How to create a Topic in Kafka through Java
Edit - Zookeeper is not required in newer version of Kafka. Please see answer by @Neeleshkumar Srinivasan Mannur for API version 0.11.0+
Original answer
I fixed it.. After a long research..
ZkClient zkClient = new ZkClient("localhost:2181", 10000, 10000);
AdminUtils.createTopic(zkClient, myTopic, 10, 1, new Properties());
From the above code, ZkClient will create a topic but this topic information will not have awareness for the kafka. So what we have to do is, we need to create object for ZkClient in following way,
First import the below statement,
import kafka.utils.ZKStringSerializer$;
and create object for ZkClient in the following way,
ZkClient zkClient = new ZkClient("localhost:2181", 10000, 10000, ZKStringSerializer$.MODULE$);
AdminUtils.createTopic(zkClient, myTopic, 10, 1, new Properties());
Edit 1: (for @ajkret comment)
The above code won't work for kafka > 0.9 since the api has been changed, Use the below code for kafka > 0.9
import java.util.Properties;
import kafka.admin.AdminUtils;
import kafka.utils.ZKStringSerializer$;
import kafka.utils.ZkUtils;
import org.I0Itec.zkclient.ZkClient;
import org.I0Itec.zkclient.ZkConnection;
public class KafkaTopicCreationInJava
{
public static void main(String[] args) throws Exception {
ZkClient zkClient = null;
ZkUtils zkUtils = null;
try {
String zookeeperHosts = "192.168.20.1:2181"; // If multiple zookeeper then -> String zookeeperHosts = "192.168.20.1:2181,192.168.20.2:2181";
int sessionTimeOutInMs = 15 * 1000; // 15 secs
int connectionTimeOutInMs = 10 * 1000; // 10 secs
zkClient = new ZkClient(zookeeperHosts, sessionTimeOutInMs, connectionTimeOutInMs, ZKStringSerializer$.MODULE$);
zkUtils = new ZkUtils(zkClient, new ZkConnection(zookeeperHosts), false);
String topicName = "testTopic";
int noOfPartitions = 2;
int noOfReplication = 3;
Properties topicConfiguration = new Properties();
AdminUtils.createTopic(zkUtils, topicName, noOfPartitions, noOfReplication, topicConfiguration);
} catch (Exception ex) {
ex.printStackTrace();
} finally {
if (zkClient != null) {
zkClient.close();
}
}
}
}
For those trying to achieve this in kafka v0.10.2.1 and running into issues with serialization error 'java.io.StreamCorruptedException: invalid stream header: 3139322E
' following is a sample working code with the needful imports.
import org.I0Itec.zkclient.ZkClient;
import org.I0Itec.zkclient.ZkConnection;
import org.I0Itec.zkclient.exception.ZkMarshallingError;
import org.I0Itec.zkclient.serialize.ZkSerializer;
import org.apache.kafka.clients.consumer.KafkaConsumer;
import org.apache.kafka.common.PartitionInfo;
import kafka.admin.AdminUtils;
import kafka.admin.RackAwareMode;
import kafka.utils.ZKStringSerializer;
import kafka.utils.ZkUtils;
public static void createTopic(String topicName, int numPartitions, int numReplication) {
ZkClient zkClient = null;
ZkUtils zkUtils = null;
try {
String zookeeperHosts = "199.98.916.902:2181"; // If multiple zookeeper then -> String zookeeperHosts = "192.168.20.1:2181,192.168.20.2:2181";
int sessionTimeOutInMs = 15 * 1000; // 15 secs
int connectionTimeOutInMs = 10 * 1000; // 10 secs
zkClient = new ZkClient(zookeeperHosts, sessionTimeOutInMs, connectionTimeOutInMs);
//Ref: https://gist.github.com/jjkoshy/3842975
zkClient.setZkSerializer(new ZkSerializer() {
@Override
public byte[] serialize(Object o) throws ZkMarshallingError {
return ZKStringSerializer.serialize(o);
}
@Override
public Object deserialize(byte[] bytes) throws ZkMarshallingError {
return ZKStringSerializer.deserialize(bytes);
}
});
zkUtils = new ZkUtils(zkClient, new ZkConnection(zookeeperHosts), false);
int noOfPartitions = 2;
int noOfReplication = 3;
Properties topicConfiguration = new Properties();
AdminUtils.createTopic(zkUtils, topicName, noOfPartitions, noOfReplication, topicConfiguration,
RackAwareMode.Enforced$.MODULE$);
} catch (Exception ex) {
ex.printStackTrace();
} finally {
if (zkClient != null) {
zkClient.close();
}
}
}
The process seems to be pretty much simplified in API 0.11.0+. Using that, it can be done as follows
import org.apache.kafka.clients.admin.AdminClient;
import org.apache.kafka.clients.admin.CreateTopicsResult;
import org.apache.kafka.clients.admin.NewTopic;
Properties properties = new Properties();
properties.load(new FileReader(new File("kafka.properties")));
AdminClient adminClient = AdminClient.create(properties);
NewTopic newTopic = new NewTopic("topicName", 1, (short)1); //new NewTopic(topicName, numPartitions, replicationFactor)
List<NewTopic> newTopics = new ArrayList<NewTopic>();
newTopics.add(newTopic);
adminClient.createTopics(newTopics);
adminClient.close();
The contents of kafka.properties
file are as follows
bootstrap.servers=localhost:9092
group.id=test
enable.auto.commit=true
auto.commit.interval.ms=1000
key.deserializer=org.apache.kafka.common.serialization.StringDeserializer
value.deserializer=org.apache.kafka.common.serialization.StringDeserializer
Note that the instance of the AdminClient must be closed in order to reflect the newly created topic.
Just a pointer to anyone looking at this with a updated version of Kafka (At the time of writing this, I was using Kafka v0.10.0.0).
You have to change;
AdminUtils.createTopic(zkUtils, topicName, noOfPartitions, noOfReplications, topicConfiguration);
To the following;
AdminUtils.createTopic(zkUtils, topicName, noOfPartitions, noOfReplications, true, Enforced$.MODULE$);
It is also a good idea to close the connection once finished;
zkClient.close();