Create foundation for Labgrid-based OS tests (#2812)

* Create foundation for Labgrid-based OS tests

Add foundation for Labgrid-based tests of OS builds. Currently uses just
the QEMU driver, which starts a virtual machine with pristine OS, and
generates few log reports which are saved as build artifacts.

Workflow is currently triggered either manually by specifying an OS
version, or by OS build job, which now saves an artifact of the OVA
image. This allows for some modularity. If we eventually add the
possibility to run builds on PRs, we could also add the workflow_call
trigger and turn the workflow into a reusable one.

TBD (in future PRs): some meaningful tests and possibility to test on
real hardware (either local or distributed).

* Apply suggestions from @agners

Co-authored-by: Stefan Agner <stefan@agner.ch>

* Wrap test command in a script, create venv for local tests

* Make shellcheck happy

---------

Co-authored-by: Stefan Agner <stefan@agner.ch>

.github/workflows/build.yaml

@@ -232,6 +232,14 @@ jobs:
           path: /mnt/cache/cc
           key: haos-cc-${{ matrix.board.id }}-${{ github.run_id }}
 
+      - name: Upload ova image to artifacts for test job
+        uses: actions/upload-artifact@v3
+        if: ${{ matrix.board.id == 'ova' }}
+        with:
+          name: ova-image
+          path: |
+            output/images/haos_ova*.qcow2.xz
+
   bump_version:
     name: Bump dev channel version
     if: ${{ github.repository == 'home-assistant/operating-system' }}

.github/workflows/test.yaml

@@ -0,0 +1,92 @@
+name: Test HAOS image
+run-name: "Test HAOS ${{ inputs.version || format('(OS build #{0})', github.event.workflow_run.run_number) }}"
+
+on:
+  workflow_dispatch:
+    inputs:
+      version:
+        description: Version of HAOS to test
+        required: true
+        type: string
+
+  workflow_run:
+    workflows: ["OS build"]  # must be in sync with build workflow `name`
+    types:
+      - completed
+
+jobs:
+  test:
+    if: ${{ github.event_name != 'workflow_run' || github.event.workflow_run.conclusion == 'success' }}
+
+    env:
+      NO_KVM: 1
+
+    name: Test in QEMU
+    runs-on: ubuntu-22.04
+    defaults:
+      run:
+        working-directory: ./tests
+    steps:
+      - name: Install system dependencies
+        run: |
+          sudo apt update
+          sudo apt install -y qemu-system-x86 ovmf
+
+      - name: Checkout source
+        uses: actions/checkout@v4
+        with:
+          persist-credentials: false
+
+      - name: Setup Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: 3.12
+
+      - name: Install Python requirements
+        run:
+          pip install -r requirements.txt
+
+      - name: Download HAOS image
+        if: ${{ github.event_name == 'workflow_dispatch' }}
+        run: |
+          curl -sfL -o haos.qcow2.xz  https://os-artifacts.home-assistant.io/${{github.event.inputs.version}}/haos_ova-${{github.event.inputs.version}}.qcow2.xz
+
+      - name: Get OS image artifact
+        if: ${{ github.event_name == 'workflow_run' }}
+        uses: dawidd6/action-download-artifact@v2
+        with:
+          workflow: build.yaml
+          workflow_conclusion: success
+          name: ova-image
+
+      - name: Extract OS image
+        run: |
+          unxz haos.qcow2.xz
+
+      - name: Run tests
+        run: |
+          ./run_tests.sh
+
+      - name: Archive logs
+        uses: actions/upload-artifact@v3
+        if: always()
+        with:
+          name: logs
+          path: |
+            lg_logs/*
+
+      - name: Archive JUnit reports
+        uses: actions/upload-artifact@v3
+        if: always()
+        with:
+          name: junit_reports
+          path: |
+            junit_reports/*.xml
+
+      - name: Publish test report
+        uses: mikepenz/action-junit-report@v4
+        if: always()
+        with:
+          report_paths: 'junit_reports/*.xml'
+          annotate_only: true
+          detailed_summary: true

.gitignore

@@ -9,4 +9,4 @@ output*/
 *.pem
 
 # vscode generated files
-.vscode*
+.vscode*

tests/.gitignore

@@ -0,0 +1,14 @@
+# QEMU images
+*.qcow2
+
+# Generated logs
+/junit_reports
+/lg_logs
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# Virtualenv
+venv

tests/conftest.py

@@ -0,0 +1,21 @@
+import os
+
+import pytest
+
+
+@pytest.hookimpl
+def pytest_runtest_setup(item):
+    log_dir = item.config.option.lg_log
+
+    if not log_dir:
+        return
+
+    logging_plugin = item.config.pluginmanager.get_plugin("logging-plugin")
+    logging_plugin.set_log_path(os.path.join(log_dir, f"{item.name}.log"))
+
+
+@pytest.fixture
+def shell_command(target, strategy):
+    strategy.transition("shell")
+    shell = target.get_driver("ShellDriver")
+    return shell

