"""在 47 服务器创建空数据副本并进行受限验收，不停止线上服务或删除文件。""" import argparse from contextlib import closing import hashlib import json import os from pathlib import Path import re import secrets import shutil import sqlite3 import subprocess import time from database_inventory import sqlite_inventory def run(*args): return subprocess.check_output(args, text=True).strip() def service_state(): return {name: run('systemctl', 'show', name, '--property=MainPID,NRestarts,ActiveState,InvocationID') for name in ('smail.service', 'nginx.service', 'mysql.service', 'Scert.service', 'fdm-studio.service', 'postgresql@18-main.service')} def database_integrity(root): # 验收器不能创建数据库，也不能允许第二个状态库。 data = root / 'data' assert not (data / 'smail-state.db').exists(), '单数据库实例生成了旧状态库' files = sqlite_inventory(root, [data / 'smail.db']) assert files == [data / 'smail.db'] assert (data / 'smail.db').is_file() and not (data / 'smail.db').is_symlink() with closing(sqlite3.connect((data / 'smail.db').as_uri() + '?mode=ro', uri=True)) as database: integrity = database.execute('PRAGMA integrity_check').fetchall() assert database.execute("SELECT COUNT(*) FROM sqlite_master WHERE type='table' AND name='config_records'").fetchone()[0] == 1 assert integrity == [('ok',)] return {'smail.db': integrity} def stop_test_units(names): assert all(re.fullmatch(r'smail-runtime-(review|probe|nginx|netns)-[0-9]{8}-[0-9]{6}', name) for name in names), '只能停止本轮隔离测试单元' errors = [] for name in names: try: stopped = subprocess.run(['systemctl', 'stop', name], capture_output=True, text=True, timeout=25) if stopped.returncode and 'not loaded' not in stopped.stderr: raise subprocess.CalledProcessError(stopped.returncode, 'systemctl') except (OSError, subprocess.SubprocessError) as error: errors.append({'unit': name, 'action': 'stop', 'error': type(error).__name__}) try: subprocess.run(['systemctl', 'kill', '--kill-whom=all', '--signal=SIGKILL', name], check=True, capture_output=True, timeout=15) except (OSError, subprocess.SubprocessError) as kill_error: errors.append({'unit': name, 'action': 'kill', 'error': type(kill_error).__name__}) return errors def collect_and_stop(result, unit, probe_unit, proxy_unit=None): result['resource_errors'] = [] try: for key, name in [('resources', unit)] + ([('nginx_resources', proxy_unit)] if proxy_unit else []): try: result[key] = subprocess.check_output([ 'systemctl', 'show', name, '--property=MemoryPeak,MemoryCurrent,MemorySwapPeak,Result,ExecMainStatus,NRestarts', ], text=True, timeout=15).strip() except (OSError, subprocess.SubprocessError) as error: result['resource_errors'].append({'unit': name, 'error': type(error).__name__}) finally: names = ([proxy_unit] if proxy_unit else []) + [probe_unit, unit] result['cleanup_errors'] = stop_test_units(names) def migration_kind(source, target, enabled, interrupted): if target is None: return source if not enabled or interrupted or target not in ('sqlite', 'mysql', 'postgres'): raise ValueError('跨引擎目标必须用于独立 SQL 迁移验收') return target def stop_database_bridges(bridges, result): errors = result.setdefault('cleanup_errors', []) closed = [] for port, bridge in reversed(list(bridges.items())): try: bridge.close() closed.append(port) except Exception as error: errors.append({'action': 'database_bridge', 'port': port, 'error': type(error).__name__}) result['database_bridge_ports_stopped'] = sorted(closed) result['database_bridge_stopped'] = bool(bridges) and len(closed) == len(bridges) def validate_memory_profile(hard_mib, soft_mib): if hard_mib not in (128, 192, 256) or soft_mib not in (64, 128, 192): raise ValueError('不支持的隔离内存配置') if soft_mib + 64 > hard_mib: raise ValueError('Go 内存目标与进程硬上限之间必须保留至少 64 MiB') def go_environment(raw): # 只保留本轮资源控制项，不能把其他环境变量写入证据。 