#!/usr/bin/env python3 from typing import List, Tuple import subprocess import logging import json import os import sys import time import urllib.parse import requests # type: ignore from ci_config import CI_CONFIG, BuildConfig from docker_pull_helper import get_image_with_version from env_helper import ( GITHUB_JOB, IMAGES_PATH, REPO_COPY, S3_BUILDS_BUCKET, S3_DOWNLOAD, TEMP_PATH, ) from pr_info import PRInfo from s3_helper import S3Helper from tee_popen import TeePopen from version_helper import ( ClickHouseVersion, Git, get_version_from_repo, update_version_local, ) from clickhouse_helper import ( ClickHouseHelper, prepare_tests_results_for_clickhouse, get_instance_type, ) from stopwatch import Stopwatch IMAGE_NAME = "clickhouse/binary-builder" BUILD_LOG_NAME = "build_log.log" def _can_export_binaries(build_config: BuildConfig) -> bool: if build_config.package_type != "deb": return False if build_config.sanitizer != "": return True if build_config.debug_build: return True return False def get_packager_cmd( build_config: BuildConfig, packager_path: str, output_path: str, profile_path: str, build_version: str, image_version: str, official: bool, ) -> 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 --output-dir={output_path} --profile-dir={profile_path}" f"--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" cmd += " --cache=sccache" cmd += " --s3-rw-access" cmd += f" --s3-bucket={S3_BUILDS_BUCKET}" if build_config.additional_pkgs: cmd += " --additional-pkgs" cmd += f" --docker-image-version={image_version}" cmd += f" --version={build_version}" if _can_export_binaries(build_config): cmd += " --with-binaries=tests" if official: cmd += " --official" return cmd def build_clickhouse( packager_cmd: str, logs_path: str, build_output_path: str ) -> Tuple[str, bool]: build_log_path = os.path.join(logs_path, BUILD_LOG_NAME) success = False with TeePopen(packager_cmd, build_log_path) as process: retcode = process.wait() if os.path.exists(build_output_path): build_results = os.listdir(build_output_path) else: build_results = [] if retcode == 0: if len(build_results) > 0: success = True logging.info("Built successfully") else: logging.info( "Success exit code, but no build artifacts => build failed" ) else: logging.info("Build failed") return build_log_path, success def check_for_success_run( s3_helper: S3Helper, s3_prefix: str, build_name: str, build_config: BuildConfig, ) -> None: # 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) try: # Performance artifacts are now part of regular build, so we're safe build_results = s3_helper.list_prefix(s3_prefix) 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: log_url = f"{S3_DOWNLOAD}/{S3_BUILDS_BUCKET}/{url_escaped}" else: 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 create_json_artifact( TEMP_PATH, build_name, log_url, build_urls, build_config, 0, success, ) # Fail build job if not successeded if not success: sys.exit(1) else: sys.exit(0) def create_json_artifact( temp_path: str, build_name: str, log_url: str, build_urls: List[str], build_config: BuildConfig, elapsed: int, success: bool, ) -> None: subprocess.check_call( f"echo 'BUILD_URLS=build_urls_{build_name}' >> $GITHUB_ENV", shell=True ) result = { "log_url": log_url, "build_urls": build_urls, "build_config": build_config.__dict__, "elapsed_seconds": elapsed, "status": success, "job_name": GITHUB_JOB, } json_name = "build_urls_" + build_name + ".json" print(f"Dump json report {result} to {json_name} with env build_urls_{build_name}") with open(os.path.join(temp_path, json_name), "w", encoding="utf-8") as build_links: json.dump(result, build_links) 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" 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 # PR number for anything else pr_number = f"PRs/{pr_info.number}" return pr_number, pr_number def upload_master_static_binaries( pr_info: PRInfo, build_config: BuildConfig, s3_helper: S3Helper, build_output_path: str, ) -> None: """Upload binary artifacts to a static S3 links""" static_binary_name = build_config.static_binary_name if pr_info.number != 0: return elif not static_binary_name: return elif pr_info.base_ref != "master": return s3_path = "/".join((pr_info.base_ref, static_binary_name, "clickhouse")) binary = os.path.join(build_output_path, "clickhouse") url = s3_helper.upload_build_file_to_s3(binary, s3_path) print(f"::notice ::Binary static URL: {url}") def main(): logging.basicConfig(level=logging.INFO) stopwatch = Stopwatch() build_name = sys.argv[1] build_config = CI_CONFIG.build_config[build_name] if not os.path.exists(TEMP_PATH): os.makedirs(TEMP_PATH) pr_info = PRInfo() logging.info("Repo copy path %s", REPO_COPY) s3_helper = S3Helper() version = get_version_from_repo(git=Git(True)) release_or_pr, performance_pr = get_release_or_pr(pr_info, version) s3_path_prefix = "/".join((release_or_pr, pr_info.sha, build_name)) # FIXME performance s3_performance_path = "/".join( (performance_pr, pr_info.sha, build_name, "performance.tar.zst") ) # 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, build_config) docker_image = get_image_with_version(IMAGES_PATH, IMAGE_NAME) image_version = docker_image.version logging.info("Got version from repo %s", version.string) official_flag = pr_info.number == 0 version_type = "testing" if "release" in pr_info.labels or "release-lts" in pr_info.labels: version_type = "stable" official_flag = True update_version_local(version, version_type) logging.info("Updated local files with version") logging.info("Build short name %s", build_name) build_output_path = os.path.join(TEMP_PATH, build_name) if not os.path.exists(build_output_path): os.makedirs(build_output_path) build_profile_path = os.path.join(TEMP_PATH, f"{build_name}_profile") if not os.path.exists(build_profile_path): os.makedirs(build_profile_path) packager_cmd = get_packager_cmd( build_config, os.path.join(REPO_COPY, "docker/packager"), build_output_path, build_profile_path, version.string, image_version, official_flag, ) logging.info("Going to run packager with %s", packager_cmd) logs_path = os.path.join(TEMP_PATH, "build_log") if not os.path.exists(logs_path): os.makedirs(logs_path) start = time.time() log_path, success = build_clickhouse(packager_cmd, logs_path, build_output_path) elapsed = int(time.time() - start) subprocess.check_call( f"sudo chown -R ubuntu:ubuntu {build_output_path}", shell=True ) logging.info("Build finished with %s, log path %s", success, log_path) if not success: # 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) # FIXME performance performance_urls = [] performance_path = os.path.join(build_output_path, "performance.tar.zst") if os.path.exists(performance_path): performance_urls.append( s3_helper.upload_build_file_to_s3(performance_path, s3_performance_path) ) logging.info( "Uploaded performance.tar.zst to %s, now delete to avoid duplication", performance_urls[0], ) os.remove(performance_path) build_urls = ( s3_helper.upload_build_folder_to_s3( build_output_path, s3_path_prefix, keep_dirs_in_s3_path=False, upload_symlinks=False, ) + performance_urls ) logging.info("Got build URLs %s", build_urls) print("::notice ::Build URLs: {}".format("\n".join(build_urls))) if os.path.exists(log_path): log_url = s3_helper.upload_build_file_to_s3( log_path, s3_path_prefix + "/" + os.path.basename(log_path) ) logging.info("Log url %s", log_url) else: logging.info("Build log doesn't exist") print(f"::notice ::Log URL: {log_url}") create_json_artifact( TEMP_PATH, build_name, log_url, build_urls, build_config, elapsed, success ) upload_master_static_binaries(pr_info, build_config, s3_helper, build_output_path) # Upload profile data instance_type = get_instance_type() query = urllib.parse.quote( f""" INSERT INTO build_time_trace ( pull_request_number, commit_sha, check_start_time, check_name, instance_type, 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}', * 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 JSONEachRow """ ) clickhouse_ci_logs_host = os.getenv("CLICKHOUSE_CI_LOGS_HOST") clickhouse_ci_logs_password = os.getenv("CLICKHOUSE_CI_LOGS_PASSWORD") url = f"https://ci:{clickhouse_ci_logs_password}@{clickhouse_ci_logs_host}/?query={query}" file_path = os.path.join(build_profile_path, "profile.json") print(f"::notice ::Log Uploading profile data, path: {file_path}, query: {query}") with open(file_path, "rb") as file: requests.post(url, data=file) # Upload statistics to CI database ch_helper = ClickHouseHelper() prepared_events = prepare_tests_results_for_clickhouse( pr_info, [], "success" if success else "failure", 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 not success: sys.exit(1) if __name__ == "__main__": main()