构建一个Mongo分片集群
本文中的表述，只代表本人的学习理解，表述可能不准确，不到位，更准确的信息请到官网或其他地方查询。
规划
分片集群需要涉及到这些内容：mongos、configserver、shard、replica set。为了集群的高可用，在本文中每一部分将配置多个节点（副本）。
Mongos:是集群请求的入口，负责把对应的数据请求转发到shard分片中。
ConfigServer: 存储所有数据库元信息（路由、分片）的配置。Mongos中的数据需要从这里加载。
Shard:分片，将1个collection的内容分到不同机器上，这样就解决了硬件性能和容量问题，分而治之的思想。
ReplicaSet：副本集，为了保证每个分片的高可用，每个分片配置了多个副本集。
机器节点规划
计划安装3个mongos,3个configServer,3个分片（shard）,每个分片上有2个副本集，1个仲裁。所以共需要15个节点。而实际有三台机器：vm603,vm601,vmzsh。所以在每台机器上启多个端口，以不同的端口代表不同的节点。
每个实际机器上有5个节点，这5个节点分别是：
1. mongos
2. config Server
3. 分片Shard1的副本集/仲裁
4. 分片Shard2的副本集/仲裁
5. 分片Shard3的副本集/仲裁
端口规划：
mongos为 20000， config server 为 21000， shard1为 22001 ， shard2为22002， shard3为22003.
具体如表格所示：
|
Vm603 (114.174) |
Vm601 (114.175) |
Vmzsh (114.173) |
|
Mongos：20000 |
Mongos：20000 |
Mongos：20000 |
|
Config Server：21000 |
Config Server：21000 |
Config Server：21000 |
|
Shard1-replica set：22001 |
仲裁：22001 |
Shard1-replica set：22001 |
|
Shard2-replica set：22002 |
仲裁：22002 |
Shard2-replica set：22002 |
|
Shard3-replica set：22003 |
仲裁：22003 |
Shard3-replica set：22003 |
|
|
|
|
注：114.17x,表未IP：192.168.114.17x
节点配置
需要的目录
以下操作需要每台机器都要操作
mkdir -p /usr/local/mongodb/mongos/log
mkdir -p /usr/local/mongodb/config/data
mkdir -p /usr/local/mongodb/config/log
mkdir -p /usr/local/mongodb/shard1/data
mkdir -p /usr/local/mongodb/shard1/log
mkdir -p /usr/local/mongodb/shard2/data
mkdir -p /usr/local/mongodb/shard2/log
mkdir -p /usr/local/mongodb/shard3/data
mkdir -p /usr/local/mongodb/shard3/log
启动配置服务器
以下操作需要每台机器都要操作
[root@vm603 bin]# ./mongod--port 21000 --configsvr --dbpath /usr/local/mongodb/config/data --logpath/usr/local/mongodb/config/log/config.log –fork
启动mongos服务器
需要上面3个config server都配置完成。
以下操作需要每台机器都要操作
./mongos--configdb 192.168.114.173:21000,192.168.114.174:21000,192.168.114.175:21000--port 20000 --logpath /usr/local/mongodb/mongos/log/mongos.log --fork
启动mongo的每个分片的副本集：
以下操作需要每台机器都要操作
./mongod--shardsvr --replSet shard1 --port 22001 --dbpath/usr/local/mongodb/shard1/data --logpath/usr/local/mongodb/shard1/log/shard1.log --fork --nojournal --oplogSize 10
./mongod--shardsvr --replSet shard2 --port 22002 --dbpath/usr/local/mongodb/shard2/data --logpath /usr/local/mongodb/shard2/log/shard2.log--fork --nojournal --oplogSize 10
./mongod--shardsvr --replSet shard3 --port 22003 --dbpath/usr/local/mongodb/shard3/data --logpath/usr/local/mongodb/shard3/log/shard3.log --fork --nojournal --oplogSize 10
设置分片的副本集配置
每个分片均需要连接到任意1个副本集
第1个分片
[root@vmzsh bin]# ./mongo localhost:22001
> use admin
switched to db admin
>config={_id:"shard1",members:[
... {_id:0,host:"192.168.114.173:22001"},
...{_id:1,host:"192.168.114.174:22001"},
...{_id:2,host:"192.168.114.175:22001",arbiterOnly:true}
... ]}
> rs.initiate(config);
第2个分片：
[root@vmzsh bin]# ./mongo localhost:22002
MongoDB shell version: 2.2.2
connecting to: localhost:22002/test
> use admin
switched to db admin
>config={_id:"shard2",members:[
...{_id:0,host:"192.168.114.173:22002"},
...{_id:1,host:"192.168.114.174:22002"},
...{_id:2,host:"192.168.114.175:22002",arbiterOnly:true}
... ]}
> rs.initiate(config);
第3个分片：
[root@vmzsh bin]# ./mongo localhost:22003
MongoDB shell version: 2.2.2
connecting to: localhost:22003/test
> use admin
switched to db admin
>config={_id:"shard3",members:[
...{_id:0,host:"192.168.114.173:22003"},
...{_id:1,host:"192.168.114.174:22003"},
...{_id:2,host:"192.168.114.175:22003",arbiterOnly:true}
... ]}
> rs.initiate(config);
配置分片集群
连接到其中任意一个mongos节点，并切换到admin数据库做以下配置
[root@vm603 bin]# ./mongolocalhost:20000/admin
MongoDB shell version: 2.2.2
connecting to: localhost:20000/admin
mongos>
db.runCommand({addshard:"shard1/192.168.114.173:22001,192.168.114.174:22001,192.168.114.175:22001",name:"shard1",maxSize:20480});
