Elasticsearch是一个分布式搜索服务，提供Restful API，底层基于Lucene，采用多shard的方式保证数据安全，并且提供自动resharding的功能，github等大型的站点也都采用Elasticsearch作为其搜索服务。废话在此就不多赘述了，下面记录下CentOS7下Elasticsearch集群部署过程：
代码语言：javascript复制之前在三台服务器部署Elasticsearch集群：
qd-vpc-op-es01 101.119.92.247
qd-vpc-op-es02 101.119.92.249
qd-vpc-op-es03 101.119.92.254
下面再追加一台es04机器到集群中，操作如下：
qd-vpc-op-es04 101.119.92.16
1）安装jdk和elasticsearch
jdk-8u5-linux-x64.rpm下载地址：
https://pan.baidu.com/s/1bpxtX5X (提取密码：df6s)
elasticsearch-5.5.0.rpm下载地址：
https://pan.baidu.com/s/1mibwWeG （提取密码：vtm2）
[root@qd-vpc-op-es04 ~]# cd tools/
[root@qd-vpc-op-es04 tools]# ls
elasticsearch-5.5.0.rpm jdk-8u5-linux-x64.rpm
[root@qd-vpc-op-es04 tools]# rpm -ivh elasticsearch-5.5.0.rpm
[root@qd-vpc-op-es04 tools]# rpm -ivh jdk-8u5-linux-x64.rpm
[root@qd-vpc-op-es04 tools]# java -version
java version "1.8.0_05"
Java(TM) SE Runtime Environment (build 1.8.0_05-b13)
Java HotSpot(TM) 64-Bit Server VM (build 25.5-b02, mixed mode)
2）配置elasticsearch
[root@qd-vpc-op-es04 ~]# cd /etc/elasticsearch/
[root@qd-vpc-op-es04 elasticsearch]# ls
elasticsearch.yml elasticsearch.yml.bak jvm.options log4j2.properties nohup.out scripts
[root@qd-vpc-op-es04 elasticsearch]# cat elasticsearch.yml|grep -v "#" //集群中每个节点的配置内容都基本一致
cluster.name: image_search //集群名称
node.name: image_search_node_4 //集群中本节点的节点名称，这里定义即可
path.data: /data/es/data //服务目录路径
path.logs: /data/es/logs //服务日志路径
discovery.zen.ping.unicast.hosts: ["101.119.92.247","101.119.92.249","101.119.92.254","10.111.233.16"] //这里是各节点的ip地址
network.host: 0.0.0.0 //服务绑定的网络地址
默认elasticsearch服务端口时9200
[root@qd-vpc-op-es04 elasticsearch]# cat elasticsearch.yml|grep 9200
#http.port: 9200
[root@qd-vpc-op-es04 elasticsearch]# systemctl start elasticsearch.service
[root@qd-vpc-op-es04 elasticsearch]# systemctl restart elasticsearch.service
[root@qd-vpc-op-es04 elasticsearch]# systemctl status elasticsearch.service
[root@qd-vpc-op-es04 elasticsearch]# ps -ef|grep elasticsearch
[root@qd-vpc-op-es04 elasticsearch]# lsof -i:9200
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
java 10998 elasticsearch 320u IPv4 39255 0t0 TCP *:wap-wsp (LISTEN)
检查elasticsearch的健康状态
[root@qd-vpc-op-es04 elasticsearch]# curl 'localhost:9200/_cat/indices?v'
health status index uuid pri rep docs.count docs.deleted store.size pri.store.size
green open video_filter Bx7He6ZtTEWuRBqXYC6gRw 5 1 458013 0 4.1gb 2gb
green open recommend_history_image svYo_Do4SM6wUiv6taUWug 5 1 2865902 0 24.9gb 12.4gb
green open recommend_history_gif rhN3MDN2TbuYILqEDksQSg 5 1 265731 0 2.4gb 1.2gb
green open post_images TMsMsMEoR5Sdb7UEQJsR5Q 5 1 48724932 0 407.3gb 203.9gb
green open review_images_v2 qzqnknpgTniU4rCsvXzs0w 5 1 50375955 0 61.6gb 30.9gb
green open review_images rWC4WlfMS8aGe-GOkTauZg 5 1 51810877 0 439.3gb 219.7gb
green open sensitive_images KxSrjvXdSz-y8YcqwBMsZA 5 1 13393 0 128.1mb 64mb
green open post_images_v2 FDphBV4-QuKVoD4_G3vRtA 5 1 49340491 0 55.8gb 27.8gb
从上面的命令结果中可以看出，本节点已经成功加入到名为image_search的elasticsearch集群中了，green表示节点状态很健康，数据也已经在同步中了。
3）在代码中更新elasticsearch的配置
通知开发同事，在代码中增加新增elasticsearch节点的配置，上线更新后，到新节点上查看elasticsearch日志是否有信息写入：
[root@qd-vpc-op-es04 ~]# cd /data/es/logs/
[root@qd-vpc-op-es04 ~]# chown -R elasticsearch.elasticsearch /data/es
[root@qd-vpc-op-es04 logs]# ls
image_search_deprecation.log image_search_index_indexing_slowlog.log image_search_index_search_slowlog.log image_search.log
----------------------------------------------------------------------------------------------------------------
注意：
如果往elasticsearch集群中新增一个节点，做法如下：
1）在新节点上安装jdk和elasticsearch服务，配置elasticsearch.yml文件了，启动elasticsearch服务
2）在集群中其他节点上配置elasticsearch.yml文件，不需要启动elasticsearch服务
3）在新节点上执行curl 'localhost:9200/_cat/indices?v'命令，查看健康状态以及数据同步情况
4）在代码中增加新增elasticsearch节点的配置，上线更新后，查看新增节点的elasticsearch日志是否有信息写入
----------------------------------------------------------------------------------------------------------------
最后顺便贴一下其他三个节点的elasticsearch.yml文件配置：
[root@qd-vpc-op-es01 ~]# cat /etc/elasticsearch/elasticsearch.yml |grep -v "#"
cluster.name: image_search
node.name: image_search_node_1
path.data: /data/es/data
path.logs: /data/es/logs
discovery.zen.ping.unicast.hosts: ["10.111.233.16","101.119.92.247","101.119.92.249","101.119.92.254"]
network.host: 0.0.0.0
[root@qd-vpc-op-es02 ~]# cat /etc/elasticsearch/elasticsearch.yml |grep -v "#"
cluster.name: image_search
node.name: image_search_node_2
path.data: /data/es/data
path.logs: /data/es/logs
discovery.zen.ping.unicast.hosts: ["10.111.233.16","101.119.92.247","101.119.92.249","101.119.92.254"]
network.host: 0.0.0.0
[root@qd-vpc-op-es03 ~]# cat /etc/elasticsearch/elasticsearch.yml|grep -v "#"
cluster.name: image_search
node.name: image_search_node_3
path.data: /data/es/data
path.logs: /data/es/logs
discovery.zen.ping.unicast.hosts: ["101.119.92.247","101.119.92.249","101.119.92.254","10.111.233.16"]
network.host: 0.0.0.0
----------------------------------------------------------------------------------------------------------------
如果关闭某个节点或将其从集群中提出来，需要提前通知该节点停止同步数据（迁移的时候会用到）：
# curl -XPUT 'localhost:9200/_cluster/settings' -d '{"transient" : {"cluster.routing.allocation.enable" : "none”}}’ 
