Elasticsearch Cluster 安装与配置

本文节选自《Netkiller Database 手札》作者:netkiller 网站: http://www.netkiller.cn

23.1.2. Elasticsearch Cluster

集群模式需要两个以上的节点,通常是一个 master 节点,多个 data 节点

首先在所有节点上安装 elasticsearch,然后配置各节点的配置文件,对于 5.5.1 不需要配置决定哪些节点属于 master 节点 或者 data 节点。

curl -s https://raw.githubusercontent.com/oscm/shell/master/search/elasticsearch/elasticsearch-5.x.sh | bash			

配置文件

cluster.name: elasticsearch-cluster # 配置集群名称,所有服务器服务器保持一致

node.name: node-1 # 每个节点唯一标识,每个节点只需改动这里,一次递增 node-1, node-2, node-3 ...

network.host: 0.0.0.0

discovery.zen.ping.unicast.hosts: ["172.16.0.20", "172.16.0.21","172.16.0.22"]  # 所有节点的IP 地址写在这里

discovery.zen.minimum_master_nodes: 3 # 可以作为master的节点总数,有多少个节点就写多少

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

查看节点状态,使用curl工具: curl 'http://localhost:9200/_nodes/process?pretty'

root@netkiller /var/log/elasticsearch % curl 'http://localhost:9200/_nodes/process?pretty'
{
  "_nodes" : {
    "total" : 2,
    "successful" : 2,
    "failed" : 0
  },
  "cluster_name" : "my-application",
  "nodes" : {
    "-lnKCmBXRpiwExLns0jc9g" : {
      "name" : "node-1",
      "transport_address" : "10.104.3.2:9300",
      "host" : "10.104.3.2",
      "ip" : "10.104.3.2",
      "version" : "5.5.1",
      "build_hash" : "19c13d0",
      "roles" : [
        "master",
        "data",
        "ingest"
      ],
      "process" : {
        "refresh_interval_in_millis" : 1000,
        "id" : 23669,
        "mlockall" : false
      }
    },
    "WVsgYi2HT8GWnZU1kUwFwA" : {
      "name" : "node-2",
      "transport_address" : "10.186.7.221:9300",
      "host" : "10.186.7.221",
      "ip" : "10.186.7.221",
      "version" : "5.5.1",
      "build_hash" : "19c13d0",
      "roles" : [
        "master",
        "data",
        "ingest"
      ],
      "process" : {
        "refresh_interval_in_millis" : 1000,
        "id" : 12641,
        "mlockall" : false
      }
    }
  }
}			

启动节点后回生成 cluster.name 为文件名的日志文件。

谁先启动谁讲成为master

[2017-08-11T17:42:46,018][INFO ][o.e.c.s.ClusterService   ] [node-1] new_master {node-1}{-lnKCmBXRpiwExLns0jc9g}{rZcJDIynSzq2Td3yP2kN5A}{10.104.3.2}{10.104.3.2:9300}, added {{node-2}{WVsgYi2HT8GWnZU1kUwFwA}{X13ShUpAQa2zA1Mgcsm3bQ}{10.186.7.221}{10.186.7.221:9300},}, reason: zen-disco-elected-as-master ([1] nodes joined)[{node-2}{WVsgYi2HT8GWnZU1kUwFwA}{X13ShUpAQa2zA1Mgcsm3bQ}{10.186.7.221}{10.186.7.221:9300}]			

如果master出现故障,其他节点会接管

[2017-08-11T17:44:52,797][INFO ][o.e.c.s.ClusterService   ] [node-2] master {new {node-2}{WVsgYi2HT8GWnZU1kUwFwA}{vl8kQx8sQdGVVohrNQnZOQ}{10.186.7.221}{10.186.7.221:9300}}, removed {{node-1}{-lnKCmBXRpiwExLns0jc9g}{rZcJDIynSzq2Td3yP2kN5A}{10.104.3.2}{10.104.3.2:9300},}, added {{node-1}{-lnKCmBXRpiwExLns0jc9g}{odnoG9kpQpeX1ltx5KYTSw}{10.104.3.2}{10.104.3.2:9300},}, reason: zen-disco-elected-as-master ([1] nodes joined)[{node-1}{-lnKCmBXRpiwExLns0jc9g}{odnoG9kpQpeX1ltx5KYTSw}{10.104.3.2}{10.104.3.2:9300}]
[2017-08-11T17:44:53,184][INFO ][o.e.c.r.DelayedAllocationService] [node-2] scheduling reroute for delayed shards in [59.5s] (11 delayed shards)
[2017-08-11T17:44:53,929][INFO ][o.e.c.r.a.AllocationService] [node-2] Cluster health status changed from [RED] to [YELLOW] (reason: [shards started [[information][0]] ...]).		

