Graylog Data Node Install

Overview

Installs and configures the Graylog data node (OpenSearch) on a dedicated host. Handles the full disk setup (partition, XFS format, mount) as well as package installation and configuration. All disk operations are idempotent — an already-formatted disk will not be reformatted, preserving existing data on rebuilds.

What This Role Does

1. Partitions and formats the data disk with XFS (skipped if already formatted — existing data is preserved)
2. Mounts the disk at the configured data path and adds it to fstab
3. Adds the Graylog repository and installs graylog-datanode
4. Configures datanode.conf with the shared secret, MongoDB URI, data path, and heap size (auto-calculated from available RAM)
5. Sets vm.max_map_count=262144 permanently via sysctl (required by OpenSearch)
6. Enables and starts the graylog-datanode service

Role Variables

Variable	Description
vault_graylog_password_secret	Shared secret — must be identical to the main Graylog node (from vault)
graylog_datanode_install_version	Graylog version
graylog_datanode_install_disk_device	Disk device for OpenSearch data
graylog_datanode_install_opensearch_data_location	Mount point and OpenSearch data path
graylog_datanode_install_opensearch_heap	OpenSearch heap size
graylog_datanode_install_mongodb_uri	MongoDB connection URI

Notes

• graylog-datanode is automatically restarted if any datanode.conf setting changes
• vm.max_map_count=262144 is set permanently — this is a hard requirement for OpenSearch
• The data node must be installed and running before the main Graylog node
• After a MongoDB restore on the main Graylog node, the data node reconnects automatically
• OpenSearch heap is auto-calculated from available RAM (takes half of it by default) and can be overridden per host