oVirt System Tests
Welcome to the oVirt System Tests source repository.
This repository is hosted on GitHub:ovirt-system-tests
System requirements
Operating System
Currently OST can run on RHEL or CentOS Stream at least roughly matching the version used by oVirt or RHV. Latest CentOS Stream or RHEL 8 should work.
Disk Space
ost-images packages that contain the preinstalled images are fairly large, make sure you have enough disk space, 15 GB per each distro variant in /usr/share. Additional space is needed for actual suite run, another 5GB at least in the directory where you run tests from.
Memory
The memory requirements are derived from the VM specs you'll have in the test suite, it is recommended that the host you're using will have at least 16GB of RAM, though basic suite still runs with 8GB.
How to contribute
Submitting patches
Patches are welcome!
Please submit patches to GitHub:ovirt-system-tests. If you are not familiar with the review process for GitHub patches you can read about Working with oVirt GitHub on the oVirt website.
Found a bug or documentation issue?
To submit a bug, suggest an enhancement or report a documentation issue for oVirt System Tests please join oVirt Development forum / mailing list and discuss there.
Still need help?
If you have any other questions, please join oVirt Development forum / mailing list and ask there.
Running the tests
Make sure your machine is set up with setup_for_ost.sh
You can use ost.sh
for running the complete suite on a concrete ost-images distro
E.g. ./ost.sh run basic-suite-master el8stream
You can use --custom-repo
for verifying an oVirt patch:
- On your GitHub account go to Settings/Developer settings/Personal access tokens/Generate new token, and give it the repo scopes.
- On your ost machine run
export GITHUB_TOKEN=personal_access_token
- Run the tests, e.g
./ost.sh run basic-suite-master el8stream --custom-repo=https://github.com/oVirt/ovirt-engine/pull/pr_number
You can pass this option multiple times to test more than 1 build at once.
The environment is left running after it finishes so that you can examine or further use the created environment.
It is necessary to clean the environment up aftery every run with ./ost.sh destroy
.
The environment contains inventory, VMs, live VM logs and is normally stored
in deployment
subdirectory of the repository. The location can be changed
through OST_DEPLOYMENT
environment variable.
When the environment is up you can use ./ost.sh shell
to connect to the VMs.
./ost.sh status
shows a little overview of how the OST environment is laid out on the host system.
Advanced use
ost.sh internally uses bash functions from lagofy.sh that you can source into your (bash only, other shells are not supported) environment and use for more granular control. Functions starting with ost_ and OST_ variables can be used to bring up the VMs with custom images, boot OST environment without running tests, or to run individual test cases.
You can also access the web ui easily, locate the engine's IP address in ./ost.sh status
. In case of Hosted Engine-based suites where
the engine VM is not declared in OST you can inspect the DNS entries of management network by virsh net-dumpxml
.
Once you've located the engine VM's IP, add it to /etc/hosts
. You have to use the exact FQDN (for basic suite you can also use just engine
)
Now, log in to the web-UI at:
- URL:
https://engine/ovirt-engine/webadmin/
- Username:
admin
- Password:
123456
- Profile:
internal
If you're running OST on a remote machine, you can tunnel a local port directly to the destination machine - from your local machine:
$ sudo ssh -L 443:192.168.200.2:443 USER@HOST_RUNNING_OST
---- ================= ~~~~~~~~~~~~~~~~
(*) (**) (***)
(*) - The port on the local machine that the tunnel will be available at.
(**) - The machine IP, visible from the *host running OST* . This is were the traffic is tunneled to.
Usually 192.168.200.2 is the address of ovirt-engine, but you can check that with ./ost.sh status
(***) - The host running OST which can reach the VMs network and will tunnel the connection.
Running specific tests
It's often useful to run just a single test, usually after running a suite, to test something using the existing machines. Following is an example showing how to run the he-basic-suite, and after it's finished, run again one of the tests.
First, run the suite:
./ost.sh run he-basic-suite-master el8stream
Now, you might change the code, e.g. to run a test with some change in it, or a new test you are developing. Then, run it:
. lagofy.sh
ost_init he-basic-suite-master
ost_run_tc he-basic-suite-master/test-scenarios/test_030_ovirt_auth.py
The first argument to ost_run_tc
is the test to run - a module, or a method/class inside one, e.g.:
ost_run_tc basic-suite-master/test-scenarios/test_004_basic_sanity.py::test_vdsm_recovery
An optional second argument is passed as -k
to pytest, for selecting tests based on pattern, e.g. this will run test_010_local_maintenance_cli.py
and test_012_local_maintenance_sdk.py
:
ost_run_tc he-basic-suite-master/test-scenarios local_maintenance
and this will run test_add_hosts
and test_verify_add_hosts
:
ost_run_tc basic-suite-master/test-scenarios/test_002_bootstrap.py add_hosts
See the pytest documentation for details.
Please note, that these are ran in a separate pytest session, and that several of our tests depend on previous ones to pass, and/or on a specific state of the test VMs.
Development
Make sure your machine is set up with setup_for_ost.sh
OST is designed to run concurrently under single non-root user, but in separate checkouts.
At runtime it creates "deployment" and "exported-artifacts" directories in the workspace.
Make sure that your new code doesn't depend on anything on the OST host, whenever you need something it's usually a better idea to run that inside one of the VMs.
Running tox tests
For developers of OST, a quick flake8 and pylint runs are available to identify any syntax errors, typos, unused import etc.
To run the flake8/pylint, you'll need to have tox installed. To run pylint without import errors you should also have ovirtdsk4 and ansible-runner installed.
Running linting tests:
tox -e flake8,pylint,black,broken-symlinks
or ost_linters
lagofy function.