ClickHouse/tests/integration/ci-runner.py

#!/usr/bin/env python3

from collections import defaultdict
import csv
import glob
import json
import logging
import os
import random
import shutil
import subprocess
import time
import zlib  # for crc32


MAX_RETRY = 1
NUM_WORKERS = 5
SLEEP_BETWEEN_RETRIES = 5
PARALLEL_GROUP_SIZE = 100
CLICKHOUSE_BINARY_PATH = "usr/bin/clickhouse"
CLICKHOUSE_ODBC_BRIDGE_BINARY_PATH = "usr/bin/clickhouse-odbc-bridge"
CLICKHOUSE_LIBRARY_BRIDGE_BINARY_PATH = "usr/bin/clickhouse-library-bridge"

FLAKY_TRIES_COUNT = 10
MAX_TIME_SECONDS = 3600

MAX_TIME_IN_SANDBOX = 20 * 60  # 20 minutes
TASK_TIMEOUT = 8 * 60 * 60  # 8 hours

NO_CHANGES_MSG = "Nothing to run"


def stringhash(s):
    return zlib.crc32(s.encode("utf-8"))


# Search test by the common prefix.
# This is accept tests w/o parameters in skip list.
#
# Examples:
# - has_test(['foobar'], 'foobar[param]') == True
# - has_test(['foobar[param]'], 'foobar') == True
def has_test(tests, test_to_match):
    for test in tests:
        if len(test_to_match) < len(test):
            if test[0 : len(test_to_match)] == test_to_match:
                return True
        else:
            if test_to_match[0 : len(test)] == test:
                return True
    return False


def get_changed_tests_to_run(pr_info, repo_path):
    result = set()
    changed_files = pr_info["changed_files"]

    if changed_files is None:
        return []

    for fpath in changed_files:
        if "tests/integration/test_" in fpath:
            logging.info("File %s changed and seems like integration test", fpath)
            result.add(fpath.split("/")[2])
    return filter_existing_tests(result, repo_path)


def filter_existing_tests(tests_to_run, repo_path):
    result = []
    for relative_test_path in tests_to_run:
        if os.path.exists(
            os.path.join(repo_path, "tests/integration", relative_test_path)
        ):
            result.append(relative_test_path)
        else:
            logging.info(
                "Skipping test %s, seems like it was removed", relative_test_path
            )
    return result


def _get_deselect_option(tests):
    return " ".join([f"--deselect {t}" for t in tests])


# https://stackoverflow.com/questions/312443/how-do-you-split-a-list-into-evenly-sized-chunks
def chunks(lst, n):
    """Yield successive n-sized chunks from lst."""
    for i in range(0, len(lst), n):
        yield lst[i : i + n]


def get_counters(fname):
    counters = {
        "ERROR": set([]),
        "PASSED": set([]),
        "FAILED": set([]),
        "SKIPPED": set([]),
    }

    with open(fname, "r") as out:
        for line in out:
            line = line.strip()
            # Example of log:
            #
            #     test_mysql_protocol/test.py::test_golang_client
            #     [gw0] [  7%] ERROR test_mysql_protocol/test.py::test_golang_client
            #
            # And only the line with test status should be matched
            if not (".py::" in line and " " in line):
                continue

            line_arr = line.strip().split(" ")
            if len(line_arr) < 2:
                logging.debug("Strange line %s", line)
                continue

            # Lines like:
            #     [gw0] [  7%] ERROR test_mysql_protocol/test.py::test_golang_client
            #     [gw3] [ 40%] PASSED test_replicated_users/test.py::test_rename_replicated[QUOTA]
            state = line_arr[-2]
            test_name = line_arr[-1]

            if state in counters:
                counters[state].add(test_name)
            else:
                # will skip lines like:
                #     30.76s call     test_host_ip_change/test.py::test_ip_drop_cache
                #     5.71s teardown  test_host_ip_change/test.py::test_ip_change[node1]
                # and similar
                logging.debug("Strange state in line %s", line)

    return {k: list(v) for k, v in counters.items()}


def parse_test_times(fname):
    read = False
    description_output = []
    with open(fname, "r") as out:
        for line in out:
            if read and "==" in line:
                break
            if read and line.strip():
                description_output.append(line.strip())
            if "slowest durations" in line:
                read = True
    return description_output


