ClickHouse/tests/ci/functional_test_check.py

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

427 lines
14 KiB
Python
Raw Normal View History

2021-10-27 07:21:48 +00:00
#!/usr/bin/env python3
2021-10-28 07:29:13 +00:00
2022-03-18 20:32:59 +00:00
import argparse
2021-10-27 07:21:48 +00:00
import csv
import logging
import os
import re
2022-03-18 20:32:59 +00:00
import subprocess
2021-10-27 07:21:48 +00:00
import sys
import atexit
from pathlib import Path
2022-11-16 09:01:17 +00:00
from typing import List, Tuple
2021-10-27 07:21:48 +00:00
2021-10-28 07:39:16 +00:00
from github import Github
2021-11-12 11:07:54 +00:00
from build_download_helper import download_all_deb_packages
from clickhouse_helper import (
2023-08-16 20:53:51 +00:00
CiLogsCredentials,
ClickHouseHelper,
prepare_tests_results_for_clickhouse,
)
from commit_status_helper import (
NotSet,
RerunHelper,
get_commit,
override_status,
post_commit_status,
post_commit_status_to_file,
update_mergeable_check,
)
2023-08-16 20:53:51 +00:00
from docker_pull_helper import DockerImage, get_image_with_version
from download_release_packages import download_last_release
from env_helper import TEMP_PATH, REPO_COPY, REPORTS_PATH
from get_robot_token import get_best_robot_token
from pr_info import FORCE_TESTS_LABEL, PRInfo
from report import TestResults, read_test_results
from s3_helper import S3Helper
from stopwatch import Stopwatch
2021-12-03 08:33:16 +00:00
from tee_popen import TeePopen
from upload_result_helper import upload_results
2021-10-27 07:21:48 +00:00
NO_CHANGES_MSG = "Nothing to run"
2022-03-18 12:36:45 +00:00
def get_additional_envs(check_name, run_by_hash_num, run_by_hash_total):
result = []
if "DatabaseReplicated" in check_name:
result.append("USE_DATABASE_REPLICATED=1")
if "DatabaseOrdinary" in check_name:
result.append("USE_DATABASE_ORDINARY=1")
if "wide parts enabled" in check_name:
result.append("USE_POLYMORPHIC_PARTS=1")
2023-02-03 13:34:18 +00:00
if "ParallelReplicas" in check_name:
result.append("USE_PARALLEL_REPLICAS=1")
if "s3 storage" in check_name:
result.append("USE_S3_STORAGE_FOR_MERGE_TREE=1")
2023-04-12 15:18:28 +00:00
if "analyzer" in check_name:
result.append("USE_NEW_ANALYZER=1")
2021-11-30 11:40:19 +00:00
if run_by_hash_total != 0:
result.append(f"RUN_BY_HASH_NUM={run_by_hash_num}")
result.append(f"RUN_BY_HASH_TOTAL={run_by_hash_total}")
return result
2021-10-27 07:21:48 +00:00
2021-10-27 07:21:48 +00:00
def get_image_name(check_name):
if "stateless" in check_name.lower():
return "clickhouse/stateless-test"
if "stateful" in check_name.lower():
return "clickhouse/stateful-test"
2021-10-27 08:02:30 +00:00
else:
raise Exception(f"Cannot deduce image name based on check name {check_name}")
2021-10-27 07:21:48 +00:00
2022-03-10 13:56:59 +00:00
def get_run_command(
2023-08-16 20:53:51 +00:00
check_name: str,
builds_path: str,
repo_path: str,
result_path: str,
server_log_path: str,
kill_timeout: int,
additional_envs: List[str],
ci_logs_args: str,
image: DockerImage,
flaky_check: bool,
tests_to_run: List[str],
) -> str:
additional_options = ["--hung-check"]
additional_options.append("--print-time")
2021-11-10 13:49:19 +00:00
if tests_to_run:
additional_options += tests_to_run
additional_options_str = (
'-e ADDITIONAL_OPTIONS="' + " ".join(additional_options) + '"'
)
2021-10-27 07:21:48 +00:00
envs = [
f"-e MAX_RUN_TIME={int(0.9 * kill_timeout)}",
2022-08-12 07:46:36 +00:00
# a static link, don't use S3_URL or S3_DOWNLOAD
'-e S3_URL="https://s3.amazonaws.com/clickhouse-datasets"',
]
2021-11-10 13:49:19 +00:00
if flaky_check:
