search
  • Sign In
  • Sign Up
Password reset successful

Follow the proiects vou are interested in andi aet the latestnews about them taster

Table of Contents
Understanding Apache Pulsar Topic Partitioning
How to create a partition topic
1. Automatic partition topic creation through Broker configuration
2. Use Pulsar Admin API to manually create partition topics
Key limitation: Immutability of partitioned topics
Summarize
Home Java javaTutorial Creation and management of Apache Pulsar partition topics

Creation and management of Apache Pulsar partition topics

Jan 01, 2026 am 04:12 AM

Creation and management of Apache Pulsar partition topics

This article details the two main methods of creating partitioned topics in Apache Pulsar: automatic partitioned topic creation by configuring the Broker, and manual creation of partitioned topics using the Pulsar Admin API. It is emphasized that the partition topic must be specified at the time of creation. Once a topic is created as a non-partition type, it cannot be subsequently converted to a partition topic. This is crucial for system design and planning.

Understanding Apache Pulsar Topic Partitioning

Apache Pulsar's topic partitioning is a key mechanism for achieving high throughput and scalability. A partitioned topic is logically divided into multiple independent "partitions", each of which can process messages independently. Producers can send messages to specific partitions or automatically distributed by Pulsar, and each consumer instance in the consumer group can consume one or more partitions, thereby achieving parallel consumption. Proper planning and creation of partition topics is critical to building high-performance Pulsar applications.

In Pulsar, a topic's partition properties are determined when it is created, and once created, cannot be changed. This means that a non-partitioned topic cannot be converted to a partitioned topic in a subsequent operation. Therefore, when designing the Pulsar messaging system, it is crucial to determine in advance whether the topic needs partitions and the number of partitions.

How to create a partition topic

There are two main strategies for creating partition topics: automatic creation through Broker configuration, or manual creation using the Pulsar Admin API.

1. Automatic partition topic creation through Broker configuration

When a producer or consumer attempts to connect to a topic that does not yet exist, Pulsar Broker can automatically create the topic based on its configuration. By adjusting the Broker configuration, you can ensure that the automatically created topic is the partition topic by default.

The following are the relevant Broker configuration parameters:

  • allowAutoTopicCreationType: This parameter controls the type of automatic topic creation. Setting it to partitioned means that all automatically created topics will be partitioned topics.
  • defaultNumPartitions: This parameter specifies the default number of partitions for automatically created partition topics when allowAutoTopicCreationType is set to partitioned.

Configuration example:

Add or modify the following lines in Pulsar Broker's configuration file (usually conf/broker.conf):

 allowAutoTopicCreationType = partitioned
defaultNumPartitions = 5

Things to note:

  • This method is suitable for scenarios that rely on Pulsar to automatically create themes. When your Java application first connects to a non-existent topic through PulsarClient, the topic will be automatically created with the specified number of partitions if the Broker is configured with the above parameters.
  • This global configuration affects all automatically created themes. If you need to set a different number of partitions for a specific topic, or some topics do not require partitions, you need to create them manually.

2. Use Pulsar Admin API to manually create partition topics

For scenarios that require finer control, or when you do not want to rely on the automatic creation mechanism, you can use the Pulsar Admin API to manually create partition topics. The Pulsar Admin API provides a programming interface to manage various resources in the Pulsar cluster, including topics.

The Pulsar Admin client provides the topics().createPartitionedTopic(topicName, numPartitions) method to create a topic with a specified number of partitions.

