1 MongoDB 分片(高可用)

1.1 准备工作

  • 三台虚拟机
  • 安装MongoDB
  • 虚拟机相互之间可以相互通信
  • 虚拟机与主机之间可以相互通信

1.2 安装MongoDB

在Ubuntu16.04 中安装 MongoDB 。参考步骤MongoDB官方网站

  • 安装时会报错

    E: The method driver /usr/lib/apt/methods/https could not be found.
    N: Is the package apt-transport-https installed?

    提示需要安装apt-transport-https

    sudo apt-get install -y apt-transport-https

1.3 启动MongoDB

sudo service mongod start

检查是否启动成功

sudo cat /var/log/mongodb/mongod.log

2019-04-19T15:40:52.808+0800 I NETWORK [initandlisten] waiting for connections on port 27017

2 MongoDB 分片

分片(sharding)是将数据拆分,将其分散存到不同机器上的过程。MongoDB 支持自动分片,可以使数据库架构对应用程序不可见。对于应用程序来说,好像始终在使用一个单机的 MongoDB 服务器一样,另一方面,MongoDB 自动处理数据在分片上的分布,也更容易添加和删除分片。类似于MySQL中的分库。

2.1 基础组件

其利用到了四个组件:mongos,config server,shard,replica set

mongos:数据库集群请求的入口,所有请求需要经过mongos进行协调,无需在应用层面利用程序来进行路由选择,mongos其自身是一个请求分发中心,负责将外部的请求分发到对应的shard服务器上,mongos作为统一的请求入口,为防止mongos单节点故障,一般需要对其做HA。可以理解为在微服务架构中的路由Eureka。

config server:配置服务器,存储所有数据库元数据(分片,路由)的配置。mongos本身没有物理存储分片服务器和数据路由信息,只是存缓存在内存中来读取数据,mongos在第一次启动或后期重启时候,就会从config server中加载配置信息,如果配置服务器信息发生更新会通知所有的mongos来更新自己的状态,从而保证准确的请求路由,生产环境中通常也需要多个config server,防止配置文件存在单节点丢失问题。理解为 配置中心

shard:在传统意义上来讲,如果存在海量数据,单台服务器存储1T压力非常大,无论考虑数据库的硬盘,网络IO,又有CPU,内存的瓶颈,如果多台进行分摊1T的数据,到每台上就是可估量的较小数据,在mongodb集群只要设置好分片规则,通过mongos操作数据库,就可以自动把对应的操作请求转发到对应的后端分片服务器上。真正的存储

replica set:在总体mongodb集群架构中,对应的分片节点,如果单台机器下线,对应整个集群的数据就会出现部分缺失,这是不能发生的,因此对于shard节点需要replica set来保证数据的可靠性,生产环境通常为2个副本+1个仲裁。 副本保持数据的HA

2.2 架构图

img

3 安装部署

为了节省服务器,采用多实例配置,三个mongos,三个config server,单个服务器上面运行不通角色的shard(为了后期数据分片均匀,将三台shard在各个服务器上充当不同的角色。),在一个节点内采用replica set保证高可用,对应主机与端口信息如下:

主机名 IP地址 组件mongos 组件config server shard
主节点: 22001
mongodb-1 192.168.90.130 端口:20000 端口:21000 副本节点:22002
仲裁节点:22003
主节点: 22002
mongodb-2 192.168.90.131 端口:20000 端口:21000 副本节点:22001
仲裁节点:22003
主节点: 22003
mongodb-3 192.168.90.132 端口:20000 端口:21000 副本节点:22001
仲裁节点:22002

架构

3.1 部署配置服务器集群

3.1.1 先创建对应的文件夹

mkdir -p /usr/local/mongo/mongoconf/{data,log,config}
touch mongoconfg.conf
touch mongoconf.log

3.1.2 填写对应的配置文件

# mongod.conf

# for documentation of all options, see:
#   http://docs.mongodb.org/manual/reference/configuration-options/

# Where and how to store data.
storage:
  dbPath: /usr/local/mongo/mongoconf/data
  journal:
    enabled: true
    commitIntervalMs: 200
#  engine:
#  mmapv1:
#  wiredTiger:

