debian-forge/test/cases/manifest_tests
Tomáš Hozza 0b158c3fd3 Test/manifest_tests: use temporary dir if workdir is not specified
In case the workdir is not provided to the script explicitly as an
argument, the script will use a temporary directory under /var/tmp as
its workdir. In such case, the workdir will be deleted on exit. This
should mitigate potentially confusing behavior when executing the script
multiple times with different arguments, while never specifying the
workdir.

Signed-off-by: Tomáš Hozza <thozza@redhat.com>
2025-01-31 10:18:14 +01:00

471 lines
18 KiB
Python
Executable file

#!/usr/bin/python3
"""
Test manifest building consistency using osbuild-image-info
The test case downloads the image build cache (manifests, build info and artifacts) for the osbuild/images repository
from S3. It then runs osbuild-image-info on the downloaded image artifact file and produces an image-info report.
The test case then builds the manifest using the current version of osbuild and produces the image artifact.
It runs osbuild-image-info on the newly built image artifact file and produces an image-info report.
Finally, it compares the image-info reports and fails if they are different.
"""
import argparse
import json
import os
import subprocess
import sys
import tempfile
from typing import Dict, List, Optional, Tuple
OSBUILD_IMAGES_REPO_URL = os.environ.get("OSBUILD_IMAGES_REPO_URL", "https://github.com/osbuild/images.git")
OS_RELEASE_FILE = "/etc/os-release"
def read_osrelease() -> Dict[str, str]:
"""Read Operating System Information from `os-release`
This creates a dictionary with information describing the running operating system. It reads the information from
the path array provided as `paths`. The first available file takes precedence. It must be formatted according to
the rules in `os-release(5)`.
"""
osrelease = {}
with open(OS_RELEASE_FILE, encoding="utf-8") as orf:
for line in orf:
line = line.strip()
if not line:
continue
if line[0] == "#":
continue
key, value = line.split("=", 1)
osrelease[key] = value.strip('"')
return osrelease
def get_host_distro() -> str:
"""
Get the host distro version based on data in the os-release file.
The format is <distro>-<version> (e.g. fedora-41).
"""
osrelease = read_osrelease()
return f"{osrelease['ID']}-{osrelease['VERSION_ID']}"
def get_host_arch() -> str:
"""
Get the host architecture.
"""
return os.uname().machine
def manifest_pipeline_names(path: str) -> List[str]:
"""
Read the manifest file and return a list of pipeline names
"""
with open(path, "r", encoding="utf-8") as f:
manifest = json.load(f)
return [pipeline["name"] for pipeline in manifest["pipelines"]]
def find_manifest_file(build_path: str) -> str:
"""
Return the path to the manifest file in the build directory
"""
return os.path.join(build_path, "manifest.json")
def find_image_file(build_path: str, export_name: Optional[str] = None) -> str:
"""
Find the path to the image by searching for the file under the directory named 'export_name'. If the name of the
export pipeline is not provided, determine it by reading the manifest to get the name of the last pipeline.
Raises RuntimeError if no or multiple files are found in the expected path.
"""
if export_name is None:
manifest_file = find_manifest_file(build_path)
export_name = manifest_pipeline_names(manifest_file)[-1]
files = os.listdir(os.path.join(build_path, export_name))
if len(files) > 1:
error = "Multiple files found in build path while searching for image file"
error += "\n".join(files)
raise RuntimeError(error)
if len(files) == 0:
raise RuntimeError("No found in build path while searching for image file")
return os.path.join(build_path, export_name, files[0])
def checkout_images_repo(ref, workdir: os.PathLike) -> str:
"""
Checkout the 'images' repository at a specific commit and return the path to the directory
If the repository is already checked-out, switch to the specified commit.
"""
images_path = os.path.join(workdir, "images")
if not os.path.exists(images_path):
print(f"Checking out '{OSBUILD_IMAGES_REPO_URL}' repository at ref '{ref}'")
try:
subprocess.check_call(
["git", "clone", "--depth=1", "--no-single-branch", OSBUILD_IMAGES_REPO_URL, "images"],
cwd=workdir, stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
)
except subprocess.CalledProcessError as e:
print(f"Failed to clone 'images' repository: {e.stdout.decode()}")
sys.exit(1)
else:
print(f"'images' repository is already checked-out at '{images_path}'")
subprocess.check_call(["git", "checkout", ref], cwd=images_path, stdout=subprocess.DEVNULL)
return images_path
def download_image_build_artifact(images_path: os.PathLike, build_dir: os.PathLike) -> None:
"""
Download the image build artifact from S3 for a specific image build.