2023-08-16 20:53:51 +00:00
envs.append("-e NUM_TRIES=100")
envs.append("-e MAX_RUN_TIME=1800")
2021-11-10 13:49:19 +00:00
envs += [f"-e {e}" for e in additional_envs]
2021-11-10 13:49:19 +00:00
env_str = " ".join(envs)
2023-04-18 00:56:29 +00:00
volume_with_broken_test = (
2023-08-16 20:53:51 +00:00
f"--volume={repo_path}/tests/analyzer_tech_debt.txt:/analyzer_tech_debt.txt "
2023-04-18 00:56:29 +00:00
if "analyzer" in check_name
else ""
)
2021-10-27 07:21:48 +00:00
return (
f"docker run --volume={builds_path}:/package_folder "
2023-08-16 20:53:51 +00:00
f"{ci_logs_args}"
2023-08-06 02:38:04 +00:00
f"--volume={repo_path}/tests:/usr/share/clickhouse-test "
2023-08-16 20:53:51 +00:00
f"{volume_with_broken_test}"
f"--volume={result_path}:/test_output "
f"--volume={server_log_path}:/var/log/clickhouse-server "
2021-10-27 09:58:21 +00:00
f"--cap-add=SYS_PTRACE {env_str} {additional_options_str} {image}"
)
2021-10-27 07:21:48 +00:00
2023-08-16 20:53:51 +00:00
def get_tests_to_run(pr_info: PRInfo) -> List[str]:
result = set()
2021-11-10 13:49:19 +00:00
if pr_info.changed_files is None:
return []
for fpath in pr_info.changed_files:
if re.match(r"tests/queries/0_stateless/[0-9]{5}", fpath):
logging.info("File '%s' is changed and seems like a test", fpath)
fname = fpath.split("/")[3]
2021-11-10 13:49:19 +00:00
fname_without_ext = os.path.splitext(fname)[0]
# add '.' to the end of the test name not to run all tests with the same prefix
# e.g. we changed '00001_some_name.reference'
# and we have ['00001_some_name.sh', '00001_some_name_2.sql']
# so we want to run only '00001_some_name.sh'
result.add(fname_without_ext + ".")
elif "tests/queries/" in fpath:
# log suspicious changes from tests/ for debugging in case of any problems
logging.info("File '%s' is changed, but it doesn't look like a test", fpath)
2021-11-10 13:49:19 +00:00
return list(result)
2022-11-16 09:01:17 +00:00
def process_results(
result_folder: str,
server_log_path: str,
) -> Tuple[str, str, TestResults, List[str]]:
test_results = [] # type: TestResults
2021-10-27 07:21:48 +00:00
additional_files = []
# Just upload all files from result_folder.
# If task provides processed results, then it's responsible for content of result_folder.
if os.path.exists(result_folder):
test_files = [
f
for f in os.listdir(result_folder)
if os.path.isfile(os.path.join(result_folder, f))
]
2021-10-27 07:21:48 +00:00
additional_files = [os.path.join(result_folder, f) for f in test_files]
if os.path.exists(server_log_path):
server_log_files = [
f
for f in os.listdir(server_log_path)
if os.path.isfile(os.path.join(server_log_path, f))
]
additional_files = additional_files + [
os.path.join(server_log_path, f) for f in server_log_files
]
2021-10-27 07:21:48 +00:00
2021-12-12 12:09:44 +00:00
status = []
2021-10-27 07:21:48 +00:00
status_path = os.path.join(result_folder, "check_status.tsv")
2021-12-12 12:09:44 +00:00
if os.path.exists(status_path):
logging.info("Found test_results.tsv")
with open(status_path, "r", encoding="utf-8") as status_file:
status = list(csv.reader(status_file, delimiter="\t"))
2021-10-28 07:29:13 +00:00
2021-10-27 07:21:48 +00:00
if len(status) != 1 or len(status[0]) != 2:
logging.info("Files in result folder %s", os.listdir(result_folder))
2021-10-27 07:21:48 +00:00
return "error", "Invalid check_status.tsv", test_results, additional_files
state, description = status[0][0], status[0][1]
try:
results_path = Path(result_folder) / "test_results.tsv"
if results_path.exists():
logging.info("Found test_results.tsv")
else:
logging.info("Files in result folder %s", os.listdir(result_folder))