# where to write logging data.
systemLog:
  destination: file
  logAppend: true
  path: /usr/local/mongo/mongoconf/log/mongoconf.log

# network interfaces
net:
  port: 21000
  bindIp: 0.0.0.0
  maxIncomingConnections: 1000

# how the process runs
processManagement:
  fork: true
#security:


#operationProfiling:

#replication:

replication:
  replSetName: replconf

#sharding:
sharding:
  clusterRole: configsvr
## Enterprise-Only Options:

#auditLog:

#snmp:

配置集群需要指定 data ,log ,以及对应的sharding角色

3.1.3 启动config集群

在三台机器上都配置好对应的文件后可以启动config集群

mogod -f /usr/local/mongo/mongoconf/conf/mongoconf.conf 

3.1.4 副本初始化操作

进入任意一台机器(130为例),做集群副本初始化操作

mongo 192.168.90.130

config = {
    _id:"replconf"
    members:[
        {_id:0,host:"192.168.90.130:21000"}
        {_id:0,host:"192.168.90.131:21000"}
        {_id:0,host:"192.168.90.132:21000"}
    ]
    
}
rs.initiate(config)

//查看集群状态
rs.status()

返回结果如下

{
    "set" : "replconf",
    "date" : ISODate("2019-04-21T06:38:50.164Z"),
    "myState" : 2,
    "term" : NumberLong(3),
    "syncingTo" : "192.168.90.131:21000",
    "syncSourceHost" : "192.168.90.131:21000",
    ...
    "members" : [
        {
            "_id" : 0,
            "name" : "192.168.90.130:21000",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 305,
            "optime" : {
                "ts" : Timestamp(1555828718, 1),
                "t" : NumberLong(3)
            },
            "optimeDurable" : {
                "ts" : Timestamp(1555828718, 1),
                "t" : NumberLong(3)
            },
            "optimeDate" : ISODate("2019-04-21T06:38:38Z"),
            "optimeDurableDate" : ISODate("2019-04-21T06:38:38Z"),
            "lastHeartbeat" : ISODate("2019-04-21T06:38:49.409Z"),
            "lastHeartbeatRecv" : ISODate("2019-04-21T06:38:49.408Z"),
            "pingMs" : NumberLong(0),
            "lastHeartbeatMessage" : "",
            "syncingTo" : "192.168.90.131:21000",
            "syncSourceHost" : "192.168.90.131:21000",
            "syncSourceId" : 1,
            "infoMessage" : "",
            "configVersion" : 1
        },
        {
            "_id" : 1,
            "name" : "192.168.90.131:21000",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 307,
            "optime" : {
                "ts" : Timestamp(1555828718, 1),
                "t" : NumberLong(3)
            },
            "optimeDurable" : {
                "ts" : Timestamp(1555828718, 1),
                "t" : NumberLong(3)
            },
            "optimeDate" : ISODate("2019-04-21T06:38:38Z"),
            "optimeDurableDate" : ISODate("2019-04-21T06:38:38Z"),
            "lastHeartbeat" : ISODate("2019-04-21T06:38:49.380Z"),
            "lastHeartbeatRecv" : ISODate("2019-04-21T06:38:49.635Z"),
            "pingMs" : NumberLong(0),
            "lastHeartbeatMessage" : "",
            "syncingTo" : "",
            "syncSourceHost" : "",
            "syncSourceId" : -1,
            "infoMessage" : "",
            "electionTime" : Timestamp(1555828429, 1),
            "electionDate" : ISODate("2019-04-21T06:33:49Z"),
            "configVersion" : 1
        },
        {
            "_id" : 2,
            "name" : "192.168.90.132:21000",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 319,
            "optime" : {
                "ts" : Timestamp(1555828718, 1),
                "t" : NumberLong(3)
            },
            "optimeDate" : ISODate("2019-04-21T06:38:38Z"),
            "syncingTo" : "192.168.90.131:21000",
            "syncSourceHost" : "192.168.90.131:21000",
            "syncSourceId" : 1,
            "infoMessage" : "",
            "configVersion" : 1,
            "self" : true,
            "lastHeartbeatMessage" : ""
        }
    ],
    "ok" : 1,
    ...
}

