Step-by-Step Guide to Installing and Configuring an Elasticsearch Cluster on CentOS

Introduction: In the era of information explosion, Elasticsearch (ES) provides a high‑performance, scalable search engine for enterprise applications. This guide walks developers through installing and configuring an ES service cluster from scratch on CentOS 7.

Prerequisites: OS CentOS 7‑2003, JDK 15.x, Elastic Stack 7.11.x, VMware 15.5.x.

Download sources: official, community, Huawei mirrors, or Docker images.

System preparation: disable firewalld (service firewalld stop; chkconfig firewalld off), disable SELinux (edit /etc/selinux/config, set SELINUX=disabled), set hostname, turn off swap (swapoff -a, vm.swappiness=1), adjust vm.max_map_count, configure file‑handle limits in /etc/security/limits.conf, create dedicated ES user (useradd elk; chown -R elk:elk /elk/*).

JDK configuration (optional) via /etc/profile with JAVA_HOME, JRE_HOME, PATH, CLASSPATH.

Set ES temporary directories: ES_TMPDIR and JNA_TMPDIR in environment or JVM.options.

Single‑node ES configuration: edit {ES_HOME}/config/elasticsearch.yml with cluster.name, node.name, network.host, http.port, transport.port, data and log paths, bootstrap.memory_lock, action.destructive_requires_name, node.processors, discovery.type=single-node, xpack.ml.enabled=false. Adjust JVM options in jvm.options (‑Xms1g, ‑Xmx1g, GC settings, log path).

# Example elasticsearch.yml
cluster.name: elk01
node.name: ${HOSTNAME}-9200
network.host: 192.168.86.106
http.port: 9200
transport.port: 9300
path.data: /elk/9200/data
path.logs: /elk/9200/logs
bootstrap.memory_lock: true
action.destructive_requires_name: true
node.processors: 4
discovery.type: single-node
discovery.seed_hosts: ["192.168.86.106:9300"]
cluster.initial_master_nodes: ["192.168.86.106:9300"]
xpack.ml.enabled: false

Kibana configuration: edit {KIBANA_HOME}/config/kibana.yml with server.port, server.host, elasticsearch.hosts, kibana.index, xpack.ml.enabled=false.

# Example kibana.yml
server.port: 5601
server.host: "192.168.86.106"
elasticsearch.hosts: ["http://192.168.86.106:9200"]
kibana.index: ".kibana"
xpack.ml.enabled: false

Cluster setup: At least three nodes (e.g., gpes06, gpes07, gpes08) with IP 192.168.86.106, each with its own ES and Kibana directories. Use the same cluster.name (elk02) and configure node.name, network.host, http/transport ports, discovery.seed_hosts, and cluster.initial_master_nodes accordingly.

Docker deployment: Verify Docker and docker‑compose versions, pull ES and Kibana images (docker.elastic.co/elasticsearch/elasticsearch:7.11.1, docker.elastic.co/kibana/kibana:7.11.1). Run single‑node container:

docker run -p 9200:9200 -p 9300:9300 --name es9200 -e "discovery.type=single-node" docker.elastic.co/elasticsearch/elasticsearch:7.11.1
docker start es9200

For cluster mode, mount custom elasticsearch.yml, use host network, and start containers accordingly.

Similarly, launch Kibana containers with appropriate port mappings and configuration files.