debops.elasticsearch default variables

Sections

APT packages, version
UNIX user and group
Ansible inventory layout
Firewall configuration
Elasticsearch network options
Elasticsearch cluster options
Node functions
Memory options
Paths
Elasticsearch configuration file
Plugin configuration
Configuration for other Ansible roles

APT packages, version 

The debops.elasticsearch role uses the debops.elatic_co Ansible role to configure the Elastic APT repositories and install the packages. The role also installs the Ansible facts that provide the elasticsearch version.

elasticsearch__base_packages

List of base APT packages to install.

elasticsearch__base_packages: [ 'elasticsearch' ]

elasticsearch__packages

List of additional APT packages to install with Elasticsearch.

elasticsearch__packages: []

elasticsearch__version

Store the detected Elasticsearch version in a convenient variable for conditional configuration.

elasticsearch__version: '{{ ansible_local.elasticsearch.version | d("0.0.0") }}'

UNIX user and group 

elasticsearch__user

Name of the UNIX user account used by Elasticsearch.

elasticsearch__user: 'elasticsearch'

elasticsearch__group

Name of the UNIX primary group used by Elasticsearch.

elasticsearch__group: 'elasticsearch'

Ansible inventory layout 

elasticsearch__inventory_group_all

Name of the Ansible inventory group which contains Elasticsearch host in a "homogeneous" configuration (each node having the same functions as every other node).

elasticsearch__inventory_group_all: 'debops_service_elasticsearch'

elasticsearch__inventory_group_master

Name of the Ansible inventory group which contains Elasticsearch master nodes.

elasticsearch__inventory_group_master: 'debops_service_elasticsearch_master'

elasticsearch__inventory_group_data

Name of the Ansible inventory group which contains Elasticsearch data nodes.

elasticsearch__inventory_group_data: 'debops_service_elasticsearch_data'

elasticsearch__inventory_group_ingest

Name of the Ansible inventory group which contains Elasticsearch ingest nodes.

elasticsearch__inventory_group_ingest: 'debops_service_elasticsearch_ingest'

elasticsearch__inventory_group_lb

Name of the Ansible inventory group which contains Elasticsearch load balancer nodes.

elasticsearch__inventory_group_lb: 'debops_service_elasticsearch_lb'

elasticsearch__inventory_master_hosts

List of Ansible inventory hosts which should be treated as Elasticsearch master nodes. See Elasticsearch clustering for more details.

elasticsearch__inventory_master_hosts: '{{ (groups[elasticsearch__inventory_group_master]
                                            | d(groups[elasticsearch__inventory_group_all]))
                                           if elasticsearch__allow_tcp else [] }}'

elasticsearch__initial_master_nodes

List of Elasticsearch nodes which will be bootstrapped as masters on the first cluster execution.

elasticsearch__initial_master_nodes: [ '{{ elasticsearch__node_name }}' ]

Firewall configuration 

elasticsearch__allow_http

List of IP addresses or CIDR subnets that can connect to the Elasticsearch HTTP service. This does not need to be set to allow the nodes to communicate. If this list is empty, nobody can connect to the HTTP server directly.

elasticsearch__allow_http: []

elasticsearch__allow_tcp

List of IP addresses or CIDR subnets that can connect to the Elasticsearch TCP transport port. This variable needs to be set to allow nodes to communicate. If this list is empty, nobody can connect to the transport port and the Elasticsearch service is configured in a standalone mode.

elasticsearch__allow_tcp: []

Elasticsearch network options 

elasticsearch__network_host

List of network interface names or IP addresses on which Elasticsearch should listen for connections. By default, if debops.ferm firewall is enabled, Elasticsearch will listen on localhost and private IP addresses only. See https://www.elastic.co/guide/en/elasticsearch/reference/current/modules-network.html for more details.

elasticsearch__network_host: '{{ [ "_local_", "_site_" ]
                                 if (ansible_local.ferm.enabled|d() and
                                     ansible_local.ferm.enabled|bool)
                                 else [ "_local_" ] }}'

elasticsearch__http_port

The port on which Elasticsearch will listen for HTTP connections.

elasticsearch__http_port: '9200'

elasticsearch__transport_tcp_port

The port on which Elasticsearch will listen for TCP transport connections.