可以看出config集群部署成功,且通过选举机制选定131为primary节点。

3.2 部署Shard分片集群

3.2.1 分别对三片数据存储创建对应的文件夹

mkdir -p shard1/{data,conf,log}
mkdir -p shard2/{data,conf,log}
mkdir -p shard3/{data,conf,log}

touch shard1/log/shard1.log
touch shard2/log/shard2.log
touch shard3/log/shard3.log

3.2.2 创建对应的配置文件

# mongod.conf

# for documentation of all options, see:
#   http://docs.mongodb.org/manual/reference/configuration-options/

# Where and how to store data.
storage:
  dbPath: /usr/local/mongo/shard1/data
  journal:
    enabled: true
    commitIntervalMs: 200
  mmapv1:
    smallFiles: true
#  wiredTiger:

# where to write logging data.
systemLog:
  destination: file
  logAppend: true
  path: /usr/local/mongo/shard1/log/shard1.log

# network interfaces
net:
  port: 22001
  bindIp: 0.0.0.0
  maxIncomingConnections: 1000

# how the process runs
processManagement:
  fork: true
#security:


#operationProfiling:

#replication:
replication:
  replSetName: shard1
  oplogSizeMB: 4096
#sharding:
sharding:
  clusterRole: shardsvr
## Enterprise-Only Options:

#auditLog:

#snmp:

分别建立shard1,shard2,shard3 的配置文件

3.2.3 启动对应的shard

 mongod -f /usr/local/mongo/shard1/conf/shard1.conf 
 mongod -f /usr/local/mongo/shard2/conf/shard2.conf 
 mongod -f /usr/local/mongo/shard3/conf/shard3.conf 

使用

netstat -lntup 

来查看进程

Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address           Foreign Address         State       PID/Program name
tcp        0      0 0.0.0.0:22001           0.0.0.0:*               LISTEN      1891/mongod     
tcp        0      0 0.0.0.0:22002           0.0.0.0:*               LISTEN      1974/mongod     
tcp        0      0 0.0.0.0:22003           0.0.0.0:*               LISTEN      2026/mongod     
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN      -               
tcp        0      0 127.0.0.1:6010          0.0.0.0:*               LISTEN      -               
tcp        0      0 0.0.0.0:21000           0.0.0.0:*               LISTEN      1779/mongod  

发现进程全部启动

3.2.4 指定对应shard集群的副本

在130 上登录

mongo 192.168.90.130:22001

指定集群以及对应的arbiter节点

config = {
    _id:"shard1",
    members:[
        {_id:0,host:"192.168.90.130:22001"},
        {_id:1,host:"192.168.90.131:22001"},
        {_id:2,host:"192.168.90.132:22001",arbiterOnly:true}
    ]   
}
rs.initate(config)
rs.status()

在131,132 节点登录,哪一个节点初始化 即哪一个节点第一次优先作为primary节点

config = {
    _id:"shard2",
    members:[
        {_id:0,host:"192.168.90.130:22002",arbiterOnly:true},
        {_id:1,host:"192.168.90.131:22002"},
        {_id:2,host:"192.168.90.132:22002"}
    ]   
}
rs.initate(config)
rs.status()

config = {
    _id:"shard3",
    members:[
        {_id:0,host:"192.168.90.130:22001"},
        {_id:1,host:"192.168.90.131:22001",arbiterOnly:true},
        {_id:2,host:"192.168.90.132:22001"}
    ]   
}
rs.initate(config)
rs.status()

3.3 配置mongos路由集群

3.3.1 创建对应文件

mkdir -p ./mongos/{config,log}

因为mongos不需要存储数据或者元数据信息,只负责处理请求分发,当启动时到config集群中取得元数据信息加载到内存使用。

3.3.2 配置文件

# mongod.conf

# for documentation of all options, see:
#   http://docs.mongodb.org/manual/reference/configuration-options/

# Where and how to store data.
#  engine:
#  mmapv1:
#  wiredTiger:

# where to write logging data.
systemLog:
  destination: file
  logAppend: true
  path: /usr/local/mongo/mongos/log/mongos.log

