Skip to main content

Apache Kafka 4.1.1 – Single Node KRaft Mode on CentOS 8 (No ZooKeeper)

Apache Kafka 4.1.1 Installation– (Single Node KRaft Mode, No ZooKeeper)Mode on CentOS 8 (No ZooKeeper)

This guide shows how to install and run Apache Kafka 4.1.1 in KRaft mode (no ZooKeeper) on CentOS 8, using a single-node broker + controller setup.

1. Install Java 17 (Required for Kafka 4.x)

Kafka 4.x requires Java 17+.

sudo dnf install -y java-17-openjdk java-17-openjdk-devel

Verify:

java -version

You should see something like:

openjdk version "17.x.x" ...

2. Create a Dedicated Kafka User

sudo useradd kafka -m
sudo usermod -s /bin/bash kafka

3. Download & Install Kafka 4.1.1

From your home directory:

cd ~
wget https://dlcdn.apache.org/kafka/4.1.1/kafka_2.13-4.1.1.tgz
tar -xvzf kafka_2.13-4.1.1.tgz

Move Kafka to /opt and set permissions:

sudo mv kafka_2.13-4.1.1 /opt/kafka
sudo chown -R kafka:kafka /opt/kafka

4. Create Data Directory

sudo mkdir -p /opt/kafka/data
sudo chown -R kafka:kafka /opt/kafka

5. Configure Kafka for Single-Node KRaft Mode

Edit the main Kafka config file:

sudo nano /opt/kafka/config/server.properties

Add/update:Replace the contents with the working KRaft config:

process.roles=broker,controller
node.id=1
controller.listener.names=CONTROLLER
controller.quorum.voters=1@localhost:9093

listeners=PLAINTEXT://:9092,CONTROLLER://:9093
controller.listener.names=CONTROLLERadvertised.listeners=PLAINTEXT://localhost:9092
inter.broker.listener.name=PLAINTEXT

num.partitions=1
offsets.topic.replication.factor=1
transaction.state.log.replication.factor=1
transaction.state.log.min.isr=1

log.dirs=/opt/kafka/data

Explanation (short):

  • process.roles=broker,controller → single process runs both broker + controller.
  • controller.quorum.voters=1@localhost:9093 → 1-node KRaft quorum.
  • listeners / advertised.listeners → Kafka is reachable at localhost:9092.
  • offsets.topic.* and transaction.state.* → required so internal topics (like __consumer_offsets) can be created.
  • log.dirs → where Kafka stores data.

6. Generate KRaft Cluster UUID

for

Run Storageas any user (root or your normal user is fine):

UUID=$(/opt/kafka/bin/kafka-storage.sh random-uuid)
echo $UUID

Copy the printed UUID.

7. Format the KRaft Storage

This initializes the KRaft metadata directory.

sudo -u kafka /opt/kafka/bin/kafka-storage.sh format   -t $UUID   -c /opt/kafka/config/server.properties

Expected:Expected output includes something like:

Formatting metadata directory /opt/kafka/data with metadata.version 4.1-IV1.

8. (Optional) SELinux Configuration

On CentOS 8, SELinux may block Kafka when installed under /opt.

Check mode:

getenforce

If it shows Enforcing, you can temporarily relax it:

sudo setenforce 0

For a permanent setting (optional, depends on your security policy):

sudo nano /etc/selinux/config
# Set:
SELINUX=permissive

Then reboot later to apply permanently.

9. Test Kafka Manually (Foreground)

Run Kafka as the kafka user:

sudo -u kafka /opt/kafka/bin/kafka-server-start.sh /opt/kafka/config/server.properties

If all is correct, you should see logs ending with something like:

Kafka starts,Server pressstarted in KRaft mode

Press Ctrl + C to stop.stop Kafka.

9. Fix SELinux (CentOS 8)

Check:

getenforce

If Enforcing, temporarily allow:

sudo setenforce 0

Permanent:

sudo nano /etc/selinux/config
# set:
SELINUX=permissive

10. Create a Systemd Service for Kafka

Create a systemd unit:

sudo nano /etc/systemd/system/kafka.service

Add:Paste:

[Unit]
Description=Kafka KRaft Server
After=network.target

[Service]
User=kafka
Environment="JAVA_HOME=/usr/lib/jvm/java-17-openjdk"
Environment="PATH=/usr/lib/jvm/java-17-openjdk/bin:/usr/bin:/bin"
ExecStart=/opt/kafka/bin/kafka-server-start.sh /opt/kafka/config/server.properties
ExecStop=/opt/kafka/bin/kafka-server-stop.sh
Restart=on-failure
LimitNOFILE=100000

[Install]
WantedBy=multi-user.target

Reload &systemd Start:and start Kafka:

sudo systemctl daemon-reload
sudo systemctl start kafka
sudo systemctl enable kafka
sudo systemctl status kafka

Status should show:

Active: active (running)

11. Verify Data Directory & Internal Topics

Check Kafka’s data directory:

ls -l /opt/kafka/data

You should eventually see directories like:

  • __cluster_metadata-0
  • __consumer_offsets-0
  • test-topic-0 (after you create a topic)

The presence of __consumer_offsets-0 confirms that Kafka successfully auto-created the internal offsets topic required for consumer groups.

12. Create and Test Kafkaa Topic

12.1 Create a Topic

/opt/kafka/bin/kafka-topics.sh --create   --topic test-topic   --bootstrap-server localhost:9092

List Topics

topics:

/opt/kafka/bin/kafka-topics.sh --list --bootstrap-server localhost:9092

Expected:

test-topic

12.2 Run a Producer

/opt/kafka/bin/kafka-console-producer.sh   --topic test-topic   --bootstrap-server localhost:9092

Type a few messages and press ENTER after each:

hi
hello
this is a test

12.3 Run a Consumer

In another terminal:

/opt/kafka/bin/kafka-console-consumer.sh   --topic test-topic   --bootstrap-server localhost:9092   --from-beginning

You should see:

hi
hello
this is a test

If you want to test with a new consumer group:

/opt/kafka/bin/kafka-console-consumer.sh   --topic test-topic   --bootstrap-server localhost:9092   --group debug-group-1   --from-beginning

Completed13. Debugging Tips

13.1 Consumer Shows No Messages

Check:

ls -l /opt/kafka/data

If __consumer_offsets-0 is missing:

  1. Ensure these properties exist in server.properties:

    num.partitions=1
offsets.topic.replication.factor=1
transaction.state.log.replication.factor=1
transaction.state.log.min.isr=1

  2. Restart Kafka:

    sudo systemctl restart kafka

  3. Check logs:

    journalctl -u kafka -n 200 --no-pager

13.2 Check Consumer Groups

/opt/kafka/bin/kafka-consumer-groups.sh --bootstrap-server localhost:9092 --list

Describe a group:

/opt/kafka/bin/kafka-consumer-groups.sh   --bootstrap-server localhost:9092   --group debug-group-1   --describe

14. Summary

You now have:

  • Kafka 4.1.1 isinstalled nowon runningCentOS 8
  • Running in single-node KRaft mode (no ZooKeeper)
  • Proper server.properties for KRaft
  • Kafka managed by systemd
  • Verified producer and consumer flow on topic test-topic

This setup is ideal for development, POC, and local testing without ZooKeeper.

Back to top