ClickHouse/tests/ci/build_check.py

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

455 lines
13 KiB
Python
Raw Normal View History

2021-10-21 11:09:15 +00:00
#!/usr/bin/env python3
from pathlib import Path
from typing import Tuple
2021-10-21 11:09:15 +00:00
import subprocess
import logging
import os
import sys
import time
from ci_config import CI_CONFIG, BuildConfig
from ccache_utils import CargoCache
from docker_pull_helper import get_image_with_version
from env_helper import (
GITHUB_JOB,
IMAGES_PATH,
REPO_COPY,
S3_BUILDS_BUCKET,
2022-08-11 13:01:32 +00:00
S3_DOWNLOAD,
TEMP_PATH,
)
from git_helper import Git, git_runner
2021-11-26 14:00:09 +00:00
from pr_info import PRInfo
from report import BuildResult, FAILURE, StatusType, SUCCESS
from s3_helper import S3Helper
from tee_popen import TeePopen
2022-01-12 12:17:21 +00:00
from version_helper import (
ClickHouseVersion,
get_version_from_repo,
update_version_local,
)
from clickhouse_helper import (
ClickHouseHelper,
2023-08-29 18:09:15 +00:00
CiLogsCredentials,
prepare_tests_results_for_clickhouse,
get_instance_type,
get_instance_id,
)
from stopwatch import Stopwatch
2021-10-21 11:09:15 +00:00
2022-02-08 18:12:04 +00:00
IMAGE_NAME = "clickhouse/binary-builder"
BUILD_LOG_NAME = "build_log.log"
2022-02-08 18:12:04 +00:00
2021-10-21 11:09:15 +00:00
2022-01-12 12:17:21 +00:00
def _can_export_binaries(build_config: BuildConfig) -> bool:
if build_config.package_type != "deb":
2021-10-21 11:09:15 +00:00
return False
if build_config.sanitizer != "":
2021-10-21 11:09:15 +00:00
return True
if build_config.debug_build:
2021-10-21 11:09:15 +00:00
return True
return False
def get_packager_cmd(
2022-01-12 12:17:21 +00:00
build_config: BuildConfig,
packager_path: str,
output_path: Path,
cargo_cache_dir: Path,
2022-01-12 12:17:21 +00:00
build_version: str,
image_version: str,
2022-03-30 14:29:13 +00:00
official: bool,
2022-01-12 12:17:21 +00:00
) -> str:
package_type = build_config.package_type
comp = build_config.compiler
cmake_flags = "-DENABLE_CLICKHOUSE_SELF_EXTRACTING=1"
cmd = (
f"cd {packager_path} && CMAKE_FLAGS='{cmake_flags}' ./packager "
f"--output-dir={output_path} --package-type={package_type} --compiler={comp}"
)
if build_config.debug_build:
cmd += " --debug-build"
if build_config.sanitizer:
cmd += f" --sanitizer={build_config.sanitizer}"
if build_config.tidy:
cmd += " --clang-tidy"
2023-03-24 16:24:21 +00:00
cmd += " --cache=sccache"
cmd += " --s3-rw-access"
cmd += f" --s3-bucket={S3_BUILDS_BUCKET}"
cmd += f" --cargo-cache-dir={cargo_cache_dir}"
if build_config.additional_pkgs:
2022-03-14 21:07:58 +00:00
cmd += " --additional-pkgs"
cmd += f" --docker-image-version={image_version}"
cmd += " --with-profiler"
cmd += f" --version={build_version}"
2021-10-21 11:09:15 +00:00
if _can_export_binaries(build_config):
cmd += " --with-binaries=tests"
2021-10-21 11:09:15 +00:00
2022-03-30 14:29:13 +00:00
if official:
cmd += " --official"
2021-10-21 11:09:15 +00:00
return cmd
2022-01-12 12:17:21 +00:00
def build_clickhouse(
2023-09-05 10:59:31 +00:00
packager_cmd: str, logs_path: Path, build_output_path: Path
) -> Tuple[Path, StatusType]:
build_log_path = logs_path / BUILD_LOG_NAME
2022-01-12 12:17:21 +00:00
success = False
2021-12-03 08:33:16 +00:00
with TeePopen(packager_cmd, build_log_path) as process:
retcode = process.wait()
if build_output_path.exists():
2021-12-14 10:40:03 +00:00
build_results = os.listdir(build_output_path)
else:
build_results = []
2021-10-21 11:09:15 +00:00
if retcode == 0:
2023-09-05 10:59:31 +00:00