elasticsearch__transport_tcp_port: '9300'

Elasticsearch cluster options 

elasticsearch__domain

The DNS domain of the host.

elasticsearch__domain: '{{ ansible_domain }}'

elasticsearch__cluster_name

Name of the Elasticsearch cluster, should be the same on all of the cluster nodes. By default it's based on the host domain name.

elasticsearch__cluster_name: '{{ elasticsearch__domain | replace(".","-") }}'

elasticsearch__node_name

Descriptive name of the Elasticsearch node, by default based on the hostname.

elasticsearch__node_name: '{{ ansible_hostname }}'

elasticsearch__discovery_hosts

List of Elasticsearch nodes which should be contacted for unicast cluster discovery. See Elasticsearch clustering for more details.

elasticsearch__discovery_hosts: '{{ elasticsearch__inventory_master_hosts }}'

elasticsearch__discovery_minimum_master_nodes

Minimum number of master-eligible nodes that are required to achieve quorum.

elasticsearch__discovery_minimum_master_nodes: '{{ "1" if (elasticsearch__inventory_master_hosts | count <= 2)
                                                       else ((elasticsearch__inventory_master_hosts | count / 2) | round(0, "floor") | int + 1) }}'

elasticsearch__gateway_recover_after_nodes

Number of Elasticsearch nodes required to enable cluster initial recovery after full restart.

elasticsearch__gateway_recover_after_nodes: '{{ elasticsearch__discovery_minimum_master_nodes }}'

Node functions 

These variables define the functionality of a given Elasticsearch node. See https://www.elastic.co/guide/en/elasticsearch/reference/current/modules-node.html for more details.

elasticsearch__node_master

When enabled, this node is a master-eligible node and can be elected as a master.

Warning: this variable is not taken into account for master node calculations. See Elasticsearch clustering for more details.

elasticsearch__node_master: '{{ True
                                if (elasticsearch__inventory_group_master in group_names)
                                else (False
                                      if (elasticsearch__inventory_group_data in group_names)
                                      else (False
                                            if (elasticsearch__inventory_group_ingest in group_names)
                                            else (False
                                                  if (elasticsearch__inventory_group_lb in group_names)
                                                  else True))) }}'

elasticsearch__node_data

When enabled, this node can hold the cluster data and perform various operations like searching and aggregation.

elasticsearch__node_data: '{{ True
                              if (elasticsearch__inventory_group_data in group_names)
                              else (False
                                    if (elasticsearch__inventory_group_master in group_names)
                                    else (False
                                          if (elasticsearch__inventory_group_ingest in group_names)
                                          else (False
                                                if (elasticsearch__inventory_group_lb in group_names)
                                                else True))) }}'

elasticsearch__node_ingest

When enabled, this node can perform operations on documents before indexing them using an ingest pipeline.

elasticsearch__node_ingest: '{{ True
                                if (elasticsearch__inventory_group_ingest in group_names)
                                else (False
                                      if (elasticsearch__inventory_group_master in group_names)
                                      else (False
                                            if (elasticsearch__inventory_group_data in group_names)
                                            else (False
                                                  if (elasticsearch__inventory_group_lb in group_names)
                                                  else True))) }}'

Memory options 

The variables below configure JVM memory allocation options. See https://www.elastic.co/guide/en/elasticsearch/guide/current/heap-sizing.html for more details.

elasticsearch__memory_lock

Enable or disable memory lock depending on availability of required POSIX capabilities. If this variable is enabled, systemd memlock limit is configured.

elasticsearch__memory_lock: '{{ True
                                if (not (ansible_system_capabilities_enforced|d())|bool or
                                    ((ansible_system_capabilities_enforced|d())|bool and
                                     "cap_ipc_lock" in (ansible_system_capabilities|d([]))))
                                else False }}'

elasticsearch__systemd_limit_memlock

Specify maximum size of the process memory in bytes that the process is allowed to lock in RAM and not have it stored in swap. Specify infinity to disable the maximum limit. This setting is applied through the systemd service unit.

elasticsearch__systemd_limit_memlock: 'infinity'

elasticsearch__jvm_memory_heap_size_multiplier

This variable defines a float value which will be used to select the JVM heap size depending on the size of the available system RAM.