result = {} for item in raw.split(b'\0'): name, separator, value = item.partition(b'=') if separator and name in (b'GOMEMLIMIT', b'GOMAXPROCS'): if name.decode('ascii') in result: raise ValueError('Go 资源环境项重复') result[name.decode('ascii')] = value.decode('ascii') return result def main(): parser = argparse.ArgumentParser() parser.add_argument('--main-sha', required=True) parser.add_argument('--binary', type=Path, default=Path('/opt/smail/smail')) parser.add_argument('--nginx-template', type=Path) parser.add_argument('--memory-max-mib', type=int, choices=(128, 192, 256), default=128) parser.add_argument('--go-memory-limit-mib', type=int, choices=(64, 128, 192), default=64) parser.add_argument('--restore', action='store_true') parser.add_argument('--migration', action='store_true') parser.add_argument('--interruption', action='store_true') parser.add_argument('--switch-fault', choices=('before-root', 'after-root')) parser.add_argument('--database-type', choices=('sqlite', 'mysql', 'postgres'), default='sqlite') parser.add_argument('--migration-target-type', choices=('sqlite', 'mysql', 'postgres')) args = parser.parse_args() try: validate_memory_profile(args.memory_max_mib, args.go_memory_limit_mib) target_kind = migration_kind(args.database_type, args.migration_target_type, args.migration, args.interruption) except ValueError as error: parser.error(str(error)) if args.interruption and not args.migration: parser.error('中断验收必须选择同引擎迁移独立场景') if args.switch_fault and (not args.migration or args.interruption or args.database_type == 'sqlite' or target_kind != args.database_type): parser.error('联合切换故障必须使用独立的同引擎 SQL 迁移场景') if args.migration and (args.nginx_template or args.restore): parser.error('迁移进程验收仅使用直接 TLS 的独立场景') if args.restore and args.database_type != 'sqlite': parser.error('SQL 数据库备份恢复需要独立验收，不能套用 SQLite 冷备流程') assert os.geteuid() == 0 and run('hostname') == 'ikun2' os.umask(0o027) live = Path('/opt/smail') binary = args.binary.resolve(strict=True) assert binary == args.binary and binary.is_file() if binary != live / 'smail': assert binary.parent.parent == Path('/var/tmp') and binary.parent.name.startswith('smail-review-') assert binary.stat().st_uid == 0 and binary.stat().st_mode & 0o022 == 0 assert binary.parent.stat().st_uid == 0 and binary.parent.stat().st_mode & 0o022 == 0 with binary.open('rb') as stream: assert hashlib.file_digest(stream, 'sha256').hexdigest() == args.main_sha stream.seek(0) header = stream.read(20) assert header[:5] == b'\x7fELF\x02' and header[18:20] == b'\x3e\x00' with (live / 'smail').open('rb') as stream: live_sha = hashlib.file_digest(stream, 'sha256').hexdigest() memory = dict(line.split(':', 1) for line in Path('/proc/meminfo').read_text().splitlines()) required_memory = args.memory_max_mib + (192 if args.nginx_template else 144) assert int(memory['MemAvailable'].split()[0]) >= required_memory * 1024, '可用内存不足，暂停隔离验收' stamp = time.strftime('%Y%m%d-%H%M%S') target = Path('/var/tmp') / ('smail-review-runtime-' + stamp) target.mkdir(mode=0o750) shutil.chown(target, user='smail', group='smail') root = target / 'instance' control = target / 'control' for directory in (root, control, root / 'ssl'): directory.mkdir(mode=0o750) shutil.chown(directory, user='smail', group='smail') shutil.copy2(binary, root / 'smail') (root / 'smail').chmod(0o755) shutil.copy2(Path(__file__).with_name('database_inventory.py'), target / 'database_inventory.py') (target / 'database_inventory.py').chmod(0o640) shutil.chown(target / 'database_inventory.py', user='root', group='smail') shutil.copy2(Path(__file__).with_name('runtime_probe.py'), target / 'runtime_probe.py') (target / 'runtime_probe.py').chmod(0o640) shutil.chown(target / 'runtime_probe.py', user='root', group='smail') if args.migration: shutil.copy2(Path(__file__).with_name('runtime_migration.py'), target / 'runtime_migration.py') (target / 'runtime_migration.py').chmod(0o640) shutil.chown(target / 'runtime_migration.py', user='root', group='smail') mixed = args.migration and args.database_type != target_kind and 'sqlite' in (args.database_type, target_kind) if args.database_type != 'sqlite' or mixed: shutil.copy2(Path(__file__).with_name('runtime_sql.py'), target / 'runtime_sql.py') (target / 'runtime_sql.py').chmod(0o640) shutil.chown(target / 'runtime_sql.py', user='root', group='smail') if args.migration: shutil.copy2(Path(__file__).with_name('runtime_sql_migration.py'), target / 'runtime_sql_migration.py') (target / 'runtime_sql_migration.py').chmod(0o640) shutil.chown(target / 'runtime_sql_migration.py', user='root', group='smail') if mixed: shutil.copy2(Path(__file__).with_name('runtime_mixed_migration.py'), target / 'runtime_mixed_migration.py') (target / 'runtime_mixed_migration.py').chmod(0o640) shutil.chown(target / 'runtime_mixed_migration.py', user='root', group='smail') if args.interruption or args.switch_fault: shutil.copy2(Path(__file__).with_name('runtime_interruption.py'), target / 'runtime_interruption.py') (target / 'runtime_interruption.py').chmod(0o640) shutil.chown(target / 'runtime_interruption.py', user='root', group='smail') if args.database_type != 'sqlite': shutil.copy2(Path(__file__).with_name('runtime_sql_interruption.py'), target / 'runtime_sql_interruption.py') (target / 'runtime_sql_interruption.py').chmod(0o640) shutil.chown(target / 'runtime_sql_interruption.py', user='root', group='smail') if args.switch_fault: shutil.copy2(Path(__file__).with_name('runtime_sql_switch_fault.py'), target / 'runtime_sql_switch_fault.py') (target / 'runtime_sql_switch_fault.py').chmod(0o640) shutil.chown(target / 'runtime_sql_switch_fault.py', user='root', group='smail') if args.nginx_template: shutil.copy2(Path(__file__).with_name('runtime_proxy.py'), target / 'runtime_proxy.py') (target / 'runtime_proxy.py').chmod(0o640) shutil.chown(target / 'runtime_proxy.py', user='root', group='smail') subprocess.run(['openssl', 'req', '-x509', '-newkey', 'rsa:2048', '-nodes', '-days', '2', '-keyout', str(root / 'ssl/runtime.key'), '-out', str(root / 'ssl/runtime.crt'), '-subj', '/CN=mail.runtime.invalid', '-addext', 'subjectAltName=DNS:runtime.invalid,DNS:*.runtime.invalid'], check=True, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL) for name in ('runtime.key', 'runtime.crt'): path = root / 'ssl' / name path.chmod(0o600) shutil.chown(path, user='smail', group='smail') assert not (root / 'data').exists() and not (root / 'config.json').exists() before = service_state() unit = 'smail-runtime-review-' + stamp probe_unit = 'smail-runtime-probe-' + stamp proxy_unit = 'smail-runtime-nginx-' + stamp result = {'target': str(target), 'unit': unit, 'main_sha256': args.main_sha, 'memory_max_mib': args.memory_max_mib, 'go_memory_limit_mib': args.go_memory_limit_mib, 'cpu_quota_percent': 25, 'source': str(binary), 'live_sha256': live_sha, 'before': before, 'topology': 'nginx-8001-8002' if args.nginx_template else 'direct-80-443'} service_env = [] probe_args = [] sql_fixture = None migration_target = None network_unit = None bridges = {} result['database_type'] = args.database_type