if len(build_results) > 0:
2022-01-12 12:17:21 +00:00
success = True
2021-12-14 10:40:03 +00:00
logging.info("Built successfully")
else:
logging.info(
"Success exit code, but no build artifacts => build failed"
)
2021-10-21 11:09:15 +00:00
else:
logging.info("Build failed")
return build_log_path, SUCCESS if success else FAILURE
2021-10-21 11:09:15 +00:00
2021-12-01 14:23:51 +00:00
def check_for_success_run(
s3_helper: S3Helper,
s3_prefix: str,
build_name: str,
version: ClickHouseVersion,
2022-11-15 13:03:16 +00:00
) -> None:
# TODO: Remove after S3 artifacts
# the final empty argument is necessary for distinguish build and build_suffix
logged_prefix = os.path.join(S3_BUILDS_BUCKET, s3_prefix, "")
logging.info("Checking for artifacts in %s", logged_prefix)
2021-12-01 14:23:51 +00:00
try:
# Performance artifacts are now part of regular build, so we're safe
build_results = s3_helper.list_prefix(s3_prefix)
2021-12-01 14:23:51 +00:00
except Exception as ex:
logging.info("Got exception while listing %s: %s\nRerun", logged_prefix, ex)
return
if build_results is None or len(build_results) == 0:
logging.info("Nothing found in %s, rerun", logged_prefix)
return
logging.info("Some build results found:\n%s", build_results)
build_urls = []
log_url = ""
for url in build_results:
url_escaped = url.replace("+", "%2B").replace(" ", "%20")
if BUILD_LOG_NAME in url:
2022-08-11 13:01:32 +00:00
log_url = f"{S3_DOWNLOAD}/{S3_BUILDS_BUCKET}/{url_escaped}"
else:
2022-08-11 13:01:32 +00:00
build_urls.append(f"{S3_DOWNLOAD}/{S3_BUILDS_BUCKET}/{url_escaped}")
if not log_url:
# log is uploaded the last, so if there's no log we need to rerun the build
return
success = len(build_urls) > 0
build_result = BuildResult(
build_name,
log_url,
build_urls,
version.describe,
SUCCESS if success else FAILURE,
0,
GITHUB_JOB,
)
build_result.write_json(Path(TEMP_PATH))
# Fail build job if not successeded
if not success:
sys.exit(1)
else:
sys.exit(0)
2021-12-01 14:23:51 +00:00
def get_release_or_pr(pr_info: PRInfo, version: ClickHouseVersion) -> Tuple[str, str]:
"Return prefixes for S3 artifacts paths"
# FIXME performance
# performance builds are havily relies on a fixed path for artifacts, that's why
# we need to preserve 0 for anything but PR number
# It should be fixed in performance-comparison image eventually
# For performance tests we always set PRs prefix
performance_pr = "PRs/0"
2022-01-12 12:17:21 +00:00
if "release" in pr_info.labels or "release-lts" in pr_info.labels:
# for release pull requests we use branch names prefixes, not pr numbers
return pr_info.head_ref, performance_pr
if pr_info.number == 0:
# for pushes to master - major version
return f"{version.major}.{version.minor}", performance_pr
2022-01-12 12:17:21 +00:00
# PR number for anything else
pr_number = f"PRs/{pr_info.number}"
return pr_number, pr_number
2022-01-12 12:17:21 +00:00
def upload_master_static_binaries(
pr_info: PRInfo,
build_config: BuildConfig,
s3_helper: S3Helper,
build_output_path: Path,
2022-11-15 13:03:16 +00:00
) -> None:
"""Upload binary artifacts to a static S3 links"""
static_binary_name = build_config.static_binary_name
if pr_info.number != 0:
return
2022-01-13 12:55:51 +00:00
elif not static_binary_name:
return
elif pr_info.base_ref != "master":
return
2022-01-13 12:55:51 +00:00
s3_path = "/".join((pr_info.base_ref, static_binary_name, "clickhouse"))
binary = build_output_path / "clickhouse"
url = s3_helper.upload_build_file_to_s3(binary, s3_path)
print(f"::notice ::Binary static URL: {url}")