elasticsearch__jvm_memory_heap_size_multiplier: '{{ "0.2"
                                                    if (ansible_memtotal_mb|int / 2 <= 2048)
                                                    else "0.45" }}'

elasticsearch__jvm_memory_min_heap_size

Specify the minimum JVM heap size, depending on the available system RAM.

elasticsearch__jvm_memory_min_heap_size: '{{ (((ansible_memtotal_mb|int
                                               * elasticsearch__jvm_memory_heap_size_multiplier|float)
                                               | round | int) | string + "m")
                                             if (ansible_memtotal_mb|int / 2 <= 32768)
                                             else "32600m" }}'

elasticsearch__jvm_memory_max_heap_size

Specify the maximum JVM heap size, depending on the available system RAM. This usually should be the same as the minimum heap size, for performance reasons.

elasticsearch__jvm_memory_max_heap_size: '{{ elasticsearch__jvm_memory_min_heap_size }}'

Paths 

The variables below configure paths used by Elasticsearch. See https://www.elastic.co/guide/en/elasticsearch/reference/current/important-settings.html#path-settings for more details.

elasticsearch__path_data

List of all data paths. They will be created and permissions will be updated if needed.

elasticsearch__path_data:
  - '/var/lib/elasticsearch'

Elasticsearch configuration file 

The variables below define the contents of the /etc/elasticsearch/elasticsearch.yml configuration file. See elasticsearch__configuration for the details and configuration syntax.

elasticsearch__default_configuration

The default configuration options which should be present in the main configuration file.

elasticsearch__default_configuration:

  - name: 'cluster.name'
    value: '{{ elasticsearch__cluster_name }}'
    comment: 'Use a descriptive name for your cluster'

  - name: 'node.master'
    comment: 'Type of the node'
    value: '{{ elasticsearch__node_master }}'
    state: '{{ "present" if (elasticsearch__version is version("7.9.0", "<")) else "absent" }}'

  - name: 'node.data'
    value: '{{ elasticsearch__node_data }}'
    state: '{{ "present" if (elasticsearch__version is version("7.9.0", "<")) else "absent" }}'

  - name: 'node.ingest'
    value: '{{ elasticsearch__node_ingest }}'
    state: '{{ "present" if (elasticsearch__version is version("7.9.0", "<")) else "absent" }}'

  - name: 'node.name'
    value: '{{ elasticsearch__node_name }}'
    comment: 'Use a descriptive name for the node'

  - name: 'node.attr.rack'
    value: 'r1'
    comment: 'Add custom attributes to the node'
    state: 'comment'

  - name: 'network.host'
    value: '{{ elasticsearch__network_host }}'
    comment: 'Set the bind address to a specific IP (IPv4 or IPv6)'

  - name: 'http.port'
    value: '{{ elasticsearch__http_port }}'
    comment: 'Set a custom port for HTTP'

  - name: 'transport.tcp.port'
    value: '{{ elasticsearch__transport_tcp_port }}'
    comment: 'Set a custom port for TCP transport'

  - name: '{{ "discovery.zen.ping.unicast.hosts"
              if (elasticsearch__version is version("7.0.0", "<"))
              else "discovery.seed_hosts" }}'
    value: '{{ elasticsearch__discovery_hosts }}'
    state: '{{ "present" if elasticsearch__discovery_hosts else "absent" }}'
    comment: |
      Pass an initial list of hosts to perform discovery when new node is started:
      The default list of hosts is ["127.0.0.1", "[::1]"]

  - name: 'discovery.zen.minimum_master_nodes'
    value: '{{ elasticsearch__discovery_minimum_master_nodes }}'
    comment: |
      Prevent the "split brain" by configuring the majority of nodes
      (total number of master-eligible nodes / 2 + 1)

  - name: 'cluster.initial_master_nodes'
    value: '{{ elasticsearch__initial_master_nodes }}'
    state: '{{ "absent"
               if (elasticsearch__version is version("7.0.0", "<"))
               else "present" }}'
    comment: 'Bootstrap the cluster using an initial set of master-eligible nodes:'

  - name: 'gateway.recover_after_nodes'
    comment: 'Block initial recovery after a full cluster restart until N nodes are started'
    value: '{{ elasticsearch__gateway_recover_after_nodes }}'

  - name: 'action.destructive_requires_name'
    value: True
    comment: 'Require explicit names when deleting indices'

  - name: 'bootstrap.memory_lock'
    value: '{{ True if elasticsearch__memory_lock|bool else False }}'
    comment: |
      Lock the memory on startup

      Make sure that the heap size is set to about half the memory available
      on the system and that the owner of the process is allowed to use this
      limit.