if args.database_type != 'sqlite': from runtime_sql import SQLFixture sql_fixture = SQLFixture(args.database_type) result['sql_database'] = sql_fixture.settings['name'] if args.migration and target_kind != 'sqlite': from runtime_sql_migration import MigrationTarget migration_target = MigrationTarget(target_kind) result['sql_target_database'] = migration_target.settings['name'] if args.migration: result['migration_target_type'] = target_kind mixed_source = mixed_target = None if mixed: from runtime_mixed_migration import SQLiteDatabase mixed_source = sql_fixture or SQLiteDatabase(root, {'kind': 'sqlite', 'name': 'smail.db'}) mixed_target = migration_target or SQLiteDatabase(root, {'kind': 'sqlite', 'name': 'migrated.db'}) if args.nginx_template: from runtime_proxy import prepare_nginx prepare_nginx(target, root, args.nginx_template) result['nginx_template_sha256'] = hashlib.sha256(args.nginx_template.read_bytes()).hexdigest() service_env = ['--setenv=setup_port=8001', '--setenv=SMAIL_HTTP_PORT=8001', '--setenv=SMAIL_HTTPS_PORT=8002'] probe_args = ['--proxy'] started = time.strftime('%Y-%m-%dT%H:%M:%S') password = secrets.token_urlsafe(9) try: if sql_fixture: sql_fixture.create() probe_args.append('--sql-stdin') network_args = ['--property=PrivateNetwork=yes'] if migration_target: migration_target.create() if args.migration and (sql_fixture or migration_target): # 使用独立限时单元固定隔离网络，应用换进程后仍能访问本轮数据库桥接。 network_unit = 'smail-runtime-netns-' + stamp subprocess.run([ 'systemd-run', '--unit=' + network_unit, '--property=Type=exec', '--property=User=smail', '--property=Group=smail', '--property=PrivateNetwork=yes', '--property=ProtectSystem=strict', '--property=ProtectHome=yes', '--property=NoNewPrivileges=yes', '--property=InaccessiblePaths=/opt/smail /run/smail', '--property=MemoryMax=8M', '--property=MemorySwapMax=0', '--property=TasksMax=8', '--property=RuntimeMaxSec=900s', '/usr/bin/sleep', '900', ], check=True) network_pid = run('systemctl', 'show', network_unit, '--property=MainPID', '--value') assert network_pid.isdigit() and int(network_pid) > 0 assert Path('/proc/' + network_pid + '/ns/net').readlink() != Path('/proc/self/ns/net').readlink() network_args = ['--property=NetworkNamespacePath=/proc/' + network_pid + '/ns/net'] subprocess.run([ 'systemd-run', '--unit=' + unit, '--property=Type=exec', '--property=User=smail', '--property=Group=smail', *network_args, '--property=PrivateTmp=yes', '--property=ProtectSystem=strict', '--property=ProtectHome=yes', '--property=NoNewPrivileges=yes', '--property=CapabilityBoundingSet=CAP_NET_BIND_SERVICE', '--property=AmbientCapabilities=CAP_NET_BIND_SERVICE', '--property=MemoryMax=' + str(args.memory_max_mib) + 'M', '--property=MemorySwapMax=0', '--property=CPUQuota=25%', '--property=TasksMax=128', '--property=RuntimeMaxSec=' + ('600s' if args.migration else '300s'), '--property=OOMPolicy=stop', '--property=KillMode=control-group', '--property=TimeoutStopSec=10s', '--property=BindPaths=' + str(root) + ':/opt/smail ' + str(control) + ':/run/smail', '--property=ReadWritePaths=/opt/smail /run/smail', '--property=WorkingDirectory=/opt/smail', '--setenv=SMAIL_ROOT=/opt/smail', '--setenv=GOMEMLIMIT=' + str(args.go_memory_limit_mib) + 'MiB', '--setenv=GOMAXPROCS=1', *service_env, '/opt/smail/smail', ], check=True) pid = run('systemctl', 'show', unit, '--property=MainPID', '--value') assert pid != '0' live_net = str(Path('/proc/self/ns/net').readlink()) assert str(Path('/proc/' + pid + '/ns/net').readlink()) != live_net assert Path('/proc/' + pid + '/root/opt/smail/smail').stat().st_ino == (root / 