2022-01-12 12:17:21 +00:00
def main():
2021-10-21 11:09:15 +00:00
logging.basicConfig(level=logging.INFO)
stopwatch = Stopwatch()
build_name = sys.argv[1]
2021-10-21 11:09:15 +00:00
build_config = CI_CONFIG.build_config[build_name]
2021-10-21 11:09:15 +00:00
temp_path = Path(TEMP_PATH)
os.makedirs(temp_path, exist_ok=True)
2021-10-21 11:09:15 +00:00
2021-11-26 14:00:09 +00:00
pr_info = PRInfo()
2021-10-21 11:09:15 +00:00
2022-01-12 12:17:21 +00:00
logging.info("Repo copy path %s", REPO_COPY)
2021-10-21 11:09:15 +00:00
2022-08-11 13:01:32 +00:00
s3_helper = S3Helper()
2021-12-01 14:23:51 +00:00
version = get_version_from_repo(git=Git(True))
release_or_pr, performance_pr = get_release_or_pr(pr_info, version)
2021-12-01 14:23:51 +00:00
s3_path_prefix = "/".join((release_or_pr, pr_info.sha, build_name))
# FIXME performance
s3_performance_path = "/".join(
2023-01-09 00:58:42 +00:00
(performance_pr, pr_info.sha, build_name, "performance.tar.zst")
)
2021-12-01 14:23:51 +00:00
# If this is rerun, then we try to find already created artifacts and just
# put them as github actions artifact (result)
check_for_success_run(s3_helper, s3_path_prefix, build_name, version)
2021-10-21 11:09:15 +00:00
2022-02-08 18:12:04 +00:00
docker_image = get_image_with_version(IMAGES_PATH, IMAGE_NAME)
2021-11-12 12:36:25 +00:00
image_version = docker_image.version
2021-10-21 11:09:15 +00:00
2022-01-28 13:39:23 +00:00
logging.info("Got version from repo %s", version.string)
2022-03-30 14:29:13 +00:00
official_flag = pr_info.number == 0
2022-05-13 17:22:23 +00:00
version_type = "testing"
if "release" in pr_info.labels or "release-lts" in pr_info.labels:
version_type = "stable"
2022-03-30 14:29:13 +00:00
official_flag = True
update_version_local(version, version_type)
logging.info("Updated local files with version")
2021-10-21 11:09:15 +00:00
logging.info("Build short name %s", build_name)
2021-10-21 11:09:15 +00:00
build_output_path = temp_path / build_name
os.makedirs(build_output_path, exist_ok=True)
cargo_cache = CargoCache(
temp_path / "cargo_cache" / "registry", temp_path, s3_helper
)
cargo_cache.download()
2021-10-21 11:09:15 +00:00
packager_cmd = get_packager_cmd(
build_config,
2022-01-12 12:17:21 +00:00
os.path.join(REPO_COPY, "docker/packager"),
build_output_path,
cargo_cache.directory,
2022-01-28 13:39:23 +00:00
version.string,
image_version,
2022-04-06 08:15:36 +00:00
official_flag,
)
2022-03-30 14:29:13 +00:00
2021-10-21 11:09:15 +00:00
logging.info("Going to run packager with %s", packager_cmd)
logs_path = temp_path / "build_log"
os.makedirs(logs_path, exist_ok=True)
2021-10-21 11:09:15 +00:00
2021-10-21 14:41:07 +00:00
start = time.time()
log_path, build_status = build_clickhouse(
packager_cmd, logs_path, build_output_path
)
2021-10-21 14:41:07 +00:00
elapsed = int(time.time() - start)
subprocess.check_call(
f"sudo chown -R ubuntu:ubuntu {build_output_path}", shell=True
)
logging.info("Build finished as %s, log path %s", build_status, log_path)
if build_status == SUCCESS:
cargo_cache.upload()
else:
# We check if docker works, because if it's down, it's infrastructure
try:
subprocess.check_call("docker info", shell=True)
except subprocess.CalledProcessError:
logging.error(
"The dockerd looks down, won't upload anything and generate report"
)
sys.exit(1)
2021-10-21 11:09:15 +00:00
# FIXME performance
2022-05-19 20:44:10 +00:00
performance_urls = []
performance_path = build_output_path / "performance.tar.zst"
if performance_path.exists():
2022-05-19 20:44:10 +00:00
performance_urls.append(
