Kafka Partition

Introduction to Kafka Partition

In Kafka, the data is store in the Kafka topics. The Kafka topic will further be divided into multiple partitions. The actual messages or the data will store in the Kafka partition. It is directly proportional to the parallelism. If we have increased the number of partition then we can run the multiple parallel jobs on the same Kafka topic. The partition level is also depending on the Kafka broker as well. We cannot define the “n” number of the partition to the Kafka topic. We need to define the partition as per the Kafka broker availability. In partition, the data is storing with the help of keys.

Syntax:

As such, there is no specific syntax available for the Kafka Partition. Generally, we are using the Kafka partition value while creating the new topic or defining the number of partitions on the Kafka commands.

Note 1) while working with the Kafka Partition. We are using the core Kafka commands and Kafka Partition command for the troubleshooting front.

2) At the time of Kafka Partition configuration; we are using the CLI method. But generally, we are using the UI tool only.

How Kafka Partition Works?

The Kafka Partition is useful to define the destination partition of the message. Internally the Kafka partition will work on the key bases i.e. the null key and the hash key. If it will set the null key then the messages or data will store at any partition or the specific hash key provided then the data will move on to the specific partition. Majorly the Kafka partition is deal with parallelism. If we are increasing the number of partition, it will also increase the parallel process also.

Below are the list of property and its value that we can use in the Kafka partition.

Sr No	Property	Value	Description
1	broker.id		In the Kafka partition, we need to define the broker id by the non-negative integer id. The broker’s name will include the combination of the hostname as well as the port name. We have used single or multiple brokers as per the requirement.
2	log.dirs	/tmp/kafka-logs	It will be a single or multiple Kafka data store location. While creating the new partition it will be placed in the directory.
3	port	6667	On the 6667 port no, the server will accept the client connections.
4	zookeeper.connect	null	We need to specify the zookeeper connection in the form the hostname and the port i.e. 2181. We need to define the multiple zookeeper hostname and port in the same partition command. Zookeeper is very important while communicating with the Kafka environment. It is the primary thing to communicate with the Kafka environment.
5	message.max.bytes	1000000	With the help of Kafka partition command, we can also define the maximum size of a message. The same count of messages that the server will receive. It is very important that the same property be in sync with the maximum fetch value with the consumer front.
6	num.network.threads	3	We can define the same value to handle the number of network threads. The same count, the server will uses for managing the network requests. We don’t need to change this value.
7	num.io.threads	8	We can define the same value to handle the number of input and output threads. We can define the number of thread as per the disk availability.
8	background.threads	4	It will help to manage the various background processes like the file deletion. Generally, we are not changing the same value.
9	queued.max.requests	500	With the help of this property, we can define the number of requests that can be queued up. It will help for the I/O threads.
10	host.name	null	It will help to define the property as the hostname of the Kafka broker. If we have set the same property then it will only bind with the same address. If the same value will not set then it will bind with all the present interfaces and will publish on the zookeeper.
11	advertised.host.name	null	It will help to connect with the multiple Kafka components like the consumers, producers, brokers, etc. As per the configuration, we can define the value like hostname or the ip address. In some cases, we can find it like 0.0.0.0 with the port number.
12	advertised.port	null	The advertised.port value will give out to the consumers, producers, brokers. It will help to establish the connections. We can define the value in different form as well.
13	socket.send.buffer.bytes	100 * 1024	It is also known as the SO_SNDBUFF buffer. It will prefer for server socket connections.
14	socket.receive.buffer.bytes	100 * 1024	It is also known as the SO_RCVBUFF buffer. It will prefer for server socket connections.
15	socket.request.max.bytes	100 * 1024 * 1024	The socket.request.max.bytes value will help to define the request size that the server will allow.

Examples

In the Kafka environment, we can create a topic to store the messages. As per the Kafka broker availability, we can define the multiple partitions in the Kafka topic.

Note: The default port of the Kafka broker in the cluster mode may verify depend on the Kafka environment.

• The cluster Kafka broker port is “6667”.
• Zookeeper port will be “2181”
• The single Kafka broker port is “9092”
• On TLS or SSL Kafka environment, the port will be “9093”.

Command :

./kafka-topics.sh --create --zookeeper 10.10.132.70:2181 --replication-factor 1 --partitions 3 --topic elearning_kafka

Explanation :

As per the above command, we have created the “elearning_kafka” Kafka topic with the partition value 3. In screenshot 1 (B), we have seen the 3 partition is available in the “elearning_kafka” topic.

Output:

Screenshot 1 (A)

Screenshot 1 (B)

Conclusion

We have seen the uncut concept of “Kafka Partition” with the proper example, explanation, and methods with different outputs. As per the requirement, we can create multiple partitions in the topic. It will increase the parallelism of get and put operation.

Recommended Articles

This is a guide to Kafka Partition. Here we discuss the definition, How to Works Kafka Partition, and how to implement Kafka Partition. You can also go through our other related articles to learn more –

1. Kafka Zookeeper
2. Kafka Console Producer
3. Kafka Monitoring
4. Kafka Node