Elasticsearch Single Machine Cloud Server Deployment & Op...

TL;DR

• Scenario: Single machine deployment of Elasticsearch for development and testing
• Conclusion: Single-node mode meets small company business scenarios, simple config easy to get started
• Output: Complete deployment process, JVM config, system parameters

---

ES Single Machine Deployment Overview

Single-Node Mode

Elasticsearch is a distributed full-text search engine, supports:

• Single-node mode (Single-Node Mode)
• Cluster mode (Cluster Mode)

Small company business scenarios can use Single-Node Mode.

---

Official Website

https://www.elastic.co/cn/downloads/elasticsearch

---

Download and Extract

Download

wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-7.3.0-linux-x86_64.tar.gz

Extract

cd /opt/software/
tar -zxvf elasticsearch-7.3.0-linux-x86_64.tar.gz
mv elasticsearch-7.3.0 ../servers/

---

Config Modification

Service Config

Modify elasticsearch.yml:

vim /opt/servers/elasticsearch-7.3.0/config/elasticsearch.yml

• For single machine installation, uncomment: node.name: node-1
• Modify network and port

# ------------------------------------ Node ------------------------------------
node.name: node-1

JVM Config

Modify jvm.options:

vim /opt/servers/elasticsearch-7.3.0/config/jvm.options

Note: Default 1G memory not enough, recommend setting 2G:

-Xms2g
-Xmx2g

---

User and Permissions

Create User

ES doesn’t allow root user to start by default:

useradd es_server
passwd es_server

Grant Permissions

chown -R es_server /opt/servers/elasticsearch-7.3.0

---

System Parameter Config

Modify sysctl.conf

ES needs to create many index files, need to adjust kernel parameters:

vim /etc/sysctl.conf

Add at end:

vm.max_map_count=655360

Apply:

sysctl -p

Modify limits.conf

Modify file descriptor limits:

vim /etc/security/limits.conf

Add at end:

* soft nofile 65536
* hard nofile 65536
* soft nproc 4096
* hard nproc 4096

---

Start Service

Switch User

su es_server

Start

/opt/servers/elasticsearch-7.3.0/bin/elasticsearch

Start in Background

/opt/servers/elasticsearch-7.3.0/bin/elasticsearch -d

---

Access Service

Verify

http://h121.wzk.icu:9200

Return Result

{
  "name": "node-1",
  "cluster_name": "elasticsearch",
  "cluster_uuid": "zv9WF53-RfKO1xNaIxt3uA",
  "version": {
    "number": "8.15.0",
    "build_flavor": "default",
    "build_type": "tar",
    "build_hash": "1a77947f34deddb41af25e6f0ddb8e830159c179",
    "build_date": "2024-08-05T10:05:34.233336849Z",
    "build_snapshot": false,
    "lucene_version": "9.11.1",
    "minimum_wire_compatibility_version": "7.17.0",
    "minimum_index_compatibility_version": "7.0.0"
  },
  "tagline": "You Know, for Search"
}

---

Deployment Best Practices

Cluster Planning

• Set different roles like master node, data node, coordinating node
• Set reasonable shard and replica counts to ensure performance and fault tolerance

Security

• Use X-Pack for user authentication and role management
• Configure TLS/SSL encrypted communication to ensure data transmission security

Performance Optimization

• Reasonably set index.refresh_interval to reduce overhead
• Use index.max_result_window to adjust query result count limit
• Enable shard allocation awareness to ensure shards distribute on different physical nodes

---

Error Quick Reference

Symptom	Root Cause	Fix
Memory insufficient	JVM heap set too large/physical memory insufficient	Reduce -Xms2g -Xmx2g
File descriptor too low	limits.conf not applied	Restart or re-login
max_map_count too low	Kernel parameter not set	sysctl -p to apply
Root run denied	ES doesn’t allow root to start	Create normal user
Port occupied	9200/9300 port occupied	Check port or modify config