Crabka

Quickstart

This quickstart runs one local Crabka broker and verifies it with standard Kafka tooling. It is the smallest useful loop: format a data directory, start the broker, create a topic, then produce and consume one record.

For a Kubernetes cluster, skip to Operator Deployment.

Prerequisites

  • Rust toolchain matching the workspace rust-toolchain.toml.
  • A Kafka CLI distribution on your PATH for commands such as kafka-topics.sh, kafka-console-producer.sh, and kafka-console-consumer.sh.
  • A clean local data directory.

1. Format storage

export CRABKA_CLUSTER_ID=00000000-0000-0000-0000-000000000001
rm -rf target/crabka-data

cargo run -p crabka-cli --bin crabka -- format \
  --log-dir target/crabka-data \
  --cluster-id "$CRABKA_CLUSTER_ID" \
  --standalone \
  --node-id 1 \
  --controller-listener 127.0.0.1:9093

crabka format is a one-time initialization step for an empty log directory. --standalone creates a one-node KRaft controller quorum for local evaluation.

2. Start the broker

cargo run -p crabka-broker --bin crabka-broker -- \
  --log-dir target/crabka-data \
  --cluster-id "$CRABKA_CLUSTER_ID" \
  --broker-id 1 \
  --listen-addr 127.0.0.1:9092

The Kafka listener is now 127.0.0.1:9092. Prometheus metrics are exposed on :9404.

3. Use Kafka tools unmodified

In another terminal, create a topic:

kafka-topics.sh \
  --bootstrap-server 127.0.0.1:9092 \
  --create \
  --topic orders \
  --partitions 1 \
  --replication-factor 1

Produce one record:

printf 'order-1\n' | kafka-console-producer.sh \
  --bootstrap-server 127.0.0.1:9092 \
  --topic orders

Consume it back:

kafka-console-consumer.sh \
  --bootstrap-server 127.0.0.1:9092 \
  --topic orders \
  --from-beginning \
  --max-messages 1

Next steps

When you are done, stop the broker process. To start over with a fresh cluster, remove target/crabka-data and repeat the format step.

Operator Deployment →