s3_helper.upload_build_file_to_s3(performance_path, s3_performance_path)
)
logging.info(
2023-01-09 00:58:42 +00:00
"Uploaded performance.tar.zst to %s, now delete to avoid duplication",
2022-05-19 20:44:10 +00:00
performance_urls[0],
)
os.remove(performance_path)
2022-05-19 20:44:10 +00:00
build_urls = (
2023-08-29 14:35:53 +00:00
s3_helper.upload_build_directory_to_s3(
2022-05-19 20:44:10 +00:00
build_output_path,
s3_path_prefix,
keep_dirs_in_s3_path=False,
upload_symlinks=False,
)
+ performance_urls
)
2021-10-21 12:37:19 +00:00
logging.info("Got build URLs %s", build_urls)
2021-10-21 12:46:25 +00:00
print("::notice ::Build URLs: {}".format("\n".join(build_urls)))
2021-10-21 12:46:25 +00:00
if log_path.exists():
log_url = s3_helper.upload_build_file_to_s3(
log_path, s3_path_prefix + "/" + log_path.name
)
logging.info("Log url %s", log_url)
else:
logging.info("Build log doesn't exist")
print(f"::notice ::Log URL: {log_url}")
2021-10-21 11:09:15 +00:00
build_result = BuildResult(
build_name,
log_url,
build_urls,
version.describe,
build_status,
elapsed,
GITHUB_JOB,
)
result_json_path = build_result.write_json(temp_path)
logging.info(
"Build result file %s is written, content:\n %s",
result_json_path,
result_json_path.read_text(encoding="utf-8"),
)
upload_master_static_binaries(pr_info, build_config, s3_helper, build_output_path)
# Upload profile data
ch_helper = ClickHouseHelper()
2023-08-29 18:09:15 +00:00
ci_logs_credentials = CiLogsCredentials(Path("/dev/null"))
if ci_logs_credentials.host:
instance_type = get_instance_type()
instance_id = get_instance_id()
query = f"""INSERT INTO build_time_trace
(
pull_request_number,
commit_sha,
check_start_time,
check_name,
instance_type,
instance_id,
file,
library,
time,
pid,
tid,
ph,
ts,
dur,
cat,
name,
detail,
count,
avgMs,
args_name
)
SELECT {pr_info.number}, '{pr_info.sha}', '{stopwatch.start_time_str}', '{build_name}', '{instance_type}', '{instance_id}', *
FROM input('
file String,
library String,
time DateTime64(6),
pid UInt32,
tid UInt32,
ph String,
ts UInt64,
dur UInt64,
cat String,
name String,
detail String,
count UInt64,
avgMs UInt64,
args_name String')
FORMAT JSONCompactEachRow"""
auth = {
"X-ClickHouse-User": "ci",
2023-08-29 18:09:15 +00:00
"X-ClickHouse-Key": ci_logs_credentials.password,
}
2023-08-29 18:09:15 +00:00
url = f"https://{ci_logs_credentials.host}/"
profiles_dir = temp_path / "profiles_source"
os.makedirs(profiles_dir, exist_ok=True)
logging.info("Processing profile JSON files from {GIT_REPO_ROOT}/build_docker")
git_runner(
"./utils/prepare-time-trace/prepare-time-trace.sh "
f"build_docker {profiles_dir.absolute()}"
)
profile_data_file = temp_path / "profile.json"
with open(profile_data_file, "wb") as profile_fd:
for profile_sourse in os.listdir(profiles_dir):
with open(profiles_dir / profile_sourse, "rb") as ps_fd:
profile_fd.write(ps_fd.read())
logging.info(
"::notice ::Log Uploading profile data, path: %s, size: %s, query: %s",
profile_data_file,
profile_data_file.stat().st_size,
query,
)
ch_helper.insert_file(url, auth, query, profile_data_file)
# Upload statistics to CI database
prepared_events = prepare_tests_results_for_clickhouse(
pr_info,
[],
build_status,
stopwatch.duration_seconds,
stopwatch.start_time_str,
log_url,
f"Build ({build_name})",
)
ch_helper.insert_events_into(db="default", table="checks", events=prepared_events)
# Fail the build job if it didn't succeed
if build_status != SUCCESS:
2021-11-25 10:01:29 +00:00
sys.exit(1)
2022-01-12 12:17:21 +00:00
if __name__ == "__main__":
main()