{ "shardAdded" :"shard1", "ok" : 1 }
mongos> db.runCommand({addshard:"shard2/192.168.114.173:22002,192.168.114.174:22002,192.168.114.175:22002",name:"shard2",maxSize:20480});
{ "shardAdded" :"shard2", "ok" : 1 }
mongos> db.runCommand({addshard:"shard3/192.168.114.173:22003,192.168.114.174:22003,192.168.114.175:22003",name:"shard3",maxSize:20480});
{ "shardAdded" :"shard3", "ok" : 1 }
说明：
如里shard是单台服务器，用 db.runCommand( { addshard : “[: ]”} )
如果shard是副本集，
用db.runCommand( {addshard : “replicaSetName/[:port][,serverhostname2[:port],…]”});
查看分片集群配置信息
mongos> db.runCommand({listshards:1});
{
"shards" : [
{
"_id" :"shard1",
"host" :"shard1/192.168.114.173:22001,192.168.114.174:22001",
"maxSize" :NumberLong(20480)
},
{
"_id" :"shard2",
"host" :"shard2/192.168.114.173:22002,192.168.114.174:22002",
"maxSize" :NumberLong(20480)
},
{
"_id" :"shard3",
"host" :"shard3/192.168.114.173:22003,192.168.114.174:22003",
"maxSize" : NumberLong(20480)
}
],
"ok" : 1
}
创建数据库并激活分片：
[root@vm601 bin]# ./mongolocalhost:20000/admin
mongos> db
admin
mongos> use nosql
switched to db nosql
mongos> for (var i=0;i<100;i++)db.tan.insert({uid:i,uname:'nosql'+i});
mongos> use admin
switched to db admin
激活数据库分片功能
mongos>db.runCommand({enablesharding:"nosql"});
{ "ok" : 1 }
mongos>
集合分片
mongos> use nosql
switched to db nosql
mongos>db.tan.ensureIndex({"uid":1});
mongos> use admin
switched to db admin
mongos> db.runCommand({shardcollection:"nosql.tan",key:{uid:1}});
{ "collectionsharded" :"nosql.tan", "ok" : 1 }
mongos>
mongos> use nosql
switched to db nosql
mongos> db.tan.stats();
{
"sharded" : true,
"ns" : "nosql.tan",
"count" : 100,
"numExtents" : 2,
"size" : 5600,
"storageSize" : 20480,
"totalIndexSize" : 16352,
"indexSizes" : {
"_id_" : 8176,
"uid_1" : 8176
},
"avgObjSize" : 56,
"nindexes" : 2,
"nchunks" : 1,
"shards" : {
"shard1" : {
"ns" :"nosql.tan",
"count" :100,
"size" :5600,
"avgObjSize": 56,
"storageSize": 20480,
"numExtents": 2,
"nindexes" :2,
"lastExtentSize" : 16384,
"paddingFactor" : 1,
"systemFlags": 1,
"userFlags" :0,
"totalIndexSize" : 16352,
"indexSizes": {
"_id_" : 8176,
"uid_1" : 8176
},
"ok" : 1
}
},
"ok" : 1
}
该集合只存放在1个分片上。再增加些数据，然后查看：
mongos> for(vari=101;i<1000000;i++)db.tan.insert({uid:i,uname:'nosql'+i});
mongos> db.tan.stats();
{
"sharded" : true,
"ns" : "nosql.tan",
"count" : 999999,
"numExtents" : 24,
"size" : 59960000,
"storageSize" : 104407040,
"totalIndexSize" : 60714976,
"indexSizes" : {
"_id_" : 32638592,
"uid_1" : 28076384
},
"avgObjSize" : 59.96005996005996,
"nindexes" : 2,
"nchunks" : 6,
"shards" : {
"shard1" : {
"ns" :"nosql.tan",
"count" :384748,
"size" :23084912,
"avgObjSize": 60.00008317132253,
"storageSize" :43171840,
"numExtents": 10,
"nindexes" :2,
"lastExtentSize" : 13942784,
"paddingFactor" : 1,
"systemFlags": 1,
"userFlags" :0,
"totalIndexSize" : 23448768,
"indexSizes": {
"_id_" : 12607392,
"uid_1" : 10841376
},
"ok" : 1
},
"shard2" : {
"ns" :"nosql.tan",
"count" :15598,
"size" :895884,
"avgObjSize": 57.43582510578279,
"storageSize": 2793472,
"numExtents": 5,
"nindexes" :2,
"lastExtentSize" : 2097152,
"paddingFactor" : 1,
"systemFlags": 1,
"userFlags" :0,
"totalIndexSize" : 1046528,
"indexSizes": {
"_id_" : 555968,
"uid_1" : 490560
},
"ok" : 1
},
"shard3" : {
"ns" :"nosql.tan",
"count" :599653,
"size" :35979204,
"avgObjSize": 60.00004002314672,
"storageSize": 58441728,
"numExtents": 9,
"nindexes" :2,
"lastExtentSize": 20643840,
"paddingFactor" : 1,
"systemFlags": 1,
"userFlags" :0,
"totalIndexSize" : 36219680,
"indexSizes": {
"_id_" : 19475232,
"uid_1" : 16744448
},
"ok" : 1
}
},
"ok" : 1
}
Mongo的分片集群配置完成。
JAVA*问该集群里，可同时连接到上面3个mongos上.
本文部分参考：http://www.lanceyan.com/tech/arch/mongodb_shard1.html