tests/qemu-strategy.yaml

@@ -0,0 +1,31 @@
+targets:
+  main:
+    resources: []
+
+    drivers:
+      - QEMUDriver:
+          qemu_bin: qemu-x86_64
+          machine: pc
+          cpu: qemu64
+          memory: 1G
+          extra_args: "-snapshot -accel kvm"
+          nic: user,model=virtio-net-pci
+          disk: disk-image
+          bios: bios
+      - ShellDriver:
+          login_prompt: 'homeassistant login: '
+          username: 'root'
+          prompt: '# '
+          login_timeout: 300
+      - QEMUShellStrategy: {}
+
+tools:
+  qemu-x86_64: /usr/bin/qemu-system-x86_64
+
+images:
+  disk-image: ./haos.qcow2
+  bios: /usr/share/ovmf/OVMF.fd
+
+
+imports:
+  - qemu_shell_strategy.py

tests/qemu_shell_strategy.py

@@ -0,0 +1,53 @@
+import enum
+import os
+
+import attr
+
+from labgrid import target_factory, step
+from labgrid.strategy import Strategy, StrategyError
+
+
+class Status(enum.Enum):
+    unknown = 0
+    off = 1
+    shell = 2
+
+
+@target_factory.reg_driver
+@attr.s(eq=False)
+class QEMUShellStrategy(Strategy):
+    """Strategy for starting a QEMU VM and running shell commands within it."""
+
+    bindings = {
+        "qemu": "QEMUDriver",
+        "shell": "ShellDriver",
+    }
+
+    status = attr.ib(default=Status.unknown)
+
+    def __attrs_post_init__(self):
+        super().__attrs_post_init__()
+        if "-accel kvm" in self.qemu.extra_args and os.environ.get("NO_KVM"):
+            self.qemu.extra_args = self.qemu.extra_args.replace(
+                "-accel kvm", ""
+            ).strip()
+
+    @step(args=["status"])
+    def transition(self, status, *, step):  # pylint: disable=redefined-outer-name
+        if not isinstance(status, Status):
+            status = Status[status]
+        if status == Status.unknown:
+            raise StrategyError(f"can not transition to {status}")
+        elif status == self.status:
+            step.skip("nothing to do")
+            return  # nothing to do
+        elif status == Status.off:
+            self.target.activate(self.qemu)
+            self.qemu.off()
+        elif status == Status.shell:
+            self.target.activate(self.qemu)
+            self.qemu.on()
+            self.target.activate(self.shell)
+        else:
+            raise StrategyError(f"no transition found from {self.status} to {status}")
+        self.status = status

tests/requirements.txt

@@ -0,0 +1 @@
+labgrid==23.0.3

tests/run_tests.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+set -e
+
+if [ -z "$GITHUB_ACTIONS" ] && [ -z "$VIRTUAL_ENV" ]; then
+  # Environment should be set up in separate GHA steps - which can also
+  # handle caching of the dependecies, etc.
+  python3 -m venv venv
+  # shellcheck disable=SC1091
+  source venv/bin/activate
+  pip3 install -r requirements.txt
+fi
+
+pytest --lg-env qemu-strategy.yaml --lg-log=lg_logs --junitxml=junit_reports/smoke_test.xml smoke_test

tests/smoke_test/test_basic.py

@@ -0,0 +1,48 @@
+import logging
+from time import sleep
+
+
+_LOGGER = logging.getLogger(__name__)
+
+
+def test_init(shell_command):
+    def check_container_running(container_name):
+        out = shell_command.run_check(
+            f"docker container inspect -f '{{{{.State.Status}}}}' {container_name} || true"
+        )
+        return "running" in out
+
+    # wait for important containers first
+    for _ in range(20):
+        if check_container_running("homeassistant") and check_container_running("hassio_supervisor"):
+            break
+
+        sleep(5)
+
+    # wait for system ready
+    for _ in range(20):
+        output = "\n".join(shell_command.run_check("ha os info || true"))
+        if "System is not ready" not in output:
+            break
+
+        sleep(5)
+
+    output = shell_command.run_check("ha os info")
+    _LOGGER.info("%s", "\n".join(output))
+
+
+def test_dmesg(shell_command):
+    output = shell_command.run_check("dmesg")
+    _LOGGER.info("%s", "\n".join(output))
+
+
+def test_supervisor_logs(shell_command):
+    output = shell_command.run_check("ha su logs")
+    _LOGGER.info("%s", "\n".join(output))
+
+
+def test_systemctl_status(shell_command):
+    output = shell_command.run_check(
+        "systemctl --no-pager -l status -a || true", timeout=90
+    )
+    _LOGGER.info("%s", "\n".join(output))