def get_test_times(output):
    result = defaultdict(float)
    for line in output:
        if ".py" in line:
            line_arr = line.strip().split(" ")
            test_time = line_arr[0]
            test_name = " ".join([elem for elem in line_arr[2:] if elem])
            if test_name not in result:
                result[test_name] = 0.0
            result[test_name] += float(test_time[:-1])
    return result


def clear_ip_tables_and_restart_daemons():
    logging.info(
        "Dump iptables after run %s",
        subprocess.check_output("sudo iptables -nvL", shell=True),
    )
    try:
        logging.info("Killing all alive docker containers")
        subprocess.check_output(
            "timeout -s 9 10m docker ps --quiet | xargs --no-run-if-empty docker kill",
            shell=True,
        )
    except subprocess.CalledProcessError as err:
        logging.info("docker kill excepted: " + str(err))

    try:
        logging.info("Removing all docker containers")
        subprocess.check_output(
            "timeout -s 9 10m docker ps --all --quiet | xargs --no-run-if-empty docker rm --force",
            shell=True,
        )
    except subprocess.CalledProcessError as err:
        logging.info("docker rm excepted: " + str(err))

    # don't restart docker if it's disabled
    if os.environ.get("CLICKHOUSE_TESTS_RUNNER_RESTART_DOCKER", "1") == "1":
        try:
            logging.info("Stopping docker daemon")
            subprocess.check_output("service docker stop", shell=True)
        except subprocess.CalledProcessError as err:
            logging.info("docker stop excepted: " + str(err))

        try:
            for i in range(200):
                try:
                    logging.info("Restarting docker %s", i)
                    subprocess.check_output("service docker start", shell=True)
                    subprocess.check_output("docker ps", shell=True)
                    break
                except subprocess.CalledProcessError as err:
                    time.sleep(0.5)
                    logging.info("Waiting docker to start, current %s", str(err))
            else:
                raise Exception("Docker daemon doesn't responding")
        except subprocess.CalledProcessError as err:
            logging.info("Can't reload docker: " + str(err))

    iptables_iter = 0
    try:
        for i in range(1000):
            iptables_iter = i
            # when rules will be empty, it will raise exception
            subprocess.check_output("sudo iptables -D DOCKER-USER 1", shell=True)
    except subprocess.CalledProcessError as err:
        logging.info(
            "All iptables rules cleared, %s iterations, last error: %s",
            iptables_iter,
            str(err),
        )


class ClickhouseIntegrationTestsRunner:
    def __init__(self, result_path, params):
        self.result_path = result_path
        self.params = params

        self.image_versions = self.params["docker_images_with_versions"]
        self.shuffle_groups = self.params["shuffle_test_groups"]
        self.flaky_check = "flaky check" in self.params["context_name"]
        self.bugfix_validate_check = (
            "bugfix validate check" in self.params["context_name"]
        )
        # if use_tmpfs is not set we assume it to be true, otherwise check
        self.use_tmpfs = "use_tmpfs" not in self.params or self.params["use_tmpfs"]
        self.disable_net_host = (
            "disable_net_host" in self.params and self.params["disable_net_host"]
        )
        self.start_time = time.time()
        self.soft_deadline_time = self.start_time + (TASK_TIMEOUT - MAX_TIME_IN_SANDBOX)

        if "run_by_hash_total" in self.params:
            self.run_by_hash_total = self.params["run_by_hash_total"]
            self.run_by_hash_num = self.params["run_by_hash_num"]
        else:
            self.run_by_hash_total = 0
            self.run_by_hash_num = 0

    def path(self):
        return self.result_path

    def base_path(self):
        return os.path.join(str(self.result_path), "../")

    def should_skip_tests(self):
        return []

    def get_image_with_version(self, name):
        if name in self.image_versions:
            return name + ":" + self.image_versions[name]
        logging.warn(
            "Cannot find image %s in params list %s", name, self.image_versions
        )
        if ":" not in name:
            return name + ":latest"
        return name

    def get_image_version(self, name: str):
        if name in self.image_versions:
            return self.image_versions[name]
        logging.warn(
            "Cannot find image %s in params list %s", name, self.image_versions
        )
        return "latest"

