MongoDB整理笔记のReplica Sets + Sharding

MongoDB Auto-Sharding 解决了海量存储和动态扩容的问题，但离实际生产环境所需的高可靠、高可用还有些距离，所以有了"Replica Sets + Sharding"的解决方案。

shard:

使用Replica Sets，确保每个数据节点都具有备份，自动容错转移，自动回复能力。

config:

使用3个配置服务器，确保元数据的完整性。

route:

使用3个路由进程，实现负载均衡，提高客户端接入性能。

配置Replica Sets + Sharding 架构图：

配置Replica Sets + Sharding

  （1）配置shard1所用到的Replica Sets

在server A上

[root@localhost bin]# /Apps/mongo/bin/mongod --shardsvr --replSet shard1 --port 27017
--dbpath /data/shard1_1 --logpath /data/shard1_1/shard1_1.log --logappend --fork
[root@localhost bin]# all output going to: /data/shard1_1/shard1_1.log
forked process: 18923

在server B上

[root@localhost bin]# /Apps/mongo/bin/mongod --shardsvr --replSet shard1 --port 27017
--dbpath /data/shard1_2 --logpath /data/shard1_2/shard1_2.log --logappend --fork
forked process: 18859
[root@localhost bin]# all output going to: /data/shard1_2/shard1_2.log
[root@localhost bin]#

在Server C 上

[root@localhost bin]# /Apps/mongo/bin/mongod --shardsvr --replSet shard1 --port 27017
--dbpath /data/shard1_3 --logpath /data/shard1_3/shard1_3.log --logappend --fork
all output going to: /data/shard1_3/shard1_3.log
forked process: 18768
[root@localhost bin]#

用mongo 连接其中一台机器的27017 端口的mongod，初始化Replica Sets“shard1”，执行:

[root@localhost bin]# ./mongo --port 27017
MongoDB shell version: 1.8.1
connecting to: 127.0.0.1:27017/test
> config = {_id: 'shard1', members: [
... {_id: 0, host: '192.168.3.231:27017'},
... {_id: 1, host: '192.168.3.232:27017'},
... {_id: 2, host: '192.168.3.233:27017'}]
... }
……
> rs.initiate(config)
{
"info" : "Config now saved locally. Should come online in about a minute.",
"ok" : 1
}

（2）配置shard2所用到的Replica Sets

在server A上

[root@localhost bin]# /Apps/mongo/bin/mongod --shardsvr --replSet shard2 --port 27018
--dbpath /data/shard2_1 --logpath /data/shard2_1/shard2_1.log --logappend --fork
all output going to: /data/shard2_1/shard2_1.log
[root@localhost bin]# forked process: 18993
[root@localhost bin]#

在server B上

[root@localhost bin]# /Apps/mongo/bin/mongod --shardsvr --replSet shard2 --port 27018
--dbpath /data/shard2_2 --logpath /data/shard2_2/shard2_2.log --logappend --fork
all output going to: /data/shard2_2/shard2_2.log
forked process: 18923
[root@localhost bin]#

在Server C上

[root@localhost bin]# /Apps/mongo/bin/mongod --shardsvr --replSet shard2 --port 27018
--dbpath /data/shard2_3 --logpath /data/shard2_3/shard2_3.log --logappend --fork
[root@localhost bin]# all output going to: /data/shard2_3/shard2_3.log
forked process: 18824
[root@localhost bin]#

用mongo 连接其中一台机器的27018 端口的mongod，初始化Replica Sets “shard2”，执行:

[root@localhost bin]# ./mongo --port 27018
MongoDB shell version: 1.8.1
connecting to: 127.0.0.1:27018/test
> config = {_id: 'shard2', members: [
... {_id: 0, host: '192.168.3.231:27018'},
... {_id: 1, host: '192.168.3.232:27018'},
... {_id: 2, host: '192.168.3.233:27018'}]
... }
……
> rs.initiate(config)
{
"info" : "Config now saved locally. Should come online in about a minute.",
"ok" : 1

db.runCommand({ enablesharding:"test" })
db.runCommand({ shardcollection: "test.users", key: { _id:1 }})

}

  （3）配置3 台Config Server
    在Server A、B、C上执行:

/Apps/mongo/bin/mongod --configsvr --dbpath /data/config --port 20000 --logpath
/data/config/config.log --logappend --fork

（4）配置3台Route Process
    在Server A、B、C上执行:

/Apps/mongo/bin/mongos --configdb
192.168.3.231:20000,192.168.3.232:20000,192.168.3.233:20000 --port 30000 --chunkSize 1
--logpath /data/mongos.log --logappend --fork

（5）配置Shard Cluster
    连接到其中一台机器的端口30000 的mongos 进程，并切换到admin 数据库做以下配置

[root@localhost bin]# ./mongo --port 30000
MongoDB shell version: 1.8.1
connecting to: 127.0.0.1:30000/test
> use admin
switched to db admin
>db.runCommand({addshard:"shard1/192.168.3.231:27017,192.168.3.232:27017,192.168.3.233:
27017"});
{ "shardAdded" : "shard1", "ok" : 1 }
>db.runCommand({addshard:"shard2/192.168.3.231:27018,192.168.3.232:27018,192.168.3.233:
27018"});
{ "shardAdded" : "shard2", "ok" : 1 }
>

激活数据库及集合的分片

db.runCommand({ enablesharding:"test" })
db.runCommand({ shardcollection: "test.users", key: { _id:1 }})

（6）验证Sharding正常工作
    连接到其中一台机器的端口30000 的mongos 进程，并切换到test 数据库，以便添加测试数据

use test
for(var i=1;i<=200000;i++) db.users.insert({id:i,addr_1:"Beijing",addr_2:"Shanghai"});
db.users.stats()
{
"sharded" : true,
"ns" : "test.users",
"count" : 200000,
"size" : 25600384,
"avgObjSize" : 128,
"storageSize" : 44509696,
"nindexes" : 2,
"nchunks" : 15,
"shards" : {
"shard0000" : {
……
},
"shard0001" : {
……
}
},
"ok" : 1
}

可以看到Sharding搭建成功了，跟我们期望的结果一致，至此我们就将Replica Sets与Sharding结合的架构也学习完毕了!