return "error", "Not found test_results.tsv", test_results, additional_files
test_results = read_test_results(results_path)
if len(test_results) == 0:
return "error", "Empty test_results.tsv", test_results, additional_files
except Exception as e:
return (
"error",
f"Cannot parse test_results.tsv ({e})",
test_results,
additional_files,
)
2021-10-27 07:21:48 +00:00
return state, description, test_results, additional_files
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument("check_name")
parser.add_argument("kill_timeout", type=int)
parser.add_argument(
"--validate-bugfix",
action="store_true",
help="Check that added tests failed on latest stable",
)
parser.add_argument(
"--post-commit-status",
default="commit_status",
choices=["commit_status", "file"],
help="Where to public post commit status",
)
return parser.parse_args()
def main():
2021-10-27 07:21:48 +00:00
logging.basicConfig(level=logging.INFO)
2021-11-19 14:47:04 +00:00
stopwatch = Stopwatch()
2021-11-26 14:00:09 +00:00
temp_path = TEMP_PATH
repo_path = REPO_COPY
reports_path = REPORTS_PATH
post_commit_path = os.path.join(temp_path, "functional_commit_status.tsv")
2021-10-27 07:21:48 +00:00
args = parse_args()
check_name = args.check_name
kill_timeout = args.kill_timeout
validate_bugfix_check = args.validate_bugfix
flaky_check = "flaky" in check_name.lower()
2022-03-08 15:19:05 +00:00
run_changed_tests = flaky_check or validate_bugfix_check
gh = Github(get_best_robot_token(), per_page=100)
2021-12-01 14:23:51 +00:00
# For validate_bugfix_check we need up to date information about labels, so pr_event_from_api is used
2022-10-27 11:32:42 +00:00
pr_info = PRInfo(
need_changed_files=run_changed_tests, pr_event_from_api=validate_bugfix_check
2022-10-27 11:32:42 +00:00
)
2021-12-01 14:23:51 +00:00
commit = get_commit(gh, pr_info.sha)
atexit.register(update_mergeable_check, gh, pr_info, check_name)
if not os.path.exists(temp_path):
os.makedirs(temp_path)
if validate_bugfix_check and "pr-bugfix" not in pr_info.labels:
if args.post_commit_status == "file":
post_commit_status_to_file(
post_commit_path,
2022-10-26 14:02:04 +00:00
f"Skipped (no pr-bugfix in {pr_info.labels})",
"success",
"null",
)
2022-10-26 14:02:04 +00:00
logging.info("Skipping '%s' (no pr-bugfix in %s)", check_name, pr_info.labels)
sys.exit(0)
2022-03-10 16:39:18 +00:00
if "RUN_BY_HASH_NUM" in os.environ:
2022-11-16 09:01:17 +00:00
run_by_hash_num = int(os.getenv("RUN_BY_HASH_NUM", "0"))
run_by_hash_total = int(os.getenv("RUN_BY_HASH_TOTAL", "0"))
check_name_with_group = (
check_name + f" [{run_by_hash_num + 1}/{run_by_hash_total}]"
)
else:
run_by_hash_num = 0
run_by_hash_total = 0
check_name_with_group = check_name
rerun_helper = RerunHelper(commit, check_name_with_group)
2021-12-01 14:23:51 +00:00
if rerun_helper.is_already_finished_by_status():
logging.info("Check is already finished according to github status, exiting")
sys.exit(0)
2021-10-27 07:21:48 +00:00
2021-11-10 13:49:19 +00:00
tests_to_run = []
2022-03-08 15:19:05 +00:00
if run_changed_tests:
2021-11-10 13:49:19 +00:00
tests_to_run = get_tests_to_run(pr_info)
if not tests_to_run:
state = override_status("success", check_name, validate_bugfix_check)
if args.post_commit_status == "commit_status":
post_commit_status(
commit,
state,
NotSet,
NO_CHANGES_MSG,
check_name_with_group,
pr_info,
)
elif args.post_commit_status == "file":
post_commit_status_to_file(
post_commit_path,
description=NO_CHANGES_MSG,
state=state,
report_url="null",
)
2021-11-10 13:49:19 +00:00
sys.exit(0)
2021-10-28 20:20:30 +00:00
2021-10-27 07:21:48 +00:00