    def shuffle_test_groups(self):
        return self.shuffle_groups != 0

    @staticmethod
    def get_images_names():
        return [
            "clickhouse/dotnet-client",
            "clickhouse/integration-helper",
            "clickhouse/integration-test",
            "clickhouse/integration-tests-runner",
            "clickhouse/kerberized-hadoop",
            "clickhouse/kerberos-kdc",
            "clickhouse/mysql-golang-client",
            "clickhouse/mysql-java-client",
            "clickhouse/mysql-js-client",
            "clickhouse/mysql-php-client",
            "clickhouse/postgresql-java-client",
        ]

    def _pre_pull_images(self, repo_path):
        image_cmd = self._get_runner_image_cmd(repo_path)

        cmd = (
            "cd {repo_path}/tests/integration && "
            "timeout -s 9 1h ./runner {runner_opts} {image_cmd} --pre-pull --command '{command}' ".format(
                repo_path=repo_path,
                runner_opts=self._get_runner_opts(),
                image_cmd=image_cmd,
                command=r""" echo Pre Pull finished """,
            )
        )

        for i in range(5):
            logging.info("Pulling images before running tests. Attempt %s", i)
            try:
                subprocess.check_output(
                    cmd,
                    shell=True,
                )
                return
            except subprocess.CalledProcessError as err:
                logging.info("docker-compose pull failed: " + str(err))
                continue
        logging.error("Pulling images failed for 5 attempts. Will fail the worker.")
        # We pass specific retcode to to ci/integration_test_check.py to skip status reporting and restart job
        exit(13)

    def _can_run_with(self, path, opt):
        with open(path, "r") as script:
            for line in script:
                if opt in line:
                    return True
        return False

    def _install_clickhouse(self, debs_path):
        for package in (
            "clickhouse-common-static_",
            "clickhouse-server_",
            "clickhouse-client",
            "clickhouse-common-static-dbg_",
        ):  # order matters
            logging.info("Installing package %s", package)
            for f in os.listdir(debs_path):
                if package in f:
                    full_path = os.path.join(debs_path, f)
                    logging.info("Package found in %s", full_path)
                    log_name = "install_" + f + ".log"
                    log_path = os.path.join(str(self.path()), log_name)
                    with open(log_path, "w") as log:
                        cmd = "dpkg -x {} .".format(full_path)
                        logging.info("Executing installation cmd %s", cmd)
                        retcode = subprocess.Popen(
                            cmd, shell=True, stderr=log, stdout=log
                        ).wait()
                        if retcode == 0:
                            logging.info("Installation of %s successfull", full_path)
                        else:
                            raise Exception("Installation of %s failed", full_path)
                    break
            else:
                raise Exception("Package with {} not found".format(package))
        # logging.info("Unstripping binary")
        # logging.info(
        #     "Unstring %s",
        #     subprocess.check_output(
        #         "eu-unstrip /usr/bin/clickhouse {}".format(CLICKHOUSE_BINARY_PATH),
        #         shell=True,
        #     ),
        # )

        logging.info("All packages installed")
        os.chmod(CLICKHOUSE_BINARY_PATH, 0o777)
        os.chmod(CLICKHOUSE_ODBC_BRIDGE_BINARY_PATH, 0o777)
        os.chmod(CLICKHOUSE_LIBRARY_BRIDGE_BINARY_PATH, 0o777)
        shutil.copy(
            CLICKHOUSE_BINARY_PATH, os.getenv("CLICKHOUSE_TESTS_SERVER_BIN_PATH")
        )
        shutil.copy(
            CLICKHOUSE_ODBC_BRIDGE_BINARY_PATH,
            os.getenv("CLICKHOUSE_TESTS_ODBC_BRIDGE_BIN_PATH"),
        )
        shutil.copy(
            CLICKHOUSE_LIBRARY_BRIDGE_BINARY_PATH,
            os.getenv("CLICKHOUSE_TESTS_LIBRARY_BRIDGE_BIN_PATH"),
        )

    def _compress_logs(self, dir, relpaths, result_path):
        retcode = subprocess.call(  # STYLE_CHECK_ALLOW_SUBPROCESS_CHECK_CALL
            "tar --use-compress-program='zstd --threads=0' -cf {} -C {} {}".format(
                result_path, dir, " ".join(relpaths)
            ),
            shell=True,
        )
        # tar return 1 when the files are changed on compressing, we ignore it
        if retcode in (0, 1):
            return
        # but even on the fatal errors it's better to retry
        logging.error("Fatal error on compressing %s: %s", result_path, retcode)