Conceptual Java code example:

 import org.apache.pulsar.client.admin.PulsarAdmin;
import org.apache.pulsar.client.admin.PulsarAdminException;
import org.apache.pulsar.client.api.PulsarClientException;

public class PartitionedTopicCreator {

    private static final String BROKER_SERVICE_URL = "http://localhost:8080"; // Pulsar Admin Service URL

    public static void main(String[] args) {
        PulsarAdmin pulsarAdmin = null;
        try {
            //Initialize Pulsar Admin client pulsarAdmin = PulsarAdmin.builder()
                    .serviceHttpUrl(BROKER_SERVICE_URL)
                    .build();

            String topicName = "persistent://public/default/my-partitioned-topic";
            int numPartitions = 3; // Expected number of partitions // Check if the topic already exists to avoid repeated creation if (!pulsarAdmin.topics().getPartitionedTopicMetadata(topicName).getPartitions().isEmpty()) {
                System.out.println("Topic" topicName " already exists and is partitioned.");
            } else {
                // Create a partition topic pulsarAdmin.topics().createPartitionedTopic(topicName, numPartitions);
                System.out.println("Partition topic successfully created: " topicName ", number of partitions: " numPartitions);
            }

        } catch (PulsarClientException e) {
            System.err.println("Pulsar client exception: " e.getMessage());
            e.printStackTrace();
        } catch (PulsarAdminException e) {
            System.err.println("Pulsar Admin exception: " e.getMessage());
            e.printStackTrace();
        } finally {
            if (pulsarAdmin != null) {
                try {
                    pulsarAdmin.close();
                } catch (PulsarAdminException e) {
                    System.err.println("Failed to close Pulsar Admin client: " e.getMessage());
                }
            }
        }
    }
}

illustrate:

  • PulsarAdmin.builder().serviceHttpUrl(...) is used to build the Pulsar Admin client instance.
  • pulsarAdmin.topics().createPartitionedTopic(topicName, numPartitions) is the core method used to create a topic with a specified name and number of partitions.
  • It is a good practice to check if the topic exists before creating it to avoid PulsarAdminException.

Key limitation: Immutability of partitioned topics

As highlighted earlier, once a topic is created as a non-partitioned topic, it cannot be converted to a partitioned topic. This is a core design principle of Pulsar. If you initially create a test topic without partitioning (for example, via pulsarClient.newProducer().topic("test").create() and the Broker is not configured for automatic partitioning), then the test topic will always be single-partitioned.

If subsequent business needs require the test topic to have partitioning capabilities, your only option is:

  1. Create a new partition topic (e.g. test-partitioned) and use the partitioning strategy.
  2. Migrate data (if required) from the old non-partitioned topic to the new partitioned topic.
  3. Update all producers and consumers to connect to the new partition topic.

Therefore, in the early stages of system design, it is crucial to fully evaluate the partitioning requirements of the topic to avoid complex migration work later.

Summarize

Creating partitioned topics in Apache Pulsar is fundamental to achieving high performance and scalability. You can configure the Broker to automatically create partition topics, or use the Pulsar Admin API to manually and precisely control the topic creation process. Whichever approach you choose, you must keep in mind the immutability principle of partitioned topics: partition properties must be defined when the topic is created and cannot be changed subsequently. This feature requires developers to carefully consider the theme's partitioning strategy during the system design phase to ensure that future expansion needs are met.

The above is the detailed content of Creation and management of Apache Pulsar partition topics. For more information, please follow other related articles on the PHP Chinese website!

Statement of this Website
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn

Hot AI Tools

Undress AI Tool

Undress AI Tool

Undress images for free

AI Clothes Remover

AI Clothes Remover

Online AI tool for removing clothes from photos.

Undresser.AI Undress

Undresser.AI Undress

AI-powered app for creating realistic nude photos

ArtGPT

ArtGPT

AI image generator for creative art from text prompts.

Stock Market GPT

Stock Market GPT

AI powered investment research for smarter decisions

Popular tool

Notepad++7.3.1

Notepad++7.3.1

Easy-to-use and free code editor

SublimeText3 Chinese version

SublimeText3 Chinese version

Chinese version, very easy to use

Zend Studio 13.0.1

Zend Studio 13.0.1

Powerful PHP integrated development environment

Dreamweaver CS6

Dreamweaver CS6

Visual web development tools

SublimeText3 Mac version

SublimeText3 Mac version

God-level code editing software (SublimeText3)

How to configure Spark distributed computing environment in Java_Java big data processing How to configure Spark distributed computing environment in Java_Java big data processing Mar 09, 2026 pm 08:45 PM

