How to Set Up Kafka Servers on CentOS

Create three servers with the following specifications:

Distribution: Ubuntu 18.04

Min Specs: 2GB RAM/20GB HDD/1CPU

Install the Confluent Community Package on the Broker Nodes

On all three nodes, add the GNU Privacy Guard (GPG) key plus package repository, and then install Confluent Community, plus Java. The format should look like this:

wget -qO - https://packages.confluent.io/deb/5.2/archive.key | sudo apt-key add -

sudo add-apt-repository "deb [arch=amd64] https://packages.confluent.io/deb/5.2 stable main"

sudo apt-get update && sudo apt-get install -y openjdk-8-jre-headless confluent-community-2.12

Configure ZooKeeper

On all three nodes, edit the hosts file:

sudo vi /etc/hosts

Add the following entries to the hosts file:

10.0.1.101 zoo1

10.0.1.102 zoo2

10.0.1.103 zoo3

Edit the zookeeper config file:

sudo vi /etc/kafka/zookeeper.properties

Delete the contents of the config file and replace them with the following:

tickTime=2000

dataDir=/var/lib/zookeeper/

clientPort=2181

initLimit=5

syncLimit=2

server.1=zoo1:2888:3888

server.2=zoo2:2888:3888

server.3=zoo3:2888:3888

autopurge.snapRetainCount=3

autopurge.purgeInterval=24

Set up the zookeeper ID for each server:

sudo vi /var/lib/zookeeper/myid

On each server, set the contents of /var/lib/zookeeper/myid to the server's ID. On Node 1, enter 1, on Node 2, enter 2, and lastly on Node 3, enter 3:

<server id 1, 2, or 3>

Configure Kafka

Edit the kafka config file:

sudo vi /etc/kafka/server.properties

Edit the broker.id and zookeeper.connect in the config file. Set the broker ID to the appropriate ID for each server (1 on Node 1, 2 on Node 2, and so on).

Set zookeeper.connect to zoo1:2181:

broker.id=<server id 1, 2, or 3>

...

zookeeper.connect=zoo1:2181

Start ZooKeeper and Kafka

Get the ZooKeeper and Kafka services started, plus enable them:

sudo systemctl start confluent-zookeeper

sudo systemctl enable confluent-zookeeper

sudo systemctl start confluent-kafka

sudo systemctl enable confluent-kafka

Both services should be active (running) on all three servers. Check the services to make sure that they are running:

sudo systemctl status confluent*

We can test our cluster by listing the current topics:

kafka-topics --list --bootstrap-server localhost:9092

Since we have not created any topics yet, we will only see a default topic. The output should look like this:

_confluent.support.metrics