    def _get_runner_opts(self):
        result = []
        if self.use_tmpfs:
            result.append("--tmpfs")
        if self.disable_net_host:
            result.append("--disable-net-host")
        return " ".join(result)

    def _get_all_tests(self, repo_path):
        image_cmd = self._get_runner_image_cmd(repo_path)
        out_file = "all_tests.txt"
        out_file_full = os.path.join(self.result_path, "runner_get_all_tests.log")
        cmd = (
            "cd {repo_path}/tests/integration && "
            "timeout -s 9 1h ./runner {runner_opts} {image_cmd} ' --setup-plan' "
            "| tee {out_file_full} | grep '::' | sed 's/ (fixtures used:.*//g' | sed 's/^ *//g' | sed 's/ *$//g' "
            "| grep -v 'SKIPPED' | sort -u  > {out_file}".format(
                repo_path=repo_path,
                runner_opts=self._get_runner_opts(),
                image_cmd=image_cmd,
                out_file=out_file,
                out_file_full=out_file_full,
            )
        )

        logging.info("Getting all tests with cmd '%s'", cmd)
        subprocess.check_call(  # STYLE_CHECK_ALLOW_SUBPROCESS_CHECK_CALL
            cmd, shell=True
        )

        all_tests_file_path = "{repo_path}/tests/integration/{out_file}".format(
            repo_path=repo_path, out_file=out_file
        )
        if (
            not os.path.isfile(all_tests_file_path)
            or os.path.getsize(all_tests_file_path) == 0
        ):
            if os.path.isfile(out_file_full):
                # log runner output
                logging.info("runner output:")
                with open(out_file_full, "r") as all_tests_full_file:
                    for line in all_tests_full_file:
                        line = line.rstrip()
                        if line:
                            logging.info("runner output: %s", line)
            else:
                logging.info("runner output '%s' is empty", out_file_full)

            raise Exception(
                "There is something wrong with getting all tests list: file '{}' is empty or does not exist.".format(
                    all_tests_file_path
                )
            )

        all_tests = []
        with open(all_tests_file_path, "r") as all_tests_file:
            for line in all_tests_file:
                all_tests.append(line.strip())
        return list(sorted(all_tests))

    def _get_parallel_tests_skip_list(self, repo_path):
        skip_list_file_path = "{}/tests/integration/parallel_skip.json".format(
            repo_path
        )
        if (
            not os.path.isfile(skip_list_file_path)
            or os.path.getsize(skip_list_file_path) == 0
        ):
            raise Exception(
                "There is something wrong with getting all tests list: file '{}' is empty or does not exist.".format(
                    skip_list_file_path
                )
            )

        skip_list_tests = []
        with open(skip_list_file_path, "r") as skip_list_file:
            skip_list_tests = json.load(skip_list_file)
        return list(sorted(skip_list_tests))

    def group_test_by_file(self, tests):
        result = {}
        for test in tests:
            test_file = test.split("::")[0]
            if test_file not in result:
                result[test_file] = []
            result[test_file].append(test)
        return result

    def _update_counters(self, main_counters, current_counters):
        for test in current_counters["PASSED"]:
            if (
                test not in main_counters["PASSED"]
                and test not in main_counters["FLAKY"]
            ):
                is_flaky = False
                if test in main_counters["FAILED"]:
                    main_counters["FAILED"].remove(test)
                    is_flaky = True
                if test in main_counters["ERROR"]:
                    main_counters["ERROR"].remove(test)
                    is_flaky = True
                if is_flaky:
                    main_counters["FLAKY"].append(test)
                else:
                    main_counters["PASSED"].append(test)

        for state in ("ERROR", "FAILED"):
            for test in current_counters[state]:
                if test in main_counters["FLAKY"]:
                    continue
                if test in main_counters["PASSED"]:
                    main_counters["PASSED"].remove(test)
                    main_counters["FLAKY"].append(test)
                    continue
                if test not in main_counters[state]:
                    main_counters[state].append(test)

        for state in ("SKIPPED",):
            for test in current_counters[state]:
                main_counters[state].append(test)