'smail').stat().st_ino assert (root / 'smail').stat().st_ino != (live / 'smail').stat().st_ino result['go_environment'] = go_environment(Path('/proc/' + pid + '/environ').read_bytes()) assert result['go_environment'] == {'GOMEMLIMIT': str(args.go_memory_limit_mib) + 'MiB', 'GOMAXPROCS': '1'} if sql_fixture or migration_target: from runtime_sql_bridge import DatabaseBridge ports = {fixture.port for fixture in (sql_fixture, migration_target) if fixture} for port in sorted(ports): bridges[port] = DatabaseBridge(int(pid), port) result['database_bridge_ports'] = sorted(bridges) if args.nginx_template: subprocess.run([ 'systemd-run', '--unit=' + proxy_unit, '--property=Type=exec', '--property=User=smail', '--property=Group=smail', '--property=NetworkNamespacePath=/proc/' + pid + '/ns/net', '--property=PrivateTmp=no', '--property=ProtectSystem=strict', '--property=ProtectHome=yes', '--property=NoNewPrivileges=yes', '--property=InaccessiblePaths=/opt/smail /run/smail /etc/nginx', '--property=ReadWritePaths=' + str(target / 'nginx'), '--property=CapabilityBoundingSet=CAP_NET_BIND_SERVICE', '--property=AmbientCapabilities=CAP_NET_BIND_SERVICE', '--property=MemoryMax=32M', '--property=MemorySwapMax=0', '--property=CPUQuota=25%', '--property=TasksMax=16', '--property=RuntimeMaxSec=300s', '--property=OOMPolicy=stop', '--property=KillMode=control-group', '--property=TimeoutStopSec=10s', '/usr/sbin/nginx', '-p', str(target / 'nginx'), '-c', str(target / 'nginx/nginx.conf'), '-g', 'daemon off;', ], check=True) probe = subprocess.run([ 'systemd-run', '--wait', '--pipe', '--unit=' + probe_unit, '--property=User=smail', '--property=Group=smail', '--property=NetworkNamespacePath=/proc/' + pid + '/ns/net', '--property=PrivateTmp=no', '--property=ProtectSystem=strict', '--property=ProtectHome=yes', '--property=NoNewPrivileges=yes', '--property=InaccessiblePaths=/opt/smail /run/smail', '--property=ReadWritePaths=' + str(target), '--property=MemoryMax=48M', '--property=MemorySwapMax=0', '--property=CPUQuota=25%', '--property=TasksMax=32', '--property=RuntimeMaxSec=240s', '--property=OOMPolicy=stop', '--property=KillMode=control-group', '--property=TimeoutStopSec=10s', 'python3', str(target / 'runtime_probe.py'), str(root), '--live-net', live_net, '--password-stdin', *probe_args, ], input=password + '\n' + (json.dumps(sql_fixture.settings) + '\n' if sql_fixture else ''), text=True, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, timeout=270) (target / 'probe.log').write_text(probe.stdout, encoding='utf-8') result['probe_code'] = probe.returncode if (target / 'probe-result.json').exists(): result['probe'] = json.loads((target / 'probe-result.json').read_text()) assert probe.returncode == 0, '隔离验收失败，请检查保留的 probe.log' assert result['probe']['passed'] with closing(sqlite3.connect(f'file:{live / "data/smail.db"}?mode=ro', uri=True)) as database: message_id = result['probe']['smtp']['message_id'] assert database.execute('SELECT COUNT(*) FROM email WHERE msg_id=?', (message_id,)).fetchone()[0] == 0 result['live_message_absent'] = True if args.migration: if args.interruption: if sql_fixture: from runtime_sql_interruption import verify_interruption verify_interruption(root, unit, password, result, sql_fixture) else: from runtime_interruption import verify_interruption verify_interruption(root, unit, password, result) if args.switch_fault: from runtime_sql_switch_fault import verify_switch_fault verify_switch_fault(root, unit, password, result, sql_fixture, migration_target, args.switch_fault) elif mixed: from runtime_mixed_migration import verify_migration