      Elasticsearch performs poorly when the system is swapping the memory.

  - name: 'path.data'
    value: '{{ elasticsearch__path_data }}'
    comment: |
      Path to directory where to store the data
      (separate multiple locations by comma)

  - name: 'path.logs'
    value: '/var/log/elasticsearch'
    comment: 'Path to log files'

elasticsearch__configuration

List of configuration options defined on all hosts in the Ansible inventory.

elasticsearch__configuration: []

elasticsearch__master_configuration

List of configuration options defined on Elasticsearch master nodes (needs to be placed in appropriate Ansible inventory group).

elasticsearch__master_configuration: []

elasticsearch__data_configuration

List of configuration options defined on Elasticsearch data nodes (needs to be placed in appropriate Ansible inventory group).

elasticsearch__data_configuration: []

elasticsearch__ingest_configuration

List of configuration options defined on Elasticsearch ingest nodes (needs to be placed in appropriate Ansible inventory group).

elasticsearch__ingest_configuration: []

elasticsearch__lb_configuration

List of configuration options defined on Elasticsearch load balancer nodes (needs to be placed in appropriate Ansible inventory group).

elasticsearch__lb_configuration: []

elasticsearch__group_configuration

List of configuration options defined on hosts in specific Ansible inventory group.

elasticsearch__group_configuration: []

elasticsearch__host_configuration

List of configuration options defined on specific hosts in the Ansible inventory.

elasticsearch__host_configuration: []

elasticsearch__plugin_configuration

List of configuration options defined separately for any Elasticsearch plugins. See elasticsearch__plugins for more details.

elasticsearch__plugin_configuration: '{{ lookup("template",
                                         "lookup/elasticsearch__plugin_configuration.j2")
                                         | from_yaml }}'

elasticsearch__dependent_role

A string that identifies another Ansible role that uses the debops.elasticsearch role as a dependency. This value is needed to correctly store the dependent configuration options. See Usage as a role dependency for more details.

elasticsearch__dependent_role: ''

elasticsearch__dependent_state

Specify the state of the dependent configuration options, either present (options should be included in the configuration file) or absent (options should be removed from the configuration file). See Usage as a role dependency for more details.

elasticsearch__dependent_state: 'present'

elasticsearch__dependent_configuration

List of Elasticsearch configuration options defined by another Ansible role and specified using role dependent variables.

elasticsearch__dependent_configuration: []

elasticsearch__dependent_configuration_filter

Actual variable used in the combined Elasticsearch configuration that unwraps the dependent configuration specified by other Ansible roles and converts it into format understood by the debops.elasticsearch configuration template. See Usage as a role dependency for more details.

elasticsearch__dependent_configuration_filter: '{{ lookup("template",
                                                   "lookup/elasticsearch__dependent_configuration_filter.j2")
                                                   | from_yaml }}'

elasticsearch__combined_configuration

Actual list of Elasticsearch configuration options passed to the configuration template. This list defines the order in which the options from different variables are processed.

elasticsearch__combined_configuration: '{{ lookup("flattened", (elasticsearch__default_configuration
                                           + elasticsearch__plugin_configuration
                                           + elasticsearch__dependent_configuration_filter
                                           + elasticsearch__configuration
                                           + elasticsearch__master_configuration
                                           + elasticsearch__data_configuration
                                           + elasticsearch__ingest_configuration
                                           + elasticsearch__lb_configuration
                                           + elasticsearch__group_configuration
                                           + elasticsearch__host_configuration)) }}'

elasticsearch__configuration_sections

List of sections defined in the /etc/elasticsearch/elasticsearch.yml configuration file and corresponding variable groups. See elasticsearch__configuration_sections for more details.

elasticsearch__configuration_sections:

  - name: 'Cluster'
    part: 'cluster'

  - name: 'Node'
    part: 'node'