    def _get_runner_image_cmd(self, repo_path):
        image_cmd = ""
        if self._can_run_with(
            os.path.join(repo_path, "tests/integration", "runner"),
            "--docker-image-version",
        ):
            for img in self.get_images_names():
                if img == "clickhouse/integration-tests-runner":
                    runner_version = self.get_image_version(img)
                    logging.info(
                        "Can run with custom docker image version %s", runner_version
                    )
                    image_cmd += " --docker-image-version={} ".format(runner_version)
                else:
                    if self._can_run_with(
                        os.path.join(repo_path, "tests/integration", "runner"),
                        "--docker-compose-images-tags",
                    ):
                        image_cmd += "--docker-compose-images-tags={} ".format(
                            self.get_image_with_version(img)
                        )
        else:
            image_cmd = ""
            logging.info("Cannot run with custom docker image version :(")
        return image_cmd

    def _find_test_data_dirs(self, repo_path, test_names):
        relpaths = {}
        for test_name in test_names:
            if "/" in test_name:
                test_dir = test_name[: test_name.find("/")]
            else:
                test_dir = test_name
            if os.path.isdir(os.path.join(repo_path, "tests/integration", test_dir)):
                for name in os.listdir(
                    os.path.join(repo_path, "tests/integration", test_dir)
                ):
                    relpath = os.path.join(os.path.join(test_dir, name))
                    mtime = os.path.getmtime(
                        os.path.join(repo_path, "tests/integration", relpath)
                    )
                    relpaths[relpath] = mtime
        return relpaths

    def _get_test_data_dirs_difference(self, new_snapshot, old_snapshot):
        res = set()
        for path in new_snapshot:
            if (path not in old_snapshot) or (old_snapshot[path] != new_snapshot[path]):
                res.add(path)
        return res

    def try_run_test_group(
        self, repo_path, test_group, tests_in_group, num_tries, num_workers
    ):
        try:
            return self.run_test_group(
                repo_path, test_group, tests_in_group, num_tries, num_workers
            )
        except Exception as e:
            logging.info("Failed to run {}:\n{}".format(str(test_group), str(e)))
            counters = {
                "ERROR": [],
                "PASSED": [],
                "FAILED": [],
                "SKIPPED": [],
                "FLAKY": [],
            }
            tests_times = defaultdict(float)
            for test in tests_in_group:
                counters["ERROR"].append(test)
                tests_times[test] = 0
            return counters, tests_times, []

    def run_test_group(
        self, repo_path, test_group, tests_in_group, num_tries, num_workers
    ):
        counters = {
            "ERROR": [],
            "PASSED": [],
            "FAILED": [],
            "SKIPPED": [],
            "FLAKY": [],
        }
        tests_times = defaultdict(float)

        if self.soft_deadline_time < time.time():
            for test in tests_in_group:
                logging.info("Task timeout exceeded, skipping %s", test)
                counters["SKIPPED"].append(test)
                tests_times[test] = 0
            return counters, tests_times, []

        image_cmd = self._get_runner_image_cmd(repo_path)
        test_group_str = test_group.replace("/", "_").replace(".", "_")

        log_paths = []
        test_data_dirs = {}

        for i in range(num_tries):
            logging.info("Running test group %s for the %s retry", test_group, i)
            clear_ip_tables_and_restart_daemons()

            test_names = set([])
            for test_name in tests_in_group:
                if test_name not in counters["PASSED"]:
                    test_names.add(test_name)

            if i == 0:
                test_data_dirs = self._find_test_data_dirs(repo_path, test_names)

            info_basename = test_group_str + "_" + str(i) + ".nfo"
            info_path = os.path.join(repo_path, "tests/integration", info_basename)

            test_cmd = " ".join([test for test in sorted(test_names)])
            parallel_cmd = (
                " --parallel {} ".format(num_workers) if num_workers > 0 else ""
            )
            # -r -- show extra test summary:
            # -f -- (f)ailed
            # -E -- (E)rror
            # -p -- (p)assed
            # -s -- (s)kipped
            cmd = "cd {}/tests/integration && timeout -s 9 1h ./runner {} {} -t {} {} '-rfEps --run-id={} --color=no --durations=0 {}' | tee {}".format(
                repo_path,
                self._get_runner_opts(),
                image_cmd,
                test_cmd,
                parallel_cmd,
                i,
                _get_deselect_option(self.should_skip_tests()),
                info_path,
            )

            log_basename = test_group_str + "_" + str(i) + ".log"
            log_path = os.path.join(repo_path, "tests/integration", log_basename)
            with open(log_path, "w") as log:
                logging.info("Executing cmd: %s", cmd)
                # ignore retcode, since it meaningful due to pipe to tee
                subprocess.Popen(cmd, shell=True, stderr=log, stdout=log).wait()