Spark cannot run in local mode, ClassNotFoundException: org.apache.spark.sql.SparkSession. This is the most common first step of getting stuck: even the dependencies are not correct. Only spark-core_2.12 is written in Maven, but spark-sql_2.12 is not added. SparkSession crashes as soon as it is built. The Scala version must strictly match the official Spark compiled version - Spark3.4.x uses Scala2.12 by default. If you use spark-sqljar of 2.13, the class loader cannot directly find the main class. Practical advice: Go to mvnre

How to safely map user-entered weekday string to integer value and implement date offset operation in Java How to safely map user-entered weekday string to integer value and implement date offset operation in Java Mar 09, 2026 pm 09:43 PM

This article introduces a concise and maintainable way to map the weekday string (such as "Monday") to the corresponding serial number (1-7), and use the modulo operation to realize the forward and backward offset of any number of days (such as Monday plus 4 days to get Friday), avoiding lengthy if chains and hard-coded logic.

What is exception masking (Suppressed Exceptions) in Java_Multiple resource shutdown exception handling What is exception masking (Suppressed Exceptions) in Java_Multiple resource shutdown exception handling Mar 10, 2026 pm 06:57 PM

What is SuppressedException: It is not "swallowed", but actively archived by the JVM. SuppressedException is not an exception loss, but the JVM quietly attaches the secondary exception to the main exception under the premise that "only one exception must be thrown" for you to verify afterwards. It is automatically triggered by the JVM in only two scenarios: one is that the resource closure in try-with-resources fails, and the other is that you manually call addSuppressed() in finally. The key difference is: the former is fully automatic and safe; the latter requires you to keep it to yourself, and it can be written as shadowing if you are not careful. try-

How to use Homebrew to install Java on Mac_A must-have Java tool chain for developers How to use Homebrew to install Java on Mac_A must-have Java tool chain for developers Mar 09, 2026 pm 09:48 PM

Homebrew installs the latest stable version of openjdk (such as JDK22) by default, not the LTS version; you need to explicitly execute brewinstallopenjdk@17 or brewinstallopenjdk@21 to install the LTS version, and manually configure PATH and JAVA_HOME to be correctly recognized by the system and IDE.

How to correctly implement runtime file writing in Java applications (avoiding JAR internal write failures) How to correctly implement runtime file writing in Java applications (avoiding JAR internal write failures) Mar 09, 2026 pm 07:57 PM

After a Java application is packaged as a JAR, data cannot be written directly to the resources in the JAR package (such as test.txt) because the JAR is essentially a read-only ZIP archive; the correct approach is to write variable data to an external path (such as a user directory, a temporary directory, or a configuration-specified path).

What is the underlying principle of array expansion in Java_Java memory dynamic adjustment analysis What is the underlying principle of array expansion in Java_Java memory dynamic adjustment analysis Mar 09, 2026 pm 09:45 PM

ArrayList.add() triggers expansion because grow() is called when size is equal to elementData.length. The first add allocates 10 capacity, and subsequent expansion is 1.5 times and not less than the minimum requirement, relying on delayed initialization and System.arraycopy optimization.

Complete tutorial on reading data from file and initializing two-dimensional array in Java Complete tutorial on reading data from file and initializing two-dimensional array in Java Mar 09, 2026 pm 09:18 PM

This article explains in detail how to load an integer sequence in an external text file into a Java two-dimensional array according to a specified row and column structure (such as 2500×100), avoiding manual assignment or index out-of-bounds, and ensuring accurate data order and robust and reusable code.

A concise method in Java to compare whether four byte values ​​are equal and non-zero A concise method in Java to compare whether four byte values ​​are equal and non-zero Mar 09, 2026 pm 09:40 PM

This article introduces several professional solutions for efficiently and safely comparing multiple byte type return values ​​(such as getPlayer()) in Java to see if they are all equal and non-zero. We recommend two methods, StreamAPI and logical expansion, to avoid Boolean and byte mis-comparison errors.

Related articles