冪等性と整合性
冪等性と整合性
Section titled “冪等性と整合性”分散システムでは「1回しか実行されない」は幻想。「何度実行されても最終状態が正しい」ことを保証する設計を詳しく解説します。
冪等性の重要性
Section titled “冪等性の重要性”分散システムでは、ネットワークエラー、タイムアウト、再起動などにより、同じ処理が複数回実行される可能性があります。
問題のあるコード
Section titled “問題のあるコード”// ❌ 問題のあるコード: 非冪等な処理class OrderService{ public function createOrder($orderData) { // 問題: 再実行時に注文が二重作成される return Order::create($orderData); }}なぜ問題か:
- 再実行時の二重作成: ネットワークエラーでクライアントが再送すると、注文が2つ作成される
- データの不整合: 同じ注文が複数存在し、在庫や決済に影響する
冪等性の担保
Section titled “冪等性の担保”Idempotency Keyの使用
Section titled “Idempotency Keyの使用”// ✅ 良い例: Idempotency Keyによる冪等性の担保class OrderService{ public function createOrder($orderData, $idempotencyKey) { // Idempotency Keyで既存の注文を確認 $existingOrder = Order::where('idempotency_key', $idempotencyKey)->first();
if ($existingOrder) { // 既に存在する場合は、既存の注文を返す return $existingOrder; }
// 新規作成 return Order::create(array_merge($orderData, [ 'idempotency_key' => $idempotencyKey, ])); }}
// マイグレーションでIdempotency Keyを追加Schema::table('orders', function (Blueprint $table) { $table->string('idempotency_key')->unique()->after('id');});なぜ重要か:
- 重複防止: 同じIdempotency Keyで再実行しても、同じ結果が返される
- データの整合性: 注文の重複作成を防止
トランザクション境界
Section titled “トランザクション境界”DB取引中に外部APIを呼ばない(失敗時復旧不可)。
問題のあるコード
Section titled “問題のあるコード”// ❌ 問題のあるコード: トランザクション内で外部APIを呼ぶclass OrderService{ public function createOrder($orderData) { return DB::transaction(function () use ($orderData) { // 1. 注文を作成(DBトランザクション内) $order = Order::create($orderData);
// 2. トランザクション内で外部APIを呼ぶ(問題) $response = Http::post('https://payment-api.example.com/charge', [ 'order_id' => $order->id, 'amount' => $orderData['amount'], ]);
if (!$response->successful()) { throw new PaymentException('Payment failed'); }
// 3. 決済結果を保存 $order->update(['payment_status' => 'COMPLETED']);
return $order; }); }}なぜ問題か:
- ロールバック不可: 外部APIが成功した後にトランザクションが失敗した場合、外部APIのロールバックが困難
- データの不整合: 外部APIは成功しているが、DBには注文が存在しない状態になる可能性
Outboxパターンの実装
Section titled “Outboxパターンの実装”// ✅ 良い例: Outboxパターンによる解決class OrderService{ public function createOrder($orderData, $idempotencyKey) { return DB::transaction(function () use ($orderData, $idempotencyKey) { // 1. Idempotency Keyで既存の注文を確認 $existingOrder = Order::where('idempotency_key', $idempotencyKey)->first();
if ($existingOrder) { return $existingOrder; }
// 2. トランザクション内で注文を作成 $order = Order::create(array_merge($orderData, [ 'idempotency_key' => $idempotencyKey, ]));
// 3. Outboxテーブルに外部API呼び出しを記録(トランザクション内) OutboxEvent::create([ 'event_type' => 'PAYMENT_CHARGE', 'aggregate_id' => (string)$order->id, 'payload' => json_encode([ 'order_id' => $order->id, 'amount' => $orderData['amount'], ]), 'idempotency_key' => $idempotencyKey, 'status' => 'PENDING', ]);
// 4. トランザクションをコミット(外部APIは呼ばない) return $order; }); }}
// 別プロセスでOutboxを処理class OutboxProcessor{ public static function processPendingEvents() { $pendingEvents = OutboxEvent::where('status', 'PENDING') ->limit(10) ->get();
foreach ($pendingEvents as $event) { try { // 外部APIを呼ぶ(トランザクション外) $payload = json_decode($event->payload, true);
$response = Http::withHeaders([ 'Idempotency-Key' => $event->idempotency_key, ])->post('https://payment-api.example.com/charge', $payload);
if ($response->successful()) { $event->update(['status' => 'COMPLETED']); } else { $event->increment('retry_count'); $event->update(['status' => 'FAILED']); } } catch (\Exception $e) { $event->increment('retry_count'); $event->update(['status' => 'FAILED']); Log::error("Failed to process outbox event: {$event->id}", ['exception' => $e]); } } }}
// 定期的にOutboxを処理// app/Console/Kernel.phpprotected function schedule(Schedule $schedule){ $schedule->call(function () { OutboxProcessor::processPendingEvents(); })->everyFiveSeconds();}なぜ重要か:
- トランザクションの短縮: DBのロック時間が短縮される
- 外部障害の分離: 外部APIの障害がトランザクションに影響しない
- 再実行の容易さ: Outboxテーブルから再実行可能
- 冪等性の保証: Idempotency Keyにより重複実行を防止
再送安全なフロー
Section titled “再送安全なフロー”「結果整合性で良いデータ」と「厳密整合性が必要なデータ」を明示的に分類する。
整合性レベルの分類
Section titled “整合性レベルの分類”// 厳密整合性が必要なデータ(ACIDトランザクション)class Order extends Model{ // 注文、決済、在庫など、ビジネス的に重要なデータ}
// 結果整合性で良いデータ(イベント駆動)class OrderAnalytics extends Model{ // 分析データ、ログ、通知など、最終的に整合性が取れれば良いデータ // 集計値は最終的に整合性が取れれば良い}使い分け:
- 厳密整合性: 注文、決済、在庫など、ビジネス的に重要なデータ
- 結果整合性: 分析データ、ログ、通知など、最終的に整合性が取れれば良いデータ
冪等性と整合性のポイント:
- 冪等性の担保: Idempotency Keyで再実行を安全化
- トランザクション境界: DB取引中に外部APIを呼ばない(Outboxパターンを使用)
- 再送安全なフロー: 厳密整合性と結果整合性を明示的に分類
これらの原則により、「何度実行されても最終状態が正しい」堅牢なシステムを構築できます。