Implementation Examples

Scenarios Requiring ZooKeeper in Distributed System

Here are some implementation examples to illustrate how Apache ZooKeeper can be used in various scenarios within a distributed system:

1. Configuration Management

Scenario: Centralized Configuration Storage

All services in a distributed system need to read from a central configuration store.

Implementation:

Java

import org.apache.zookeeper.*;

public class ConfigManager {
    private static ZooKeeper zk;
    private static final String CONFIG_NODE = "/config";

    public static void main(String[] args) throws Exception {
        zk = new ZooKeeper("localhost:2181", 3000, null);

        // Read configuration data
        byte[] configData = zk.getData(CONFIG_NODE, new Watcher() {
            @Override
            public void process(WatchedEvent event) {
                if (event.getType() == Event.EventType.NodeDataChanged) {
                    try {
                        byte[] newConfigData = zk.getData(CONFIG_NODE, false, null);
                        System.out.println("Configuration updated: " + new String(newConfigData));
                    } catch (Exception e) {
                        e.printStackTrace();
                    }
                }
            }
        }, null);

        System.out.println("Initial configuration: " + new String(configData));
    }
}

2. Service Discovery

Scenario: Dynamic Service Registry

Services register themselves in ZooKeeper, and other services discover them dynamically.

Implementation:

Java

import org.apache.zookeeper.*;

public class ServiceRegistry {
    private static ZooKeeper zk;
    private static final String SERVICES_NODE = "/services";

    public static void registerService(String serviceName, String serviceAddress) throws Exception {
        String path = SERVICES_NODE + "/" + serviceName;
        zk.create(path, serviceAddress.getBytes(), ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL);
        System.out.println("Service registered: " + serviceName + " at " + serviceAddress);
    }

    public static void discoverServices() throws Exception {
        zk.getChildren(SERVICES_NODE, new Watcher() {
            @Override
            public void process(WatchedEvent event) {
                if (event.getType() == Event.EventType.NodeChildrenChanged) {
                    try {
                        System.out.println("Services updated: " + zk.getChildren(SERVICES_NODE, false));
                    } catch (Exception e) {
                        e.printStackTrace();
                    }
                }
            }
        });
    }

    public static void main(String[] args) throws Exception {
        zk = new ZooKeeper("localhost:2181", 3000, null);
        zk.create(SERVICES_NODE, new byte[0], ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.PERSISTENT);
        
        // Register a service
        registerService("service1", "192.168.1.1:8080");
        
        // Discover services
        discoverServices();
    }
}

3. Leader Election

Scenario: Master Node Election

Nodes in a cluster elect a master node to coordinate tasks.

Implementation:

Java

import org.apache.zookeeper.*;

public class LeaderElection {
    private static ZooKeeper zk;
    private static final String ELECTION_NODE = "/election";
    private String currentNode;

    public void participateInElection() throws Exception {
        currentNode = zk.create(ELECTION_NODE + "/n_", new byte[0], ZooDefs.Ids.OPEN_ACL_UNSAFE, CreateMode.EPHEMERAL_SEQUENTIAL);
        checkLeadership();
    }

    private void checkLeadership() throws Exception {
        List<String> children = zk.getChildren(ELECTION_NODE, false);
        Collections.sort(children);
        String leaderNode = ELECTION_NODE + "/" + children.get(0);

        if (currentNode.equals(leaderNode)) {
            System.out.println("I am the leader");
        } else {
            System.out.println("I am not the leader. Leader is: " + leaderNode);
            zk.exists(leaderNode, new Watcher() {
                @Override
                public void process(WatchedEvent event) {
                    if (event.getType() == Event.EventType.NodeDeleted) {
                        try {
                            checkLeadership();
                        } catch (Exception e) {
                            e.printStackTrace();
                        }
                    }
                }
            });
        }
    }

    public static void main(String[] args) throws Exception {
        zk = new ZooKeeper("localhost:2181", 3000, null);
        LeaderElection le = new LeaderElection();
        le.participateInElection();
    }
}

When Does a Distributed System Need ZooKeeper?

In the constantly changing world of distributed computing, making sure that all the different parts work well together can be tough. As systems get more complicated, it’s super important to have strong tools to handle all the challenges. Apache ZooKeeper is one of the best tools for dealing with these issues.

It helps with things like managing settings, making sure everything is synced up, and coordinating actions across multiple computers.
In this article, we’ll look at specific situations where ZooKeeper is useful. We’ll see how it keeps things organized and reliable in the often chaotic world of distributed systems.

Important Topics to Understand the Need of ZooKeeper in Distributed System

What is Apache ZooKeeper?
Key Features and Capabilities of Apache Zookeeper
Use Cases and Applications of Apache Zookeeper
Scenarios Requiring ZooKeeper in Distributed System
Implementation Examples
Alternatives to ZooKeeper

Implementation Examples

1. Configuration Management

Scenario: Centralized Configuration Storage

Implementation:

2. Service Discovery

Scenario: Dynamic Service Registry

Implementation:

3. Leader Election

Scenario: Master Node Election

Implementation:

When Does a Distributed System Need ZooKeeper?

Similar Reads