            extra_logs_names = [log_basename]
            log_result_path = os.path.join(
                str(self.path()), "integration_run_" + log_basename
            )
            shutil.copy(log_path, log_result_path)
            log_paths.append(log_result_path)

            for pytest_log_path in glob.glob(
                os.path.join(repo_path, "tests/integration/pytest*.log")
            ):
                new_name = (
                    test_group_str
                    + "_"
                    + str(i)
                    + "_"
                    + os.path.basename(pytest_log_path)
                )
                os.rename(
                    pytest_log_path,
                    os.path.join(repo_path, "tests/integration", new_name),
                )
                extra_logs_names.append(new_name)

            dockerd_log_path = os.path.join(repo_path, "tests/integration/dockerd.log")
            if os.path.exists(dockerd_log_path):
                new_name = (
                    test_group_str
                    + "_"
                    + str(i)
                    + "_"
                    + os.path.basename(dockerd_log_path)
                )
                os.rename(
                    dockerd_log_path,
                    os.path.join(repo_path, "tests/integration", new_name),
                )
                extra_logs_names.append(new_name)

            if os.path.exists(info_path):
                extra_logs_names.append(info_basename)
                new_counters = get_counters(info_path)
                for state, tests in new_counters.items():
                    logging.info(
                        "Tests with %s state (%s): %s", state, len(tests), tests
                    )
                times_lines = parse_test_times(info_path)
                new_tests_times = get_test_times(times_lines)
                self._update_counters(counters, new_counters)
                for test_name, test_time in new_tests_times.items():
                    tests_times[test_name] = test_time

            test_data_dirs_new = self._find_test_data_dirs(repo_path, test_names)
            test_data_dirs_diff = self._get_test_data_dirs_difference(
                test_data_dirs_new, test_data_dirs
            )
            test_data_dirs = test_data_dirs_new

            if extra_logs_names or test_data_dirs_diff:
                extras_result_path = os.path.join(
                    str(self.path()),
                    "integration_run_" + test_group_str + "_" + str(i) + ".tar.zst",
                )
                self._compress_logs(
                    os.path.join(repo_path, "tests/integration"),
                    extra_logs_names + list(test_data_dirs_diff),
                    extras_result_path,
                )
                log_paths.append(extras_result_path)

            if len(counters["PASSED"]) + len(counters["FLAKY"]) == len(tests_in_group):
                logging.info("All tests from group %s passed", test_group)
                break
            if (
                len(counters["PASSED"]) + len(counters["FLAKY"]) >= 0
                and len(counters["FAILED"]) == 0
                and len(counters["ERROR"]) == 0
            ):
                logging.info(
                    "Seems like all tests passed but some of them are skipped or "
                    "deselected. Ignoring them and finishing group."
                )
                break
        else:
            # Mark all non tried tests as errors, with '::' in name
            # (example test_partition/test.py::test_partition_simple). For flaky check
            # we run whole test dirs like "test_odbc_interaction" and don't
            # want to mark them as error so we filter by '::'.
            for test in tests_in_group:
                if (
                    test not in counters["PASSED"]
                    and test not in counters["ERROR"]
                    and test not in counters["SKIPPED"]
                    and test not in counters["FAILED"]
                    and "::" in test
                ):
                    counters["ERROR"].append(test)

        return counters, tests_times, log_paths

    def run_flaky_check(self, repo_path, build_path, should_fail=False):
        pr_info = self.params["pr_info"]

        tests_to_run = get_changed_tests_to_run(pr_info, repo_path)
        if not tests_to_run:
            logging.info("No integration tests to run found")
            return "success", NO_CHANGES_MSG, [(NO_CHANGES_MSG, "OK")], ""