  - name: 'Paths'
    part: 'path'

  - name: 'Memory'
    part: 'bootstrap'

  - name: 'Network'
    parts: [ 'network', 'http', 'transport' ]

  - name: 'Discovery'
    part: 'discovery'

  - name: 'Gateway'
    part: 'gateway'

  - name: 'X-Pack'
    part: 'xpack'

  - name: 'Search Guard'
    part: 'searchguard'

  - name: 'ReadonlyREST'
    part: 'readonlyrest'

Plugin configuration 

These variables define lists of Elasticsearch plugins to install/remove, as well as additional configuration options for them. See elasticsearch__plugins for more details.

elasticsearch__plugins

List of Elasticsearch plugins to manage on all hosts in the Ansible inventory.

elasticsearch__plugins: []

elasticsearch__master_plugins

List of Elasticsearch plugins to manage on master nodes (variable needs to be defined in appropriate Ansible inventory group).

elasticsearch__master_plugins: []

elasticsearch__data_plugins

List of Elasticsearch plugins to manage on data nodes (variable needs to be defined in appropriate Ansible inventory group).

elasticsearch__data_plugins: []

elasticsearch__ingest_plugins

List of Elasticsearch plugins to manage on ingest nodes (variable needs to be defined in appropriate Ansible inventory group).

elasticsearch__ingest_plugins: []

elasticsearch__lb_plugins

List of Elasticsearch plugins to manage on load balancer nodes (variable needs to be defined in appropriate Ansible inventory group).

elasticsearch__lb_plugins: []

elasticsearch__group_plugins

List of Elasticsearch plugins to manage on hosts in specific Ansible inventory group.

elasticsearch__group_plugins: []

elasticsearch__host_plugins

List of Elasticsearch plugins to manage on specific hosts in the Ansible inventory.

elasticsearch__host_plugins: []

elasticsearch__combined_plugins

Actual list of Elasticsearch plugins that combines other plugin variables and is used in the Ansible tasks and the configuration template.

elasticsearch__combined_plugins: '{{ lookup("flattened", (elasticsearch__plugins
                                     + elasticsearch__master_plugins
                                     + elasticsearch__data_plugins
                                     + elasticsearch__ingest_plugins
                                     + elasticsearch__lb_plugins
                                     + elasticsearch__group_plugins
                                     + elasticsearch__host_plugins)) }}'

Configuration for other Ansible roles 

elasticsearch__etc_services__dependent_list

Configuration for the debops.etc_services Ansible role.

elasticsearch__etc_services__dependent_list:

  - name: 'elasticsearch-http'
    port: '{{ elasticsearch__http_port }}'

  - name: 'elasticsearch-tcp'
    port: '{{ elasticsearch__transport_tcp_port }}'

elasticsearch__sysctl__dependent_parameters

Configuration for the debops.sysctl Ansible role.

elasticsearch__sysctl__dependent_parameters:

  - name: 'elasticsearch'
    weight: 80
    options:

        # This parameter is set by default by the Elasticsearch .deb package,
        # but this breaks 'sysctl' usage in containers. The original file will
        # be diverted, and the role will configure the relevant setting for us,
        # automatically commenting it out inside of a container.
      - name: 'vm.max_map_count'
        comment: |
          Elasticsearch uses a mmapfs directory by default to store its
          indices. The default operating system limits on mmap counts is likely
          to be too low, which may result in out of memory exceptions.
          Ref: https://www.elastic.co/guide/en/elasticsearch/reference/current/vm-max-map-count.html
        value: 262144

elasticsearch__extrepo__dependent_sources

Configuration for the debops.extrepo Ansible role.

elasticsearch__extrepo__dependent_sources:
  - 'elastic'

elasticsearch__ferm__dependent_rules

Configuration for the debops.ferm Ansible role.

elasticsearch__ferm__dependent_rules:

  - name: 'elasticsearch_http'
    type: 'accept'
    dport: '{{ elasticsearch__http_port }}'
    saddr: '{{ elasticsearch__allow_http }}'
    accept_any: False

  - name: 'elasticsearch_tcp'
    type: 'accept'
    dport: '{{ elasticsearch__transport_tcp_port }}'
    saddr: '{{ elasticsearch__allow_tcp }}'
    accept_any: False