verify_migration(root, unit, password, result, mixed_source, mixed_target) elif migration_target: from runtime_sql_migration import verify_migration verify_migration(root, unit, password, result, sql_fixture, migration_target) else: from runtime_migration import verify_migration verify_migration(root, unit, password, result) finally: try: collect_and_stop(result, unit, probe_unit, proxy_unit if args.nginx_template else None) result['after'] = service_state() with (live / 'smail').open('rb') as stream: result['live_unchanged'] = (result['after'] == before and hashlib.file_digest(stream, 'sha256').hexdigest() == live_sha) try: if mixed: from runtime_mixed_migration import migration_integrity result['integrity'] = migration_integrity(root, mixed_source, mixed_target) elif migration_target: from runtime_sql_migration import assert_active active_database = sql_fixture if args.switch_fault == 'before-root' else migration_target assert_active(root, active_database) result['integrity'] = active_database.integrity(root) elif args.migration: from runtime_migration import migration_integrity result['integrity'] = migration_integrity(root, interrupted=args.interruption) else: result['integrity'] = sql_fixture.integrity(root) if sql_fixture else database_integrity(root) except (AssertionError, sqlite3.Error, RuntimeError, OSError, ValueError, subprocess.SubprocessError) as error: result['integrity'] = {'error': type(error).__name__} finally: result.setdefault('cleanup_errors', []) if bridges: stop_database_bridges(bridges, result) if sql_fixture: try: sql_fixture.lock() result['test_database_account_locked'] = True except Exception as error: result['cleanup_errors'].append({'action': 'database_account', 'error': type(error).__name__}) if migration_target: try: migration_target.lock() result['migration_target_account_locked'] = True except Exception as error: result['cleanup_errors'].append({'action': 'migration_target_account', 'error': type(error).__name__}) if network_unit: errors = stop_test_units([network_unit]) result['cleanup_errors'].extend(errors) result['network_anchor_stopped'] = not errors (target / 'service.log').write_text(run('journalctl', '-u', unit, '--no-pager'), encoding='utf-8') kernel = run('journalctl', '-k', '--since', started, '--no-pager') result['oom_events'] = sum('oom-kill:constraint=' in line for line in kernel.splitlines()) (target / 'result.json').write_text(json.dumps(result, ensure_ascii=False, indent=2), encoding='utf-8') print(json.dumps(result, ensure_ascii=False, indent=2)) assert result['live_unchanged'] and result['oom_events'] == 0 assert not result['cleanup_errors'], '测试单元未全部停止，需要核查保留的结果' assert not result['resource_errors'], '资源证据不完整，需要核查保留的结果' if mixed: assert result['migration']['passed'] and result['network_anchor_stopped'] assert result['integrity']['config_records'] > 0 elif sql_fixture: assert result['integrity'].get('sqlite_files') == 0 and result['integrity']['config_records'] > 0 if args.migration: assert result['migration']['passed'] and result['network_anchor_stopped'] elif args.migration: assert result['migration']['passed'] and result['integrity']['active_database'] == 'migrated.db' else: assert result['integrity'] == {'smail.db': [('ok',)]} if args.restore: from runtime_restore import verify_backup_restore restored = verify_backup_restore(root, result, password, args.nginx_template is not None) (target / 'restore-result.json').write_text(json.dumps(restored, ensure_ascii=False, indent=2), encoding='utf-8') print(json.dumps({'restore': restored}, ensure_ascii=False, indent=2)) if __name__ == '__main__': main()