acopar / ansible-elasticsearch-role

Ansible role to install and configure Elasticsearch 6.0 on RHEL/CentOS 7.

Elasticsearch

Ansible role to install and configure Elasticsearch 6.x on RHEL/CentOS 7.

NOTE: This version of the role only supports Elasticsearch 6.x. To use the role with Elasticsearch 2.4 or 5.6, use version 1.0.0 or 2.0.0, respectively.

Main features:

• Installs Elasticsearch from upstream's RPM repositories.
• Upgrades Elastcisearch from version 5.x (if necessary).
• Configures Elasticsearch according to the official Elasticsearch Reference for version 6.0
• Configures system's performance settings with Tuned if it is enabled (default for RHEL/CentOS 7).
• Validates important Elasticsearch system and JVM configuration.

Requirements

This role requires Ansible 2.4 or higher.

Role Variables

Variables in defaults/main.yml:

Name	Type	Description	Default
elasticsearch_data_dir	string	Path of Elasticsearch data directory. For more information, see: Configuring Elasticsearch.	/var/lib/elasticsearch
elasticsearch_log_dir	string	Path of Elasticsearch log directory. For more information, see: Configuring Elasticsearch.	/var/log/elasticsearch
elasticsearch_version	string	Major version to install. The actual version that is installed is the latest minor (and patch) version of the chosen major version. It can be one of the following: "6"	"6"
elasticsearch_sysctl_file	string	Path of Elasticsearch's sysctl configuration file. NOTE: It should be in one of the locations that are searched for by the systemd-sysctl.service. For more information, see sysctl.d's man page.	/etc/sysctl.d/00-elasticsearch.conf
elasticsearch_heap_size_mem_frac	float	Fraction of total system memory to use for JVM's heap. NOTE: Before changing this value, read the documentation on setting heap size. Regardless of the chosen fraction value, the minimal and maximal heap sizes are 256 MB and 31 GB (respectively).	0.5
elasticsearch_cluster_name	string	Name of Elasticsearch cluster. For more information, see: cluster.name documentation. NOTE: A node can only join a cluster when it shares the same cluster name with all other nodes in the cluster.	elasticsearch
elasticsearch_node_name	string	Name of Elasticsearch node. For more information, see: node.name documentation.	"{{ ansible_nodename }}"
elasticsearch_network_host	string	Name of host or IP address to bind to. For more information, see: network settings documentation. NOTE: One can't use special values, e.g. _local_, _site_, since the variable is also used by Ansible which doesn't understand Elasticsearch's special syntax.	localhost
elasticsearch_http_port	integer	Port to bind to. For more information, see: network settings documentation.	9200
elasticsearch_discovery_zen_hosts	list	List of nodes to contact initially to join a cluster. One can specify host names or IP addresses. For more information, see: network settings documentation.	["127.0.0.1", "[::1]"]
elasticsearch_only_one_node_per_system	boolean	Limit Elasticsearch to one node per system. For more information, see: node.max_local_storage_nodes documentation.	true
elasticsearch_logging_to_file	boolean	Save main logging events to a file. system. NOTE: Main logging events are always stored via systemd-journald.service.	false

Variables in distribution-specific files under vars/:

Name	Type	Description	Default
elasticsearch_java_package	string	Name of the package that provides a JVM. NOTE: Elasticsearch 6.0 requires Java 1.8. For more information, see: Product and JVM matrix.	java-1.8.0-openjdk-headless

Dependencies

None.

Example Playbook

- hosts: all

  vars:
    elasticsearch_data_dir: /mnt/fast_storage/elasticsearch

  roles:
    - ansible-elasticsearch-role

License

GPLv3

Author Information

Tadej Janež

Domen Blenkuš