# network interfaces
net:
  port: 20000
  bindIp: 0.0.0.0
  maxIncomingConnections: 1000

# how the process runs
processManagement:
  fork: true
sharding:
  configDB: replconf/mongo0:21000,mongo1:21000,mongo2:21000
## Enterprise-Only Options:


#auditLog:


3.3.3 启动mongos

mongos -f /usr/local/mongo/mongos/conf/mongos.conf 

查看集群启动情况

Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address           Foreign Address         State       PID/Program name
tcp        0      0 0.0.0.0:22001           0.0.0.0:*               LISTEN      1891/mongod     
tcp        0      0 0.0.0.0:22002           0.0.0.0:*               LISTEN      1974/mongod     
tcp        0      0 0.0.0.0:22003           0.0.0.0:*               LISTEN      2026/mongod     
tcp        0      0 0.0.0.0:22              0.0.0.0:*               LISTEN      -               
tcp        0      0 127.0.0.1:6010          0.0.0.0:*               LISTEN      -               
tcp        0      0 0.0.0.0:20000           0.0.0.0:*               LISTEN      2174/mongos     
tcp        0      0 0.0.0.0:21000           0.0.0.0:*               LISTEN      1779/mongod     

3.3.4 在admin表中填写shard分片信息

登录任意一台mongos,在admin表中加入信息

use admin

db.runCommand(
    {
        addShard:"shard1/192.168.90.130:22001,192.168.90.131:22001,192.168.90.132:22001"
    }
)

db.runCommand(
    {
        addShard:"shard2/192.168.90.130:22002,192.168.90.131:22002,192.168.90.132:22002"
    }
)
db.runCommand(
    {
        addShard:"shard3/192.168.90.130:22003,192.168.90.131:22003,192.168.90.132:22003"
    }
)

sh.status()

返回信息:

--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "minCompatibleVersion" : 5,
    "currentVersion" : 6,
    "clusterId" : ObjectId("5cba82d33290d8f4fb3ac8f7")
  }
  shards:
        {  "_id" : "shard1",  "host" : "shard1/192.168.90.130:22001,192.168.90.132:22001",  "state" : 1 }
        {  "_id" : "shard2",  "host" : "shard2/192.168.90.130:22002,192.168.90.131:22002",  "state" : 1 }
        {  "_id" : "shard3",  "host" : "shard3/192.168.90.131:22003,192.168.90.132:22003",  "state" : 1 }
  active mongoses:
....

3.3.5 给对应的表进行分片

use admin
db.runCommand({
    shardCollection:"lishubindb.table1",key:{_id:"hashed"}
})
db.runCommand({
    listshards:1
})

3.4 测试

加入10W条数据到table1集合中

use lishubindb
for(var i=0;i<100000;i++){
    db.table1.insert({
        "name":"lishubin"+i,"num":i
    })
}

观察分片情况

db.table1.status()

    "ns" : "lishubindb.table1",
    "count" : 100000,
    "size" : 5888890,
    "storageSize" : 2072576,
    "totalIndexSize" : 4694016,
    "indexSizes" : {
        "_id_" : 1060864,
        "_id_hashed" : 3633152
    },
    "avgObjSize" : 58,
    "maxSize" : NumberLong(0),
    "nindexes" : 2,
    "nchunks" : 6,
    "shards" : {
        "shard3" : {
            "ns" : "lishubindb.table1",
            "size" : 1969256,
            "count" : 33440,
            "avgObjSize" : 58,
            "storageSize" : 712704,
            "capped" : false,
            ...
        },
        "shard2" : {
            "ns" : "lishubindb.table1",
            "size" : 1973048,
            "count" : 33505,
            "avgObjSize" : 58,
            "storageSize" : 708608,
            "capped" : false,
            ...
        },
        "shard1" : {
            "ns" : "lishubindb.table1",
            "size" : 1946586,
            "count" : 33055,
            "avgObjSize" : 58,
            "storageSize" : 651264,
            "capped" : false,
            ...
        },

4 参考文献

版权声明:本文为lishubin原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接和本声明。
本文链接:https://www.cnblogs.com/lishubin/p/11662960.html