workspace dir run:
$ stestr --regex '(?!.*\[.*\bslow\b.*\])(^patrole_tempest_plugin\.tests\.api))'
will run the same set of tests as the default gate jobs.
You can also run Patrole tests using tox, but as Patrole needs access to global packages use
--sitepackages
argument. To do so,cd
into the Tempest directory and run:$ tox -eall --sitepackages -- patrole_tempest_plugin.tests.api
Note
It is possible to run Patrole via
tox -eall
in order to run Patrole isolated from other plugins. This can be accomplished by including the installation of services that currently use policy in code -- for example, Nova and Keystone. For example:$ tox -evenv-tempest -- pip install /opt/stack/patrole /opt/stack/keystone /opt/stack/nova $ tox -eall -- patrole_tempest_plugin.tests.api
Log information from tests is captured in
tempest.log
under the Tempest repository. Some Patrole debugging information is captured in that log related to expected test results and Role Overriding.More detailed RBAC testing log output is emitted to
patrole.log
under the Patrole repository. To configure Patrole's logging, see the Patrole Configuration Guide.
RBAC Tests
To change the roles that the patrole tests are being run as, edit
rbac_test_roles
in the patrole
section of
tempest.conf: :
[patrole]
rbac_test_roles = member,reader
...
Note
The rbac_test_roles
is service-specific. member, for
example, is an arbitrary role, but by convention is used to designate
the default non-admin role in the system. Most Patrole tests should be
run with admin and member roles.
However, other services may use entirely different roles or role
combinations.
For more information about RBAC, reference the rbac-overview documentation page.
For information regarding which projects Patrole offers RBAC testing for, reference the HACKING documentation page.
Unit Tests
Patrole also has a set of unit tests which test the Patrole code itself. These tests can be run by specifying the test discovery path:
$ stestr --test-path ./patrole_tempest_plugin/tests/unit run
By setting --test-path
option to
./patrole_tempest_plugin/tests/unit
it specifies that test
discovery should only be run on the unit test directory.
Alternatively, there are the py27 and py35 tox jobs which will run the unit tests with the corresponding version of Python.
One common activity is to just run a single test; you can do this with tox simply by specifying to just run py27 or py35 tests against a single test:
$ tox -e py27 -- -n patrole_tempest_plugin.tests.unit.test_rbac_utils.RBACUtilsTest.test_override_role_with_missing_admin_role
Or all tests in the test_rbac_utils.py file:
$ tox -e py27 -- -n patrole_tempest_plugin.tests.unit.test_rbac_utils
You may also use regular expressions to run any matching tests:
$ tox -e py27 -- test_rbac_utils
For more information on these options and details about stestr, please see the stestr documentation.
Release Versioning
Patrole Release Notes shows which changes have been released for each version.
Patrole's release versioning follows Tempest's conventions. Like Tempest, Patrole is branchless and uses versioning instead.
Storyboard
Bugs and enhancements are tracked via Patrole's Storyboard Page.