"""
cmd = ["./test/scripts/dl-one-image-build-cache", build_dir]
print(" ".join(cmd))
try:
subprocess.check_call(cmd, cwd=images_path, env=os.environ)
except subprocess.CalledProcessError as _:
print("⚠️ Failed to download image build cache")
sys.exit(1)
def download_image_build_cache_md(
images_path, output_dir, distros: List[str], arch: str, configs: Optional[List[str]] = None,
image_types: Optional[List[str]] = None, skip_img_types: Optional[List[str]] = None) -> None:
"""
Download the image build cache metadata from S3 for a specific distro / arch / configs.
The image artifacts is not downloaded, only the metadata.
"""
cmd = [
"./test/scripts/dl-image-build-cache",
"--arch", arch,
"--output", output_dir,
]
for distro in distros:
cmd += ["--distro", distro]
for config in configs or []:
cmd += ["--config", config]
# The image_types and skip_img_types are mutually exclusive,
# but this is enforced in the argument parser.
for imag_type in image_types or []:
cmd += ["--image-type", imag_type]
for skip_img_type in skip_img_types or []:
cmd += ["--skip-image-type", skip_img_type]
print(" ".join(cmd))
try:
subprocess.check_call(cmd, cwd=images_path, env=os.environ)
except subprocess.CalledProcessError as _:
print("⚠️ Failed to download image build cache")
sys.exit(1)
def gen_image_info_report(image_path: str, report_path: str) -> None:
"""
Run osbuild-image-info on the image file and return the parsed JSON output
"""
cmd = ["sudo", "osbuild-image-info", image_path]
output = subprocess.check_output(cmd)
with open(report_path, "w", encoding="utf-8") as f:
f.write(output.decode())
def diff_files(file1: str, file2: str) -> str:
"""
Run diff on two files and return the output
"""
cmd = [
"diff", "-u",
"-I", r'"[0-9a-f]\{8\}-\([0-9a-f]\{4\}-\)\{3\}[0-9a-f]\{12\}"', # Ignore UUIDs
"-I", r'"[0-9a-zA-Z]\{6\}-\([0-9a-zA-Z]\{4\}-\)\{5\}[0-9a-zA-Z]\{6\}"', # Ignore LVM quasi-UUIDs
"-I", r'"volid": "[0-9a-f]\{8\}"', # Ignore volid
file1, file2,
]
print(" ".join(cmd))
run = subprocess.run(cmd, stdout=subprocess.PIPE, check=False)
return run.stdout.decode()
class OSBuild:
def __init__(self, store, outdir):
self.store = store
self.outdir = outdir
def run(self, manifest, exports, checkpoints=None) -> Tuple[int, str, str]:
cmd = [
"sudo",
"osbuild",
"--cache-max-size", "unlimited",
"--store", os.fspath(self.store),
"--output-directory", os.fspath(self.outdir),
os.fspath(manifest)
]
for checkpoint in checkpoints or []:
cmd += [
"--checkpoint", checkpoint
]
for export in exports:
cmd += [
"--export", export
]
print(" ".join(cmd))
run = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, check=False)
return run.returncode, run.stdout.decode(), run.stderr.decode()
def run_manifest_behavior_test(
build_dir: str, results_dir: str, osbuild_store: str, rm_artifacts_after_test: bool = False) -> None:
"""
This function implements the test case for a single image build cache directory which contains:
- the manifest.json file
- the image artifact file built by an older version of osbuild pinned in the images repository
The test case does the following:
- runs osbuild-image-info on the downloaded image artifact file and produces image-info report
- builds the manifest using the current version of osbuild and produces image artifact
- runs osbuild-image-info on the newly built image artifact file and produces image-info report
- compares the image-info reports
The test case fails if the image-info reports are different
The function can optionally remove the downloaded and rebuilt image artifacts after the test case
to save disk space.