image_name = get_image_name(check_name)
docker_image = get_image_with_version(reports_path, image_name)
2021-10-27 07:21:48 +00:00
packages_path = os.path.join(temp_path, "packages")
if not os.path.exists(packages_path):
os.makedirs(packages_path)
if validate_bugfix_check:
download_last_release(packages_path)
else:
download_all_deb_packages(check_name, reports_path, packages_path)
2021-11-12 11:07:54 +00:00
2021-10-27 07:21:48 +00:00
server_log_path = os.path.join(temp_path, "server_log")
if not os.path.exists(server_log_path):
os.makedirs(server_log_path)
result_path = os.path.join(temp_path, "result_path")
if not os.path.exists(result_path):
os.makedirs(result_path)
2022-12-26 15:29:32 +00:00
run_log_path = os.path.join(result_path, "run.log")
2021-10-27 07:21:48 +00:00
additional_envs = get_additional_envs(
check_name, run_by_hash_num, run_by_hash_total
)
if validate_bugfix_check:
additional_envs.append("GLOBAL_TAGS=no-random-settings")
2023-08-16 20:53:51 +00:00
ci_logs_credentials = CiLogsCredentials(Path(temp_path) / "export-logs-config.sh")
ci_logs_args = ci_logs_credentials.get_docker_arguments(
pr_info, stopwatch.start_time_str, check_name
)
run_command = get_run_command(
2023-04-18 00:12:30 +00:00
check_name,
packages_path,
2023-08-06 02:38:04 +00:00
repo_path,
result_path,
server_log_path,
kill_timeout,
additional_envs,
2023-08-16 20:53:51 +00:00
ci_logs_args,
docker_image,
flaky_check,
tests_to_run,
)
2021-10-27 07:21:48 +00:00
logging.info("Going to run func tests: %s", run_command)
2021-12-03 08:33:16 +00:00
with TeePopen(run_command, run_log_path) as process:
retcode = process.wait()
if retcode == 0:
logging.info("Run successfully")
else:
logging.info("Run failed")
2021-10-27 07:21:48 +00:00
2023-06-19 16:17:34 +00:00
try:
subprocess.check_call(f"sudo chown -R ubuntu:ubuntu {temp_path}", shell=True)
except subprocess.CalledProcessError:
logging.warning("Failed to change files owner in %s, ignoring it", temp_path)
2021-10-27 07:21:48 +00:00
2023-08-16 20:53:51 +00:00
ci_logs_credentials.clean_ci_logs_from_credentials(Path(run_log_path))
2022-08-11 13:01:32 +00:00
s3_helper = S3Helper()
2022-01-10 16:45:17 +00:00
state, description, test_results, additional_logs = process_results(
result_path, server_log_path
)
state = override_status(state, check_name, invert=validate_bugfix_check)
2021-11-19 14:47:04 +00:00
ch_helper = ClickHouseHelper()
report_url = upload_results(
s3_helper,
pr_info.number,
pr_info.sha,
test_results,
[run_log_path] + additional_logs,
check_name_with_group,
)
2021-11-12 12:36:25 +00:00
2022-03-18 12:36:45 +00:00
print(f"::notice:: {check_name} Report url: {report_url}")
if args.post_commit_status == "commit_status":
post_commit_status(
commit, state, report_url, description, check_name_with_group, pr_info
)
elif args.post_commit_status == "file":
post_commit_status_to_file(
post_commit_path,
description,
state,
report_url,
)
2022-03-18 12:36:45 +00:00
else:
raise Exception(
f'Unknown post_commit_status option "{args.post_commit_status}"'
)
prepared_events = prepare_tests_results_for_clickhouse(
pr_info,
test_results,
state,
stopwatch.duration_seconds,
stopwatch.start_time_str,
report_url,
check_name_with_group,
)
2022-03-29 19:06:50 +00:00
ch_helper.insert_events_into(db="default", table="checks", events=prepared_events)
2021-11-25 10:01:29 +00:00
if state != "success":
if FORCE_TESTS_LABEL in pr_info.labels:
2022-04-21 14:33:46 +00:00
print(f"'{FORCE_TESTS_LABEL}' enabled, will report success")
2021-12-01 09:50:36 +00:00
else:
sys.exit(1)
if __name__ == "__main__":
main()