在本地搭建 Kafka

Let's run Kafka on your own machine — perfect for learning, testing, and development. A local setup is your personal sandbox, no cloud required.

First, the prerequisites: JDK 8+ (Kafka runs on the JVM) and any major OS — Linux, macOS, or Windows via WSL. Check with java -version.

Now download the Kafka binary from the official Apache site — grab the latest stable release. The shown curl command pulls it straight to your machine.

Next, extract the downloaded .tgz archive with tar -xzf, then cd into the new folder. You're now inside the Kafka install directory.

ZooKeeper is the distributed coordinator Kafka has long used to track brokers and store metadata. Newer Kafka uses KRaft instead, but ZooKeeper is still common for learning.

Start ZooKeeper (bundled with Kafka) using the shown command. It'll spew logs — keep this terminal open, since ZooKeeper has to stay running.

A Kafka broker is the server running the Kafka process — the heart of the cluster. It stores messages in topics and serves producers and consumers. We'll run one locally.

With ZooKeeper up, open a new terminal and start the broker with the shown command. Keep it running too — you now have a single-node Kafka cluster live.

Confirm it works: in a new terminal, create a test topic with the shown kafka-topics.sh command. No errors means your local Kafka is running.

You've just learned the steps to get Kafka running locally. Can you put them in the correct order?

You did it — a local Kafka cluster is running. You installed Java, extracted the binaries, started ZooKeeper and the broker, and verified with a test topic. Next: producing and consuming.