声明：本示例主要作为测试用，生产请慎重。

最近公司突发奇想，想让我们搞个单机多实例的 es 的集群，看看其性能咋样。通常来说，es 作为搜索引擎，应用场景不乏日志分析、网络安全、搜索引擎等，有时也会用作日志数据库使用，毕竟其出色的搜索查询性能，不是同等量级关系型数据库可以比拟的，主要还是因为其倒排索引的特殊性，这里不讨论倒排索引与 B+ Tree 的性能，我们主要看看这种集群怎么组建的。

环境

环境准备：

ubuntu，24核，64G
docker 20.10.2

因为是 es 集群，我们准备通过 docker 来创建实例，所以之前你还得先 pull es、kibana 的 image:

docker pull elasticsearch:6.8.23
docker pull kibana:6.8.23

相关配置

如果你的容器有限，可以直接通过脚本运行 docker run，但是如果容器数量多还有相关依赖，建议通过容器编排起容器，当然数量更大的情况下，建议通过 k8s 部署。

我们的集群主要包含 3 个 es 节点，外加一个 kibana 作为观测，所以通过 docker-compose 作为容器编排，相对合适。

接下来是我们的编排定义： docker-compose.yml

yaml 复制代码

version: "2.3"
services:
  es-0:
    image: elasticsearch:6.8.23
    hostname: es-0
    container_name: es-0
    environment:
      - bootstrap.memory_lock=true
    ulimits:
      memlock:
        soft: -1
        hard: -1
    cap_add:
      - IPC_LOCK
    volumes:
      - /var/xxx/es_cluster/es-0:/usr/share/elasticsearch/data # 容器数据映射
      - ./es_cluster/es-0/elasticsearch:/etc/default/elasticsearch # elasticsearch 文件映射
      - ./es_cluster/es-0/config:/usr/share/elasticsearch/config # 配置映射，主要是 elasticsearch.yaml 和 jvm.options
      - /var/log/es_cluster/es-0/logs:/usr/share/elasticsearch/logs # 日志映射
      - /usr/share/zoneinfo/Asia/Shanghai:/etc/localtime # 时间
      - /etc/timezone:/etc/timezone
      #- ./elasticsearch/jvm.options:/etc/elasticsearch/jvm.options
    ports:
      - "9200:9200" # 端口映射
    command: elasticsearch
    logging:
      options:
        max-size: "200M"
        max-file: "5"
    networks:
      app_net:
        ipv4_address: 172.238.238.219
    healthcheck:
        test: ["CMD", "curl", "-f", "-s", "http://172.238.238.219:9200/_cluster/health?wait_for_status=yellow&timeout=50s"]
        interval: 30s
        timeout: 10s
        retries: 20
    restart: always

  es-1:
    image: elasticsearch:6.8.23
    hostname: es-1
    container_name: es-1
    environment:
      - bootstrap.memory_lock=true
    ulimits:
      memlock:
        soft: -1
        hard: -1
    cap_add:
      - IPC_LOCK
    volumes:
      - /var/xxx/es_cluster/es-1:/usr/share/elasticsearch/data
      - ./es_cluster/es-1/elasticsearch:/etc/default/elasticsearch
      - ./es_cluster/es-1/config:/usr/share/elasticsearch/config
      - /var/log/es_cluster/es-1/logs:/usr/share/elasticsearch/logs
      - /usr/share/zoneinfo/Asia/Shanghai:/etc/localtime
      - /etc/timezone:/etc/timezone
      #- ./elasticsearch/jvm.options:/etc/elasticsearch/jvm.options
    ports:
      - "9201:9201"
    command: elasticsearch
    logging:
      options:
        max-size: "200M"
        max-file: "5"
    networks:
      app_net:
        ipv4_address: 172.238.238.229
    healthcheck:
      test: ["CMD", "curl", "-f", "-s", "http://172.238.238.229:9200/_cluster/health?wait_for_status=yellow&timeout=50s"]
      interval: 30s
      timeout: 10s
      retries: 20
    restart: always

  es-2:
    image: elasticsearch:6.8.23
    hostname: es-2
    container_name: es-2
    environment:
      - bootstrap.memory_lock=true
    ulimits:
      memlock:
        soft: -1
        hard: -1
    cap_add:
      - IPC_LOCK
    volumes:
      - /var/xxx/es_cluster/es-2:/usr/share/elasticsearch/data
      - ./es_cluster/es-2/elasticsearch:/etc/default/elasticsearch
      - ./es_cluster/es-2/config:/usr/share/elasticsearch/config
      - /var/log/es_cluster/es-2/logs:/usr/share/elasticsearch/logs
      - /usr/share/zoneinfo/Asia/Shanghai:/etc/localtime
      - /etc/timezone:/etc/timezone
      #- ./elasticsearch/jvm.options:/etc/elasticsearch/jvm.options
    ports:
      - "9202:9202"
    command: elasticsearch
    logging:
      options:
        max-size: "200M"
        max-file: "5"
    networks:
      app_net:
        ipv4_address: 172.238.238.239
    healthcheck:
      test: ["CMD", "curl", "-f", "-s", "http://172.238.238.239:9200/_cluster/health?wait_for_status=yellow&timeout=50s"]
      interval: 30s
      timeout: 10s
      retries: 20
    restart: always

