0Pricing
RabbitMQ Messaging & Async Systems · レッスン

永続メッセージと永続キュー

Broker の再起動後もメッセージとキューを保持する方法を理解します。変動の大きい環境でも重要なデータが失われないようにします。

「永続メッセージと永続キュー」はCoddyKit上の無料RabbitMQ Messaging & Async Systemsレッスンです。 これはレッスン1/4です。 下記で完全なレッスンを無料で読むことができます。その後、ブラウザ内の組み込みコードエディタと24時間対応のAIチューターでハンズオン演習できます。 これはRabbitMQ Messaging & Async Systems学習パスの一部であり、ウェブとCoddyKitアプリ全体で進捗が同期されます。 RabbitMQ Messaging & Async Systemsコースには全4レッスンが含まれています。

このレッスンの一部はまだ翻訳されておらず、英語で表示されています。

Why Persistence Matters

Imagine your RabbitMQ broker suddenly restarts. What happens to the messages and queues you've set up?

By default, many things are temporary. If your broker crashes or restarts, non-persistent messages and non-durable queues might be lost forever. This is where persistence comes in.

Persistence ensures that critical data—your queues and messages—survive broker restarts, making your messaging system robust and reliable.

Making Queues Durable

The first step to reliability is making your queues durable. A durable queue will survive a RabbitMQ broker restart.

When you declare a queue, you pass a durable parameter. Setting this to true tells RabbitMQ to save the queue's definition to disk.

  • Durable: Queue definition saved; survives restarts.
  • Non-durable: Queue definition lost on restart.

Remember, once a queue is declared durable (or not), you can't change its durability later without deleting and re-creating it.

Code: Declaring a Durable Queue

Here's how you declare a durable queue in Java. Notice the true for the second parameter in queueDeclare.

Try running this! Even if you stop and restart RabbitMQ, this queue named 'my_durable_queue' will still be there.

import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;

public class DeclareDurableQueue {
  private final static String QUEUE_NAME = "my_durable_queue";

  public static void main(String[] argv) throws Exception {
    ConnectionFactory factory = new ConnectionFactory();
    factory.setHost("localhost");
    try (Connection connection = factory.newConnection();
         Channel channel = connection.createChannel()) {
      
      // Declare a durable queue
      channel.queueDeclare(QUEUE_NAME, true, false, false, null);
      System.out.println("Queue '" + QUEUE_NAME + "' declared as durable.");
    }
  }
}

Understanding Message Persistence

Even if a queue is durable, messages sent to it are not automatically saved to disk. For messages to survive a broker restart, they must also be marked as persistent.

When a message is persistent, RabbitMQ writes it to disk as soon as it's received by the broker, before delivering it to consumers.

This ensures that if the broker goes down, the message can be recovered from disk when the broker restarts.

Code: Sending a Persistent Message

To send a persistent message, you need to set the deliveryMode to 2 (which means persistent) or use MessageProperties.PERSISTENT_TEXT_PLAIN for convenience.

Run this producer, then try to restart your RabbitMQ broker. After restarting, if you run a consumer, you should still receive this message!

import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;
import com.rabbitmq.client.MessageProperties;

public class SendPersistentMessage {
  private final static String QUEUE_NAME = "my_durable_queue";

  public static void main(String[] argv) throws Exception {
    ConnectionFactory factory = new ConnectionFactory();
    factory.setHost("localhost");
    try (Connection connection = factory.newConnection();
         Channel channel = connection.createChannel()) {
      
      // Ensure the queue is durable first
      channel.queueDeclare(QUEUE_NAME, true, false, false, null);

      String message = "Hello, persistent world!";
      // Mark message as persistent
      channel.basicPublish("", QUEUE_NAME, 
                           MessageProperties.PERSISTENT_TEXT_PLAIN, 
                           message.getBytes("UTF-8"));
      System.out.println(" [x] Sent '" + message + "'");
    }
  }
}

The Persistence 'Gotcha'

It's crucial to understand a subtlety: even with durable queues and persistent messages, a message can still be lost.