master 节点恢复上线会提示

[2017-08-11T17:44:52,855][INFO ][o.e.c.s.ClusterService   ] [node-1] detected_master {node-2}{WVsgYi2HT8GWnZU1kUwFwA}{vl8kQx8sQdGVVohrNQnZOQ}{10.186.7.221}{10.186.7.221:9300}, added {{node-2}{WVsgYi2HT8GWnZU1kUwFwA}{vl8kQx8sQdGVVohrNQnZOQ}{10.186.7.221}{10.186.7.221:9300},}, reason: zen-disco-receive(from master [master {node-2}{WVsgYi2HT8GWnZU1kUwFwA}{vl8kQx8sQdGVVohrNQnZOQ}{10.186.7.221}{10.186.7.221:9300} committed version [44]])			

23.1.3. 负载均衡配置

因为 elasticsearch 没有用户认证机制我们通常在内网访问他。如果对外提供服务需要增加用户认证。

			$ printf "john:$(openssl passwd -crypt s3cr3t)n" > /etc/nginx/passwords 			

创建 nginx 配置文件 /etc/nginx/conf.d/elasticsearch.conf

upstream elasticsearch {
	server 172.16.0.10:9200;
	server 172.16.0.20:9200;
	server 172.16.0.30:9200;

	keepalive 15;
}

server {
	listen 9200;
	auth_basic "Protected Elasticsearch";
	auth_basic_user_file passwords;

	location ~* ^(/_cluster|/_nodes) {
		return 403;
		break;
	} 

	location / {
    
		if ($request_filename ~ _shutdown) {
		    return 403;
		    break;
		}

		proxy_pass http://elasticsearch;
		proxy_http_version 1.1;
		proxy_set_header Connection "Keep-Alive";
		proxy_set_header Proxy-Connection "Keep-Alive";
	}

}		

反复使用下面方法请求,最终你会发现 total_opened 会达到你的nginx 配置数量

$ curl 'http://test:test@localhost:9200/_nodes/stats/http?pretty' | grep total_opened
# "total_opened" : 15			

23.2.1. elasticsearch-analysis-ik

安装插件

root@netkiller ~ % /usr/share/elasticsearch/bin/elasticsearch-plugin install https://github.com/medcl/elasticsearch-analysis-ik/releases/download/v5.5.1/elasticsearch-analysis-ik-5.5.1.zip
-> Downloading https://github.com/medcl/elasticsearch-analysis-ik/releases/download/v5.5.1/elasticsearch-analysis-ik-5.5.1.zip
[=================================================] 100%   
-> Installed analysis-ik			
curl -XPOST http://localhost:9200/index/fulltext/_mapping -d'
{
        "properties": {
            "content": {
                "type": "text",
                "analyzer": "ik_max_word",
                "search_analyzer": "ik_max_word"
            }
        }
    
}'			

23.3. 节点管理

23.3.1. 查看索引

root@netkiller ~ % curl 'http://localhost:9200/_cat/indices?v'         
health status index        uuid                   pri rep docs.count docs.deleted store.size pri.store.size
yellow open   information  oygxIi-dR1eB9NoIZtJrxQ   5   1         45           42      731kb          731kb
green  open   .kibana      9jBBaOomTO2EakZlZqnE-g   1   1          5            1     62.5kb         31.2kb
green  open   logstash-api WHXZhn3vRWiuVbhR8rGoEg   5   1        565            0      3.8mb          1.9mb			

