bee34dbb86
Previousy, when shutting down a libvirt domain before evacuation, we would wait until it disappeared from the list of libvirt domains, but we just merrily continued on even if it was still present after 60 seconds. I suspect the domain remained present, and our next steps were to evacuate the instance and start nova-compute back up again, at which point it tried to delete evacuated instances, and fail with the following: File "/opt/stack/nova/nova/compute/manager.py", line 839, in _destroy_evacuated_instances <snip> libvirt.libvirtError: Failed to terminate process 55183 with SIGKILL: Device or resource busy This patch makes us wait for the domain to disappear, and raise an error if it doesn't disappear in time. It also refactors the shutdown code slightly to avoid an extra API quey to get the server. Change-Id: I76885987842440f9b690c198e6b8149bdd85f899 |
||
---|---|---|
devstack | ||
playbooks | ||
tools | ||
whitebox_tempest_plugin | ||
.gitignore | ||
.gitreview | ||
.pre-commit-config.yaml | ||
.stestr.conf | ||
.zuul.yaml | ||
dictionary.txt | ||
LICENSE | ||
README.rst | ||
requirements.txt | ||
setup.cfg | ||
setup.py | ||
test-requirements.txt | ||
tox.ini |
Whitebox Tempest plugin
This is a Tempest plugin for whitebox testing. While Tempest's scope is limited to only the REST APIs, whitebox allows tests to peak behind the curtain, similar to how a cloud admin might. Examining things on the compute host(s) and/or the controller(s) is not only allowed, it's required for a test to be in whitebox's scope. Whitebox tests must still be REST API-driven, however their assertions can involve things like the instance XML (if the Nova libvirt driver is in use) or the database.
- Bugs: https://storyboard.openstack.org/#!/project/1162
- IRC: #openstack-qa on OFTC
Requirements
While Tempest is cloud-agnostic because all clouds expose the same OpenStack APIs (with some caveats around extensions), whitebox peaks behind the curtain, and thus is coupled to the way the cloud was deployed. Currently, devstack and TripleO (with undercloud and overcloud) are supported, with only devstack being tested in CI.
Some tests have specific hardware requirements. These should be documented as config options, and tests are expected to skip if their hardware requirements are not declared in the configuration.
Install, configure and run
Tempest needs to be installed and configured.
Install the plugin.
This should be done from source. :
WORKSPACE=/some/directory cd $WORKSPACE git clone https://opendev.org/openstack/whitebox-tempest-plugin sudo pip install whitebox-tempest-plugin
Configure Tempest.
The exact configuration will depend on the deployment. There is no configuration reference yet, have a look at
whitebox_tempest_plugin/config.py
instead. As an example, here is a configuration for a multinode TripleO deployment. :[whitebox] ctlplane_addresses = compute-0.localdomain:192.168.24.6,compute-1.localdomain:192.168.24.12 ctlplane_ssh_username = heat-admin ctlplane_ssh_private_key_path = /home/stack/.ssh/id_rsa containers = true max_compute_nodes = 2 # Some tests depend on there being a single (available) compute node
Here is an example for a two-node DevStack deployment:
[whitebox] nodes_yaml = /opt/stack/whitebox-tempest-plugin/nodes.yaml ctlplane_ssh_username = vagrant ctlplane_ssh_private_key_path = /home/vagrant/.ssh/id_rsa
with a
nodes.yaml
file that looks something like:controller: services: libvirt: start-command: 'systemctl start libvirtd' stop_command: 'systemctl stop libvirtd' nova-compute: config_path: '/etc/nova/nova-cpu.conf' start_command: 'systemctl start devstack@n-cpu' stop_command: 'systemctl stop devstack@n-cpu' compute1: services: libvirt: start-command: 'systemctl start libvirtd' stop_command: 'systemctl stop libvirtd' nova-compute: config_path: '/etc/nova/nova-cpu.conf' start_command: 'systemctl start devstack@n-cpu' stop_command: 'systemctl stop devstack@n-cpu'
where
controller
is the hostname of the controller node andcompute1
is the hostname of the second node running nova-compute.Execute the tests. :
tempest run --serial --regex whitebox_tempest_plugin.
Important
Whitebox expects its tests to run one at a time. Make sure to pass
--serial
or--concurrency 1
totempest run
.
How to add a new test
Tests should fit whitebox's scope. If a test intereacts with REST
APIs and nothing else, it is better suited for Tempest itself. New tests
should be added in their respective subdirectories. For example, tests
that use the compute API live in
whitebox_tempest_plugin/api/compute
. Test code does not
need unit tests, but helpers or utilities do. Unit tests live in
whitebox_tempest_plugin/tests
. Whitebox does not adhere to
the Tempest plugin interface
<https://docs.openstack.org/tempest/latest/plugin.html>. As
mentioned, whitebox tests run one at a time, so it's safe for a test to
modify the environment and/or be destructive, as long as it cleans up
after itself. For example, changing Nova configuration values and/or
restarting services is acceptable, as long as the original values and
service state are restored.