"""
manifest = find_manifest_file(build_dir)
manifest_pipelines = manifest_pipeline_names(manifest)
downloaded_image = find_image_file(build_dir)
downloaded_image_iminfo = os.path.join(results_dir, "downloaded_image_iminfo.json")
print(f"📜 Generating image info report for downloaded image '{downloaded_image}' to '{downloaded_image_iminfo}'")
gen_image_info_report(downloaded_image, downloaded_image_iminfo)
if rm_artifacts_after_test:
print("🗑️ Removing downloaded image artifact")
os.remove(downloaded_image)
rebuild_dir = os.path.join(build_dir, "rebuild")
os.makedirs(rebuild_dir, exist_ok=True)
osbuild = OSBuild(osbuild_store, rebuild_dir)
print("Rebuilding the image artifact using installed osbuild version")
retcode, stdout, stderr = osbuild.run(manifest, [manifest_pipelines[-1]], ["build"])
with open(os.path.join(rebuild_dir, "image_rebuild_osbuild.log"), "w", encoding="utf-8") as f:
f.write(stdout)
f.write(stderr)
if retcode != 0:
raise RuntimeError(f"Failed to rebuild the image artifact:\nSTDOUT:\n{stdout}\nSTDERR:\n{stderr}")
rebuilt_image = find_image_file(rebuild_dir, manifest_pipelines[-1])
rebuilt_image_iminfo = os.path.join(results_dir, "rebuilt_image_iminfo.json")
print(f"📜 Generating image info report for rebuilt image: '{rebuilt_image}' to '{rebuilt_image_iminfo}'")
gen_image_info_report(rebuilt_image, rebuilt_image_iminfo)
if rm_artifacts_after_test:
print("🗑️ Removing the dir with rebuilt image artifact")
# NB: use sudo to remove the dir, because the files are owned by root after osbuild run
subprocess.run(["sudo", "rm", "-rf", rebuild_dir], check=False)
diff = diff_files(downloaded_image_iminfo, rebuilt_image_iminfo)
with open(os.path.join(results_dir, "iminfo.diff"), "w", encoding="utf-8") as f:
f.write(diff)
if diff:
raise RuntimeError(f"Image info reports are different:\n{diff}")
def get_argparser():
class ExtendAction(argparse.Action):
"""
Custom argparse action to append multiple values to a list option
to prevent overwriting the list with each new value.
This may be removed when Python 3.8 is the minimum supported version.
"""
def __call__(self, parser, namespace, values, option_string=None):
items = getattr(namespace, self.dest) or []
items.extend(values)
setattr(namespace, self.dest, items)
parser = argparse.ArgumentParser(description=__doc__)
parser.register('action', 'extend', ExtendAction)
parser.add_argument(
"--distro",
metavar="DISTRO",
action="extend",
nargs="+",
help="Distro to test. Default is the host distro."
)
parser.add_argument(
"--image-type", action="extend", metavar="TYPE", nargs="+",
help="Image type for which the image build cache is downloaded. Can be specified multiple times. " +
"If not provided, all image types are downloaded. " +
"The option is mutually exclusive with --skip-image-type.",
)
parser.add_argument(
"--skip-image-type", action="extend", metavar="TYPE_GLOB", nargs="+",
help="Image types to skip when downloading the image build cache. Can be specified multiple times. " +
"This is useful for image types which can't be analyzed using osbuild-image-info.",
)
parser.add_argument(
"--config", action="extend", metavar="NAME_GLOB", nargs="+",
help="Config name globs used to filter which image build cache files are downloaded. " +
"Can be specified multiple times. If not provided, all configs are downloaded.",
)
parser.add_argument(
"--images-ref",
metavar="REF",
default="main",
help="Git ref to checkout in the osbuild/images repository."
)
parser.add_argument(
"--osb-store",
metavar="PATH",
type=os.path.abspath,
default=None,
help="Directory where intermediary os trees are stored."
)
parser.add_argument(
"--workdir",
metavar="PATH",
type=os.path.abspath,
help="Working directory where the images repository is checked out and the image build cache is downloaded. " +
"If not provided, a temporary directory will be used and deleted on exit."
)
parser.add_argument(
"--results-dir",
metavar="PATH",
type=os.path.abspath,
default="./osbuild-manifest-tests-results",
help="Directory where the test results are stored."