        self._install_clickhouse(build_path)
        logging.info("Found '%s' tests to run", " ".join(tests_to_run))
        result_state = "success"
        description_prefix = "No flaky tests: "
        start = time.time()
        logging.info("Starting check with retries")
        final_retry = 0
        logs = []
        tires_num = 1 if should_fail else FLAKY_TRIES_COUNT
        for i in range(tires_num):
            final_retry += 1
            logging.info("Running tests for the %s time", i)
            counters, tests_times, log_paths = self.try_run_test_group(
                repo_path, "bugfix" if should_fail else "flaky", tests_to_run, 1, 1
            )
            logs += log_paths
            if counters["FAILED"]:
                logging.info("Found failed tests: %s", " ".join(counters["FAILED"]))
                description_prefix = "Failed tests found: "
                result_state = "failure"
                if not should_fail:
                    break
            if counters["ERROR"]:
                description_prefix = "Failed tests found: "
                logging.info("Found error tests: %s", " ".join(counters["ERROR"]))
                # NOTE "error" result state will restart the whole test task,
                # so we use "failure" here
                result_state = "failure"
                if not should_fail:
                    break
            assert len(counters["FLAKY"]) == 0 or should_fail
            logging.info("Try is OK, all tests passed, going to clear env")
            clear_ip_tables_and_restart_daemons()
            logging.info("And going to sleep for some time")
            if time.time() - start > MAX_TIME_SECONDS:
                logging.info("Timeout reached, going to finish flaky check")
                break
            time.sleep(5)

        test_result = []
        for state in ("ERROR", "FAILED", "PASSED", "SKIPPED", "FLAKY"):
            if state == "PASSED":
                text_state = "OK"
            elif state == "FAILED":
                text_state = "FAIL"
            else:
                text_state = state
            test_result += [
                (
                    c + " (✕" + str(final_retry) + ")",
                    text_state,
                    "{:.2f}".format(tests_times[c]),
                )
                for c in counters[state]
            ]
        status_text = description_prefix + ", ".join(
            [
                str(n).lower().replace("failed", "fail") + ": " + str(len(c))
                for n, c in counters.items()
            ]
        )

        return result_state, status_text, test_result, logs

    def run_impl(self, repo_path, build_path):
        if self.flaky_check or self.bugfix_validate_check:
            return self.run_flaky_check(
                repo_path, build_path, should_fail=self.bugfix_validate_check
            )

        self._install_clickhouse(build_path)

        logging.info("Pulling images")
        runner._pre_pull_images(repo_path)

        logging.info(
            "Dump iptables before run %s",
            subprocess.check_output("sudo iptables -nvL", shell=True),
        )
        all_tests = self._get_all_tests(repo_path)

        if self.run_by_hash_total != 0:
            grouped_tests = self.group_test_by_file(all_tests)
            all_filtered_by_hash_tests = []
            for group, tests_in_group in grouped_tests.items():
                if stringhash(group) % self.run_by_hash_total == self.run_by_hash_num:
                    all_filtered_by_hash_tests += tests_in_group
            all_tests = all_filtered_by_hash_tests

        parallel_skip_tests = self._get_parallel_tests_skip_list(repo_path)
        logging.info(
            "Found %s tests first 3 %s", len(all_tests), " ".join(all_tests[:3])
        )
        filtered_sequential_tests = list(
            filter(lambda test: has_test(all_tests, test), parallel_skip_tests)
        )
        filtered_parallel_tests = list(
            filter(
                lambda test: not has_test(parallel_skip_tests, test),
                all_tests,
            )
        )
        not_found_tests = list(
            filter(
                lambda test: not has_test(all_tests, test),
                parallel_skip_tests,
            )
        )
        logging.info(
            "Found %s tests first 3 %s, parallel %s, other %s",
            len(all_tests),
            " ".join(all_tests[:3]),
            len(filtered_parallel_tests),
            len(filtered_sequential_tests),
        )
        logging.info(
            "Not found %s tests first 3 %s",
            len(not_found_tests),
            " ".join(not_found_tests[:3]),
        )

        grouped_tests = self.group_test_by_file(filtered_sequential_tests)
        i = 0
        for par_group in chunks(filtered_parallel_tests, PARALLEL_GROUP_SIZE):
            grouped_tests[f"parallel{i}"] = par_group
            i += 1
        logging.info("Found %s tests groups", len(grouped_tests))

        counters = {
            "ERROR": [],
            "PASSED": [],
            "FAILED": [],
            "SKIPPED": [],
            "FLAKY": [],
        }
        tests_times = defaultdict(float)
        tests_log_paths = defaultdict(list)

        items_to_run = list(grouped_tests.items())

        logging.info("Total test groups %s", len(items_to_run))
        if self.shuffle_test_groups():
            logging.info("Shuffling test groups")
            random.shuffle(items_to_run)

