Integration Testing Harness for Openstack Snaps
Python Shell
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
example
snapstack
tests
.gitignore
.travis.yml
LICENSE
README.md
admin-openrc
requirements.txt
setup.py
snap-deploy
snap-destroy
test-requirements.txt
tox.ini

README.md

Snapstack

Snapstack is an integration testing harness for Openstack Snaps.

Snapstack also includes scripts for deploying a minimal snap-based OpenStack cloud.

Using Snapstack

Deploying a Cloud

If you just want to deploy and configure a minimal snap-based OpenStack cloud (keystone, neutron, glance, nova, nova-hypervisor) you can run:

snap-deploy
source admin-openrc

To destroy the cloud you can run:

snap-destroy

Testing a Cloud

The main purpose of snapstack, however, is that it provides a harness for doing basic CI on a snap. To use it, you'll want to pass one or more Steps into a Plan object, and then run the Plan. Steps typically specify a snap to install, config files to write, and executable Python or bash scripts that will configure and perform a basic smoke test on the snap.

The Setup class in snapstack/base.py defines the Steps for the default base OpenStack cloud. While the steps for the base OpenStack cloud exist in the snapstack code base, steps for individual snaps will usually live in the corresponding snap's git repo.

The following example is a basic test that installs and tests an example snap, after letting the Plan object install and configure the default base OpenStack (keystone, neutron, glance, nova, and nova-hypervisor) from the snap store.

example = Step(
    snap='example',
    script_loc='./tests/',
    scripts=['example.sh'],
    files=['example.json'],
    snap_store=False
)
plan = Plan(tests=[example])
plan.run()

Take a look in the example directory to see the source snippet above in context.

The following example is a bit different than the first. While it also installs and configures the default base OpenStack, it does so while overriding the default base keystone with a locally built keystone snap. It then runs the specified test scripts. This is taken from the keystone snap's git repo.

setup = Setup()
setup.add_steps(('keystone', Step(
    snap='keystone',
    script_loc='./tests/',
    scripts=['keystone.sh'],
    files=['etc/snap-keystone/keystone/keystone.conf.d/database.conf'],
    snap_store=False)))
plan = Plan(base_setup=setup.steps())
plan.run()

In the previous examples, the script location (script_loc) is a local relative path, however this could also be a remote URL location. If you take a closer look at the Setup class in snapstack/base.py, you'll see that the script_loc for each of the base snaps corresponds to each individual snap on github. For example, the glance script is located at: https://raw.githubusercontent.com/openstack/snap-glance/master/tests/glance.sh

This codebase is meant to be a fairly lightweight Python wrapper around your shell scripts. The purpose is to test the snap, rather than to extensively test the underlying source, so basic tests will usually suffice.

Flags and Other Notes

If you are in a network restricted environment, and need to go through a proxy to make HTTP requests, set SNAPSTACK_HTTPS_PROXY and SNAPSTACK_HTTP_PROXY in your terminal environment. This will set HTTP/S_PROXY for the snap build steps, and give you an environment variable that you can reference in your scripts.