)
parser.add_argument(
"--chunk-size", nargs=2, metavar=("CHUNK_NUMBER", "TOTAL_CHUNKS"),
type=int, default=None,
help="Run tests only for a specific chunk of the image build cache. " +
"The first argument is the chunk number (starting from 1) and the second argument is the total " +
"number of chunks. This is useful for running the tests in parallel."
)
parser.add_argument(
"--rm-artifacts-after-test",
action="store_true",
default=False,
help="Remove the downloaded and rebuilt image artifacts after the test case. " +
"This is useful for saving disk space."
)
return parser
def main():
parser = get_argparser()
args = parser.parse_args()
if not args.distro:
args.distro = [get_host_distro()]
chunk_number, total_chunks = None, None
if args.chunk_size:
chunk_number, total_chunks = args.chunk_size
if chunk_number < 1 or total_chunks < 1:
parser.error("Both values for '--chunk-size' must be greater than zero.")
if chunk_number > total_chunks:
parser.error("The chunk number must be less than or equal to the total number of chunks.")
if args.image_type and args.skip_image_type:
parser.error("Options --image-type and --skip-image-type are mutually exclusive.")
try:
tmpdir = tempfile.TemporaryDirectory(dir='/var/tmp', prefix='osbuild-manifest-tests-workdir')
workdir = args.workdir or tmpdir
os.makedirs(workdir, exist_ok=True)
print(f"👷 Using working directory: {workdir}")
os.chdir(workdir)
osbuild_store_dir = args.osb_store or os.path.join(workdir, "osbuild-store")
os.makedirs(osbuild_store_dir, exist_ok=True)
print(f"💾 Using osbuild store directory: {osbuild_store_dir}")
# Checkout the images repository, since we will need scripts from it to download the image build cache
images_path = checkout_images_repo(args.images_ref, workdir)
# Create the directory where the image build cache will be downloaded
image_build_cache = os.path.join(workdir, "image-build-cache")
os.makedirs(image_build_cache, exist_ok=True)
download_image_build_cache_md(
images_path, image_build_cache, args.distro, get_host_arch(),
args.config, args.image_type, args.skip_image_type
)
# The test case is run for every directory in the image build cache directory
test_cases = sorted(os.listdir(image_build_cache))
if not test_cases:
print("⚠️ No image build cache directories found -> nothing to test", file=sys.stderr)
sys.exit(1)
print(f"📦 Found {len(test_cases)} image build cache directories")
if chunk_number and total_chunks:
all_test_cases = test_cases
print(f"📦 Will run subset of tests for chunk {chunk_number} of {total_chunks}:")
chunk_size = len(test_cases) // total_chunks
chunk_size_remainder = len(test_cases) % total_chunks
# determine the chunk index range
start = 0
end = 0
for i in range(chunk_number):
current_chunk_size = chunk_size + 1 if i < chunk_size_remainder else chunk_size
start = end
end = start + current_chunk_size
test_cases = test_cases[start:end]
idx_range = range(start, end)
for i, test_case in enumerate(all_test_cases):
print(f" {'🟢' if i in idx_range else '🚫'} {test_case}")
# Dictionary holding the test case name as key and a boolean indicating if the test case failed as value
test_cases_failed: Dict[str, bool] = {}
print(f"🏃 Running {len(test_cases)} test cases:\n{os.linesep.join(test_cases)}")
for test_case in test_cases:
print(f"🏃 Running test case for {test_case}")
test_case_build_dir = os.path.join(image_build_cache, test_case)
download_image_build_artifact(images_path, test_case_build_dir)
test_case_results_dir = os.path.join(args.results_dir, test_case)
os.makedirs(test_case_results_dir, exist_ok=True)
try:
run_manifest_behavior_test(
test_case_build_dir, test_case_results_dir, osbuild_store_dir, args.rm_artifacts_after_test)
# pylint: disable=broad-exception-caught
except Exception as e:
print(f"{test_case} FAILED")
print(e)
test_cases_failed[test_case] = True
else:
print(f"{test_case} PASSED")
test_cases_failed[test_case] = False
finally:
# We can't use the context manager here, because some files in it may be owned by root
subprocess.run(["sudo", "rm", "-rf", tmpdir.name], check=False)
print("Test results:")
for test_case, failed in test_cases_failed.items():
print(f" {'' if failed else ''} {test_case}")
if any(test_cases_failed.values()):
sys.exit(1)
if __name__ == "__main__":
main()