        for group, tests in items_to_run:
            logging.info("Running test group %s containing %s tests", group, len(tests))
            group_counters, group_test_times, log_paths = self.try_run_test_group(
                repo_path, group, tests, MAX_RETRY, NUM_WORKERS
            )
            total_tests = 0
            for counter, value in group_counters.items():
                logging.info(
                    "Tests from group %s stats, %s count %s", group, counter, len(value)
                )
                counters[counter] += value
                logging.info(
                    "Totally have %s with status %s", len(counters[counter]), counter
                )
                total_tests += len(counters[counter])
            logging.info("Totally finished tests %s/%s", total_tests, len(all_tests))

            for test_name, test_time in group_test_times.items():
                tests_times[test_name] = test_time
                tests_log_paths[test_name] = log_paths

            if len(counters["FAILED"]) + len(counters["ERROR"]) >= 20:
                logging.info("Collected more than 20 failed/error tests, stopping")
                break

        if counters["FAILED"] or counters["ERROR"]:
            logging.info(
                "Overall status failure, because we have tests in FAILED or ERROR state"
            )
            result_state = "failure"
        else:
            logging.info("Overall success!")
            result_state = "success"

        test_result = []
        for state in ("ERROR", "FAILED", "PASSED", "SKIPPED", "FLAKY"):
            if state == "PASSED":
                text_state = "OK"
            elif state == "FAILED":
                text_state = "FAIL"
            else:
                text_state = state
            test_result += [
                (c, text_state, "{:.2f}".format(tests_times[c]), tests_log_paths[c])
                for c in counters[state]
            ]

        failed_sum = len(counters["FAILED"]) + len(counters["ERROR"])
        status_text = "fail: {}, passed: {}, flaky: {}".format(
            failed_sum, len(counters["PASSED"]), len(counters["FLAKY"])
        )

        if self.soft_deadline_time < time.time():
            status_text = "Timeout, " + status_text
            result_state = "failure"

        counters["FLAKY"] = []
        if not counters or sum(len(counter) for counter in counters.values()) == 0:
            status_text = "No tests found for some reason! It's a bug"
            result_state = "failure"

        if "(memory)" in self.params["context_name"]:
            result_state = "success"

        for res in test_result:
            # It's not easy to parse output of pytest
            # Especially when test names may contain spaces
            # Do not allow it to avoid obscure failures
            if " " not in res[0]:
                continue
            logging.warning("Found invalid test name with space: %s", res[0])
            status_text = "Found test with invalid name, see main log"
            result_state = "failure"

        return result_state, status_text, test_result, []


def write_results(results_file, status_file, results, status):
    with open(results_file, "w") as f:
        out = csv.writer(f, delimiter="\t")
        out.writerows(results)
    with open(status_file, "w") as f:
        out = csv.writer(f, delimiter="\t")
        out.writerow(status)


if __name__ == "__main__":
    logging.basicConfig(level=logging.INFO, format="%(asctime)s %(message)s")

    repo_path = os.environ.get("CLICKHOUSE_TESTS_REPO_PATH")
    build_path = os.environ.get("CLICKHOUSE_TESTS_BUILD_PATH")
    result_path = os.environ.get("CLICKHOUSE_TESTS_RESULT_PATH")
    params_path = os.environ.get("CLICKHOUSE_TESTS_JSON_PARAMS_PATH")

    params = json.loads(open(params_path, "r").read())
    runner = ClickhouseIntegrationTestsRunner(result_path, params)

    logging.info("Running tests")

    # Avoid overlaps with previous runs
    logging.info("Clearing dmesg before run")
    subprocess.check_call(  # STYLE_CHECK_ALLOW_SUBPROCESS_CHECK_CALL
        "dmesg --clear", shell=True
    )

    state, description, test_results, _ = runner.run_impl(repo_path, build_path)
    logging.info("Tests finished")

    # Dump dmesg (to capture possible OOMs)
    logging.info("Dumping dmesg")
    subprocess.check_call(  # STYLE_CHECK_ALLOW_SUBPROCESS_CHECK_CALL
        "dmesg -T", shell=True
    )

    status = (state, description)
    out_results_file = os.path.join(str(runner.path()), "test_results.tsv")
    out_status_file = os.path.join(str(runner.path()), "check_status.tsv")
    write_results(out_results_file, out_status_file, test_results, status)
    logging.info("Result written")