Integration tests

Writing e2e tests

We aim to run e2e tests in parallel by default. As such the following rules should be followed:

• Use cirros and alpine VMs for testing wherever possible. If you have to use another OS, discuss on the PR why it is needed.
• Stay within the boundary of the Testnamespaces which we prove. If you have to create resources outside the test namespaces, discuss potential solutions on such a PR.
• If you really have to run tests serial (destructive tests, infra-tests, ...), add the Serial decorator to the test.
• If tests are not using the default cleanup code, additional custom preparations may be necessary.

Additional suggestions:

• The [Disruptive] tag is recognized by the test suite but is not yet mandatory. Feel free to set it on destructive tests.
• Conformance tests need to be marked with a decorators.Conformance label.
• We try to mark tests that require advanced/special storage capabilities with [storage-req],
  So they are easy to skip when lanes with new storage solutions are introduced.
  At the point of writing this we use rook-ceph-block which certainly qualifies for running them.

Test Namespaces

If tests are executed in parallel, every test gets its unique set of namespaces to test in. If you write a test and reference the namespaces test.NamespaceTestDefault, test.NamespaceTestAlternative or tests.NamespaceTestOperator, you get different values, based on the ginkgo execution node.

For as long as test resources are created by referencing these namespaces, there is no test conflict to expect.

Running integration tests

Integration tests require a running Kubevirt cluster. Once you have a running Kubevirt cluster, you can use the -master and the -kubeconfig flags to point the tests to the cluster.

Running networking tests for outside connectivity

When running the tests with no internet connection, some networking tests ,that test outside connectivity, might fail, and you might want to skip them. For that some additional flags are needed to be passed. In addition, if you'd like to test outside connectivity using different addresses than the default (8.8.8.8, 2001:db8:1::1 and google.com), you can achieve that with the designated flags as well.

For each method detailed below, there is note about the needed flags to the outside connectivity tests and how to pass them.

Run them on an arbitrary KubeVirt installation

cd tests # from the git repo root folder
go test -kubeconfig=path/to/my/config -config=default-config.json

outside connectivity tests: The tests will run by default. To skip the outside connectivity tests add -ginkgo.skip='\[outside_connectivity\]' To your go command. To change the IPV4, IPV6 or DNS used for outside connectivity tests, add conn-check-ipv4-address, conn-check-ipv6-address or conn-check-dns to your go command, with the desired value. For example:

go test -kubeconfig=$KUBECONFIG -config=default-config.json \
-conn-check-ipv4-address=8.8.4.4 -conn-check-ipv6-address=2620:119:35::35 \
-conn-check-dns=amazon.com \

Run them on one of the core KubeVirt providers

There is a make target to run this with the config taken from hack/config.sh:

# from the git repo root folder
make functest

outside connectivity tests: The tests will run by default. To skip the tests export KUBEVIRT_E2E_SKIP='\[outside_connectivity\]' environment variable before running the tests. To change the IPV4, IPV6 or DNS used for outside connectivity tests, you can export CONN_CHECK_IPV4_ADDRESS, CONN_CHECK_IPV6_ADDRESS and
CONN_CHECK_DNS with the desired values. For example:

export CONN_CHECK_IPV4_ADDRESS=8.8.4.4
export CONN_CHECK_IPV6_ADDRESS=2620:119:35::35
export CONN_CHECK_DNS=amazon.com

Run them anywhere inside of container

# Create directory for data / results / kubectl binary
mkdir -p /tmp/kubevirt-tests-data
# Make sure that everybody can write there
setfacl -m d:o:rwx /tmp/kubevirt-tests-data
setfacl -m o:rwx /tmp/kubevirt-tests-data

docker run \
    -v /tmp/kubevirt-tests-data:/home/kubevirt-tests/data:rw,z --rm \
    kubevirt/tests:latest \
        --kubeconfig=data/openshift-master.kubeconfig \
        --container-tag=latest \
        --container-prefix=quay.io/kubevirt \
        --test.timeout 180m \
        --junit-output=data/results/junit.xml \
        --deploy-testing-infra \
        --path-to-testing-infra-manifests=data/manifests

Skipping tests when cluster requirements aren't met

KubeVirt provides tests decorators to indicate the cluster requirements for certain tests. If your cluster doesn't meet these requirements, you can choose to skip the relevant tests by explicitly setting the KUBEVIRT_E2E_SKIP environment variable with the appropriate decorators.

For example, to skip tests that require more than one schedulable node, which is useful for single-node clusters, run:

export KUBEVIRT_E2E_SKIP=requires-two-schedulable-nodes

The KUBEVIRT_E2E_SKIP variable supports regular expressions. For example, to skip tests requiring two schedulable nodes and specific storage, run:

export KUBEVIRT_E2E_SKIP=requires-two-schedulable-nodes|storage-req

For more information on filtering tests when using Ginkgo directly, refer to Ginkgo documentation.

Filtering tests by labels

You can filter tests by labels using the KUBEVIRT_LABEL_FILTER environment variable. This variable allows you to specify labels that tests must have to be executed.
For example, to run only tests with the Conformance label, run:

export KUBEVIRT_LABEL_FILTER=Conformance

The KUBEVIRT_LABEL_FILTER variable supports regular expressions.

Focusing on specific tests

To focus on specific tests, you can use the KUBEVIRT_E2E_FOCUS environment variable. This variable supports regular expressions to match the test descriptions.
For example, to run only tests related to networking, run:

export KUBEVIRT_E2E_FOCUS=networking

The KUBEVIRT_LABEL_FILTER variable supports regular expressions.