MongoDB的分片架构搭建

一、我们为什么需要分片？

在Mongodb里面存在另一种集群，就是分片技术,可以满足MongoDB数据量大量增长的需求。
当MongoDB存储海量的数据时，一台机器可能不足以存储数据，也可能不足以提供可接受的读写吞吐量。这时，我们就可以通过在多台机器上分割数据，使得数据库系统能存储和处理更多的数据。
​

• 复制所有的写入操作到主节点
• 延迟的敏感数据会在主节点查询
• 单个副本集限制在12个节点
• 当请求量巨大时会出现内存不足。
• 本地磁盘不足
• 垂直扩展价格昂贵
  

二、分片架构

上图中主要有如下所述三个主要组件：

• Shard: 用于存储实际的数据块，实际生产环境中一个shard server角色可由几台机器组个一个relica set承担，防止主机单点故障
• Config Server: mongod实例，存储了整个 ClusterMetadata，其中包括 chunk信息。从3.4版本开始，config Server必须部署为副本集架构（CSRS）
• Query Routers: 前端路由，客户端由此接入，且让整个集群看上去像单一数据库，前端应用可以透明使用。​
  

三、测试规划

角色	IP:PORT
​	172.16.113.129:27017
分片(shard)	172.16.113.129:27018
	172.16.113.129:27019
配置(config)	172.16.113.137:27017
路由(router)	172.16.113.137:27018

备注：机器有限，shard、config、router我们都使用单节点，不用副本集。但是生产禁用

四、搭建

4.1 shard

4.1.1 配置文件

[root@node01 mongo]# cat mongod.conf 
dbpath=/usr/bin/mongo/data  
logpath=/usr/bin/mongo/logs/mongodb.log 
pidfilepath=/usr/bin/mongo/mongod.pid  
directoryperdb=true  
logappend=true  
bind_ip=127.0.0.1,172.16.113.129 
#keyFile=/usr/bin/mongo/mongodb-keyfile
port=27017  
fork=true 
#auth=true
#replSet=rs0
oplogSize=128

[root@node01 mongo]# cat mongod2.conf 
dbpath=/usr/bin/mongo/data2  
logpath=/usr/bin/mongo/logs2/mongodb.log 
pidfilepath=/usr/bin/mongo/data2/mongod.pid  
directoryperdb=true  
logappend=true  
bind_ip=127.0.0.1,172.16.113.129 
#keyFile=/usr/bin/mongo/mongodb-keyfile
port=27018  
fork=true 
#auth=true
#replSet=rs0
oplogSize=128

[root@node01 mongo]# cat mongod3.conf  
dbpath=/usr/bin/mongo/data3  
logpath=/usr/bin/mongo/logs3/mongodb.log 
pidfilepath=/usr/bin/mongo/data3/mongod.pid  
directoryperdb=true  
logappend=true  
bind_ip=127.0.0.1,172.16.113.129 
#keyFile=/usr/bin/mongo/mongodb-keyfile
port=27019
fork=true 
#auth=true
#replSet=rs0
oplogSize=128
[root@node01 mongo]# 

4.1.2 启动

 mongod -f /usr/bin/mongo/mongod.conf
 mongod -f /usr/bin/mongo/mongod2.conf
 mongod -f /usr/bin/mongo/mongod3.conf

4.2 config

4.2.1 配置文件

dbpath=/usr/bin/mongo/config  
logpath=/usr/bin/mongo/configlogs/mongodb.log 
pidfilepath=/usr/bin/mongo/config/mongod.pid  
directoryperdb=true  
logappend=true  
bind_ip=127.0.0.1,172.16.113.137 
configsvr=true
port=27017  
fork=true  
oplogSize=100

4.2.2 启动

mongod -f /usr/bin/mongo/config.conf

4.3 mongos

4.3.1 配置文件

cat router.conf 
logpath=/usr/bin/mongo/routerlogs/mongodb.log
pidfilepath=/usr/bin/mongo/router/mongod.pid  
logappend=true
bind_ip=127.0.0.1,172.16.113.137
port=40000
configdb=172.16.113.137:27017 #关联配置服务
fork=true
chunkSize=200 #mongos启动参数中，chunkSize这一项是用来指定chunk的大小的，单位是MB，默认大小为200MB.