23.3.2. 节点健康状态

root@netkiller ~ % curl 'http://localhost:9200/_cat/health?v' 
epoch      timestamp cluster        status node.total node.data shards pri relo init unassign pending_tasks max_task_wait_time active_shards_percent
1502445967 18:06:07  my-application yellow          2         2     17  11    0    0        5             0                  -                 77.3%			
root@netkiller ~ % curl 'http://localhost:9200/_cluster/health'  
{"cluster_name":"my-application","status":"yellow","timed_out":false,"number_of_nodes":2,"number_of_data_nodes":2,"active_primary_shards":11,"active_shards":17,"relocating_shards":0,"initializing_shards":0,"unassigned_shards":5,"delayed_unassigned_shards":0,"number_of_pending_tasks":0,"number_of_in_flight_fetch":0,"task_max_waiting_in_queue_millis":0,"active_shards_percent_as_number":77.27272727272727} 			

23.3.3. 节点http状态

root@VM_3_2_centos ~ % curl 'localhost:9200/_nodes/stats/http?pretty'
{
  "_nodes" : {
    "total" : 2,
    "successful" : 2,
    "failed" : 0
  },
  "cluster_name" : "my-application",
  "nodes" : {
    "-lnKCmBXRpiwExLns0jc9g" : {
      "timestamp" : 1502446878773,
      "name" : "node-1",
      "transport_address" : "10.104.3.2:9300",
      "host" : "10.104.3.2",
      "ip" : "10.104.3.2:9300",
      "roles" : [
        "master",
        "data",
        "ingest"
      ],
      "http" : {
        "current_open" : 4,
        "total_opened" : 29
      }
    },
    "WVsgYi2HT8GWnZU1kUwFwA" : {
      "timestamp" : 1502446878782,
      "name" : "node-2",
      "transport_address" : "10.186.7.221:9300",
      "host" : "10.186.7.221",
      "ip" : "10.186.7.221:9300",
      "roles" : [
        "master",
        "data",
        "ingest"
      ],
      "http" : {
        "current_open" : 0,
        "total_opened" : 2
      }
    }
  }
}			

原文发布于微信公众号 - Netkiller(netkiller-ebook)

原文发表时间:2017-08-15

本文参与腾讯云自媒体分享计划,欢迎正在阅读的你也加入,一起分享。

发表于

我来说两句

0 条评论
登录 后参与评论

相关文章

来自专栏文渊之博

简述SQL2008部署多实例集群(学习)

数据库集群     集群的存在意义是为了保证高可用、数据安全、扩展性以及负载均衡。 什么是集群?      由二台或更多物理上独立的服务器共同组成的"虚拟"服...

18510
来自专栏运维一切

ceph基础运维操作 原

保存退出 开始安装ceph Yum install ceph –y 完毕之后初始化目录

982
来自专栏数据和云

一个空格引发的血案

系统运维从来就是一个精细化的工作,除了规则与规范的约束之外,运维人员的严谨、谨慎也必不可少,有时候一个简单的错误就会导致一场灾难,小到一个字符,一个空格。 本文...

34211
来自专栏乐沙弥的世界

RMAN 配置保留策略

      RMAN保留策略关乎数据的完整性,因此事关重大,由用户定义的、基于用户数据恢复所能承受的容忍度来设置。也就是说根据恢复的需要,用户需要保留几天的数据...

643
来自专栏比原链

Derek解读Bytom源码-protobuf生成比原核心代码

Gitee地址:https://gitee.com/BytomBlockchain/bytom

581
来自专栏纯洁的微笑

分布式文件系统FastDFS详解

上一篇文章《一次FastDFS并发问题的排查经历》介绍了一次生产排查并发问题的经历,可能有些人对FastDFS不是特别的了解,因此计划写几篇文章完整的介绍一下这...

3165
来自专栏玩转JavaEE

elasticsearch安装与配置

念叨了两天,第一篇安装教程总算来啦!。好了,废话不多说,关于elasticsearch是什么这里我就不做过多介绍了,咱们直接上干货。

631
来自专栏猛牛哥的博客

iptables回流方法

3384
来自专栏鬼谷君

构建NTP时间服务器

862
来自专栏YG小书屋

keepalived保证mysql主从自动切换

3384

扫码关注云+社区