humio.kafka

DEPRECATED. This repository will no longer receive updates. Please consider using humio-operator to install and manage Logscale: Installation Guide.

Kafka Installer for Use with Humio

Requirements

This role requires access to a Zookeeper installation.

Role Variables

Check the defaults.

If the machine has no internet access, set the kafka_mirror variable to "master" to copy the Kafka tarball from the master's files directory:

kafka_mirror: "master"

Rack Awareness

Kafka can recognize different racks if you specify which rack each Kafka node is in. You can set this up by using the broker.rack setting in Kafka. This is done in this Ansible role using the kafka_broker_rack variable. To set this up correctly, define each rack or data center as a group in your inventory and assign the right machines to it. Then, you can either set the kafka_broker_rack variable in your inventory or create a group_vars file for each rack/data center group to define it properly.

By default, all Kafkas are set to a single rack called dc1.

Dependencies

You need to have Java 9 installed. The humio.java role is recommended.

Example Playbook

- hosts: servers
  become: true
  vars:
    zookeeper_hosts:
      - zookeeper_id: 1
        ip: "{{ ansible_default_ipv4.address }}"
  roles:
    - role: humio.java
    - role: AnsibleShipyard.ansible-zookeeper
    - role: humio.kafka

Troubleshooting

If you need to stop the Ansible run during the Install Kafka from remote step, you may face an issue where the tarball being fetched and unarchived gets interrupted. If this happens while fetching, you can simply run the playbook again. However, if it happens during the unarchiving, you will need to manually delete the /usr/lib/kafka_{{ kafka_scala_version }}-{{ kafka_version }} directory before running the playbook again, as this step checks for that directory's existence to skip the process in future runs.

License

Apache 2.0