4.3.2 启动

[root@node02 mongo]# mongos -f /usr/bin/mongo/router.conf
2021-09-28T15:28:29.906+0800 W SHARDING [main] Running a sharded cluster with fewer than 3 config servers should only be done for testing purposes and is not recommended for production.
about to fork child process, waiting until server is ready for connections. 
forked process: 2693
child process started successfully, parent exiting

五、配置分片

5.1 添加分片节点

接下来，我们使用MongoDB Shell登录到mongos，添加Shard节点

mongos> use admin
switched to db admin
mongos> db.runCommand({addshard:"172.16.113.129:27017"})
{ "shardAdded" : "shard0000", "ok" : 1 }
mongos> db.runCommand({addshard:"172.16.113.129:27018"})
{ "shardAdded" : "shard0001", "ok" : 1 }
mongos> db.runCommand({addshard:"172.16.113.129:27019"})
{ "shardAdded" : "shard0001", "ok" : 1 }

查看状态

mongos> sh.status()
--- Sharding Status --- 
  sharding version: {
        "_id" : 1,
        "minCompatibleVersion" : 5,
        "currentVersion" : 6,
        "clusterId" : ObjectId("6152c41db4c49a97e985606f")
}
  shards:
        {  "_id" : "shard0001",  "host" : "172.16.113.129:27019" }
        {  "_id" : "shard0002",  "host" : "172.16.113.129:27017" }
        {  "_id" : "shard0003",  "host" : "172.16.113.129:27018" }
  active mongoses:
        "3.2.10" : 1
  balancer:
        Currently enabled:  yes
        Currently running:  no
        Failed balancer rounds in last 5 attempts:  0
        Migration Results for the last 24 hours: 
                No recent migrations
  databases:
        {  "_id" : "gengjin",  "primary" : "shard0001",  "partitioned" : false }
        {  "_id" : "hello",  "primary" : "shard0001",  "partitioned" : false }
        {  "_id" : "kobe",  "primary" : "shard0001",  "partitioned" : false }
        {  "_id" : "mydb",  "primary" : "shard0001",  "partitioned" : false }
        {  "_id" : "nihao",  "primary" : "shard0001",  "partitioned" : false }

5.2 模拟测试

#1.创建用户
mongos> use admin
switched to db admin
mongos> db.createUser({"user":"admin","pwd":"admin","customData":{"name":"gengjin"},roles:[{role:"userAdminAnyDatabase",db:"admin"}]})
#2.设置分片存储的数据库
mongos> db.runCommand({ enablesharding:"test" })
{ "ok" : 1 }
mongos> 
#3.设置分片存储的数据库集合
mongos> sh.shardCollection("test.user", { _id : "hashed" } )
{ "collectionsharded" : "test.user", "ok" : 1 }
mongos> 

实例

mongos> db.gengjin.ensureIndex({name:1})
{
        "raw" : {
                "172.16.113.129:27019" : {
                        "createdCollectionAutomatically" : true,
                        "numIndexesBefore" : 1,
                        "numIndexesAfter" : 2,
                        "ok" : 1
                }
        },
        "ok" : 1
}
mongos> sh.shardCollection("test.gengjin", {name:1})
{ "collectionsharded" : "test.gengjin", "ok" : 1 }
mongos> 

#1.查看集合状态
mongos> db.gengjin.stats()
{
        "sharded" : true,
        "capped" : false,
        "ns" : "test.gengjin",
        "count" : 0,
        "size" : 0,
        "storageSize" : 4096,
        "totalIndexSize" : 8192,
        "indexSizes" : {
                "_id_" : 4096,
                "name_1" : 4096
        },
        .....
#2.查看分片状态
db.printShardingStatus()/sh.status()
#3.在mongos节点插入数据
for (var i=0; i<1000; i++) db.gengjin.insert({name:"zhangsan"+i, age:i})
#4.查看shard01、shard02、shard03三个节点，可以看到数据已经分布在这三个节点上了(mongos节点上是不保存插入数据的)
db.user.find()

​