Skip to content
This repository has been archived by the owner on Nov 15, 2024. It is now read-only.

humio/ansible-server

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ansible-humio

DEPRECATED. This repo will no longer get any updates. Please consider using humio-operator to install and manage Logscale https://library.humio.com/falcon-logscale-self-hosted/installation-containers-kubernetes.html.

Installer for Humio. Current features

  • Launch one Humio instance per CPU socket

Requirements

A running Zookeeper and Kafka installation.

Role Variables

humio_host_id: 0
humio_public_url: "http://{{ ansible_default_ipv4.address }}"
humio_socket_bind: "0.0.0.0"
humio_http_bind: "{{ ansible_eth0.ipv4.address }}"
humio_config:
  all: |
    AUTHENTICATION_METHOD=byproxy
    AUTH_BY_PROXY_HEADER_NAME=name-of-http-header
  "0": |
    HUMIO_HTTP_BIND={{ ansible_eth0_0.ipv4.address }}
  "1": |
    HUMIO_HTTP_BIND={{ ansible_eth0_1.ipv4.address }}
zookeeper_hosts:
  - ip: "{{ ansible_default_ipv4.address }}"
kafka_hosts:
  - ip: "{{ ansible_default_ipv4.address }}"

humio_total_memory_mb can be used to override the memory used by Humio. This defaults to using all memory available on the machine. The JVM memory settings will be determined automatically based on that number using our recommended formula. If you want more fine-grained control, you can set the JVM memory options individually. See the defaults/main.yml file for which variables to override.

Example Playbook

To run a singlenode Humio instance pointing to existing zookeeper and kafka instances.

- hosts: humios
  become: true
  vars:
    kafka_hosts:
      - ip: "172.17.0.2"
    zookeeper_hosts:
      - ip: "172.17.0.2"
  roles:
    - role: humio.server

License

Apache 2.0

Author Information

If you're having issues with this module please to not hesitate to reach out to us on the #ansible channel on our Slack community team.