This happens if the message is in transit from the producer to the broker, or if it has arrived at the broker but hasn't yet been flushed to disk, and the broker crashes at that exact moment.

For true reliability, where you know the broker has definitely received and saved your message, you need Publisher Confirms. We'll cover this in a future lesson!

When to Use Persistence

Persistence isn't free; it adds overhead because RabbitMQ has to write data to disk. So, when should you use it?

  • Critical Tasks: Orders, financial transactions, user sign-ups.
  • Long-running Processes: Image processing, video encoding.
  • System Integration: Ensuring data transfer between services is reliable.

If message loss is acceptable (e.g., real-time monitoring data that's quickly outdated), you can skip persistence for better performance.

Performance Impact

Writing data to disk takes more time than keeping it in memory. Therefore, enabling persistence for queues and messages will reduce your message throughput (messages per second).

  • Durable Queues: Minor impact, as it's just the queue definition.
  • Persistent Messages: More significant impact, as each message is written.

It's a trade-off: higher reliability versus higher performance. Always choose persistence for messages that you absolutely cannot afford to lose.

Recap: Persistence in Action

Let's briefly review the steps for a fully persistent setup:

  1. Declare your queue as durable (durable=true).
  2. Send your messages with a persistent delivery mode (MessageProperties.PERSISTENT_TEXT_PLAIN).
  3. For ultimate reliability (knowing the broker received it), combine with Publisher Confirms (covered later).

Quick Check: Durable Queues

You want to ensure that a queue and its messages survive if the RabbitMQ broker restarts. Which of the following is absolutely necessary for the queue itself to survive a restart?

Recap & Next Steps

Great job! You've learned how to make your RabbitMQ queues and messages persistent, a crucial step towards building reliable messaging systems.

We covered:

  • The importance of persistence.
  • How to declare durable queues.
  • How to send persistent messages.
  • The performance trade-offs and the 'gotcha' with in-flight messages.

In the next lesson, we'll dive into Publisher Confirms, which provide an even higher level of reliability by ensuring the broker has truly received and processed your messages.

よくある質問

「永続メッセージと永続キュー」レッスンは無料ですか？

はい。「永続メッセージと永続キュー」の完全なテキストはこのウェブで無料で読めます。インタラクティブに演習し（組み込みコードエディタと24時間対応のAIチューター）、RabbitMQ Messaging & Async Systemsコースの残りをアンロックするには、CoddyKit PROにアップグレードしてください。 RabbitMQ Messaging & Async Systemsコースには全4レッスンが含まれています。

「永続メッセージと永続キュー」で何を学びますか？

Broker の再起動後もメッセージとキューを保持する方法を理解します。変動の大きい環境でも重要なデータが失われないようにします。 ブラウザで直接実行するハンズオンコードでRabbitMQ Messaging & Async Systemsを演習し、24時間対応のAIチューターがレッスンを進める中での質問に答えます。

RabbitMQ Messaging & Async Systemsを始めるのに経験は必要ですか？

事前経験は必要ありません。CoddyKitのRabbitMQ Messaging & Async Systemsは初級者から上級者向けに構成されているため、ここから始めるか最初から始めて、自分のペースで進むことができます。 これはレッスン1/4です。

「永続メッセージと永続キュー」レッスンにはどのくらい時間がかかりますか？

ほとんどのCoddyKitレッスンは約5～10分かかります。各レッスンはコンパクトでインタラクティブなので、着実に進歩し、ウェブとアプリ全体で正確に前回の場所から再開できます。

このRabbitMQ Messaging & Async Systemsレッスンでコードを書いて実行できますか？

はい。すべてのRabbitMQ Messaging & Async Systemsレッスンに組み込みコードエディタが含まれているため、ブラウザでリアルコードを書いて実行し、即座のAIフィードバックを取得できます。ローカル設定は不要です。

このコースのすべてのレッスン

  1. 永続メッセージと永続キュー
  2. 信頼性を高めるPublisher Confirms
  3. Consumer Acknowledgementsと再キューイング
  4. トランザクションとPublisher Confirmの比較
← RabbitMQ Messaging & Async Systemsに戻る