Ansible role to install and configure Apache Kafka 2.8.0
Apache Kafka is a distributed event streaming platform using publish-subscribe topics. Applications and streaming components can produce and consume messages by subscribing to these topics. Kafka is extremely fast, handling megabytes of reads and writes per second from thousands of clients. Messages are persisted and replicated to prevent data loss. Data streams are partitioned and can be elastically scaled with no downtime.
- RedHat 6
- RedHat 7
- RedHat 8
- Debian 10.x
- Ubuntu 18.04.x
- Ubuntu 20.04.x
- Apache ZooKeeper
- Java 8 / 11
The below Apache ZooKeeper role from Ansible Galaxy can be used if one is needed.
ansible-galaxy install sleighzy.zookeeperAnsible 2.9.16 or 2.10.4 are the minimum required versions to workaround an
issue with certain kernels that have broken the systemd status check. The
error message "Service is in unknown state" will be output when attempting to
start the service via the Ansible role and the task will fail. The service will
start as expected if the systemctl start command is run on the physical host.
See ansible/ansible#71528 for more information.
| Variable | Default |
|---|---|
| kafka_download_base_url | http://www-eu.apache.org/dist/kafka |
| kafka_version | 2.8.0 |
| kafka_scala_version | 2.13 |
| kafka_create_user_group | true |
| kafka_user | kafka |
| kafka_group | kafka |
| kafka_root_dir | /opt |
| kafka_dir | {{ kafka_root_dir }}/kafka |
| kafka_start | yes |
| kafka_restart | yes |
| kafka_log_dir | /var/log/kafka |
| kafka_broker_id | 0 |
| kafka_java_heap | -Xms1G -Xmx1G |
| kafka_background_threads | 10 |
| kafka_listener_protocol | PLAINTEXT |
| kafka_listener_hostname | localhost |
| kafka_listener_port | 9092 |
| kafka_num_network_threads | 3 |
| kafka_num_io_threads | 8 |
| kafka_num_replica_fetchers | 1 |
| kafka_socket_send_buffer_bytes | 102400 |
| kafka_socket_receive_buffer_bytes | 102400 |
| kafka_socket_request_max_bytes | 104857600 |
| kafka_replica_socket_receive_buffer_bytes | 65536 |
| kafka_data_log_dirs | /var/lib/kafka/logs |
| kafka_num_partitions | 1 |
| kafka_num_recovery_threads_per_data_dir | 1 |
| kafka_log_cleaner_threads | 1 |
| kafka_offsets_topic_replication_factor | 1 |
| kafka_transaction_state_log_replication_factor | 1 |
| kafka_transaction_state_log_min_isr | 1 |
| kafka_log_retention_hours | 168 |
| kafka_log_segment_bytes | 1073741824 |
| kafka_log_retention_check_interval_ms | 300000 |
| kafka_auto_create_topics_enable | false |
| kafka_delete_topic_enable | true |
| kafka_default_replication_factor | 1 |
| kafka_group_initial_rebalance_delay_ms | 0 |
| kafka_zookeeper_connect | localhost:2181 |
| kafka_zookeeper_connection_timeout | 6000 |
| kafka_bootstrap_servers | localhost:9092 |
| kafka_consumer_group_id | kafka-consumer-group |
- The Kafka service can be started via:
systemctl start kafka - The Kafka service can be stopped via:
systemctl stop kafka
- The Kafka service can be started via:
service kafka start - The Kafka service can be stopped via:
service kafka stop
| Property | Value |
|---|---|
| ZooKeeper connection | localhost:2181 |
| Kafka bootstrap servers | localhost:9092 |
| Kafka consumer group Id | kafka-consumer-group |
| Kafka broker id | 0 |
| Number of partitions | 1 |
| Data log file retention period | 168 hours |
| Enable auto topic creation | false |
| Enable topic deletion | true |
| Port | Description |
|---|---|
| 9092 | Kafka listener port |
| Directory / File | |
|---|---|
| Kafka installation directory (symlink to installed version) | /opt/kafka |
| Kafka configuration directory (symlink to /opt/kafka/config) | /etc/kafka |
| Directory to store data files | /var/lib/kafka/logs |
| Directory to store logs files | /var/log/kafka |
| Kafka service | /usr/lib/systemd/system/kafka.service |
Add the below to a playbook to run those role against hosts belonging to the
kafka-nodes group.
- hosts: kafka-nodes
roles:
- sleighzy.kafkaLinting should be done using ansible-lint.
pip3 install ansible-lint --userThis module uses the [Ansible Molecule] testing framework. This test suite creates a Kafka and ZooKeeper cluster consisting of three nodes running within Docker containers. Each container runs a different OS to test the supported platforms for this Ansible role.
As per the Molecule Installation guide this should be done using a virtual environment. The commands below will create a Python virtual environment and install Molecule including the Docker driver.
$ python3 -m venv molecule-venv
$ source molecule-venv/bin/activate
(molecule-venv) $ python3 -m pip install molecule-docker "molecule[docker,lint]"Run playbook and tests. Linting errors need to be corrected before Molecule will execute any tests. This will run all tests and then destroy the Docker containers.
molecule testThe below command can be used to run the playbook without the tests. This can be run multiple times when making changes to the role, and ensuring that operations are idempotent.
molecule convergeThe below commands can be used to just run the tests without tearing everything
down. The command molecule verify can be repeated for each test run.
molecule create
molecule converge
molecule verifyTear down Molecule tests and Docker containers.
molecule destroy