  kibana:
    image: kibana:6.8.23
    hostname: kibana
    container_name: kibana
    volumes:
      - ./kibana/kibana.yml:/usr/share/kibana/config/kibana.yml
      - /usr/share/zoneinfo/Asia/Shanghai:/etc/localtime
      - /etc/timezone:/etc/timezone
      - ./kibana/kibana.keystore:/usr/share/kibana/data/kibana.keystore
    ports:
      - "5601:5601"
    networks:
      app_net:
        ipv4_address: 172.238.238.242
    restart: always
    logging:
      options:
        max-size: "200M"
        max-file: "5"

networks:
  app_net:
    driver: bridge
    ipam:
      driver: default
      config:
      - subnet: 172.238.238.0/24

这里限定 docker 的网络网段。

然后我们要看看对应的其他准备。

主要看我们的对应到主机中的 data 目录，所以参考 yml 中的相关映射，注意创建相关目录。

这里我们主要看看相关的 elasticsearch.yaml 和 jvm.options。

elasticsearch.yml

yaml 复制代码

cluster:
  name: logserver
#################
node.name: es-0 # 其他节点类似，修改 node name
discovery.zen.ping.unicast.hosts: ["es-0", "es-1", "es-2"]
network.host: 0.0.0.0
discovery.zen.minimum_master_nodes: 2
gateway.recover_after_nodes: 3
http.port: 9200
transport.tcp.port: 9300
node.master: true
node.data: true

http.host: 0.0.0.0

http:
  enabled: true
  compression: true
  cors:
    enabled: true
    allow-origin: "*"

bootstrap.memory_lock: true
bootstrap.system_call_filter: false
path.data: /usr/share/elasticsearch/data

#cluster.routing.allocation.disk.threshold_enabled: true
#cluster.routing.allocation.disk.watermark.flood_stage: 80gb
#cluster.routing.allocation.disk.watermark.high: 100gb
#cluster.routing.allocation.disk.watermark.low: 120gb


##Lock memory, do not write swap
#bootstrap.mlockall: true
##The cache type is set to Soft Reference,
##and is reclaimed only if there is not enough memory
#index.cache.field.max_size: 50000
#index.cache.field.expire: 10m
#index.cache.field.type: soft
##Weighing the performance of the index and the timeliness of retrieval
#index.translog.flush_threshold_ops: 10000
#index.refresh_interval: 1
#number_of_replicas: 0

#indices.lifecycle.poll_interval: 5m

xpack.ml.enabled: false

jvm.options

shell 复制代码

...
# Xms represents the initial size of total heap space
# Xmx represents the maximum size of total heap space

-Xms16g # 主机内存64G，每个实例分配16G
-Xmx16g

################################################################
## Expert settings
################################################################
##
...

这两个文件按照上面内容修改。

接下来是 kibana.yml:

erlang 复制代码

...
elasticsearch.url: "http://172.xxx.xxx.xxx:9200" # 根据实际情况，填入自己的主机ip
...

接下来通过 docker-compose 命令就可以起容器了。

bash 复制代码

docker-compsoe up -d # 后台运行容器

docker-compose ps # 查看容器运行状态

docker-compose down # 停掉容器

结果查看

接下来看看容器状态：

bash 复制代码

# docker-compose ps
 Name               Command                  State                         Ports
---------------------------------------------------------------------------------------------------
es-0     /usr/local/bin/docker-entr ...   Up (healthy)   0.0.0.0:9200->9200/tcp, 9300/tcp
es-1     /usr/local/bin/docker-entr ...   Up (healthy)   9200/tcp, 0.0.0.0:9201->9201/tcp, 9300/tcp
es-2     /usr/local/bin/docker-entr ...   Up (healthy)   9200/tcp, 0.0.0.0:9202->9202/tcp, 9300/tcp
kibana   /usr/local/bin/kibana-docker     Up             0.0.0.0:5601->5601/tcp

通过 kibana 查看容器状态：

其他看板：

可以看到，es 集群已经顺利起来了，集群实例就演示到这里，希望对你有用。

elasticsearch 组件基于单机的多实例集群

环境

相关配置

结果查看