NoSQL数据库:MongoDB初探

跟着时下炒得火热的NOSQL潮流，学习了一下mongodb，记录在此，希望与感兴趣的同学一起研究！

MongoDB概述

mongodb由C＋＋写就，其名字来自humongous这个单词的中间部分，是由10gen开发并维护的,关于它的一个最简洁描述为：scalable, high-performance, open source, schema-free, document-oriented database。MongoDB的主要目标是在键/值存储方式（提供了高性能和高度伸缩性）以及传统的RDBMS系统（丰富的功能）架起一座桥梁，集两者的优势于一身。

MongoDB特性：

l  面向文档存储

l  全索引支持,扩展到内部对象和内嵌数组

l  复制和高可用

l  自动分片支持云级扩展性

l  查询记录分析

l  动态查询

l  快速,就地更新

l  支持Map/Reduce操作

l  GridFS文件系统

l  商业支持,培训和咨询

官网: http://www.mongodb.org/

配置

Master-slaves 模式

机器IP角色test001192.168.1.1mastertest002192.168.1.2slavetest003192.168.1.3slavetest004192.168.1.4slavetest005192.168.1.5slavetest006192.168.1.6slave

启动master:

1

./mongod -dbpath=/mongodb/data/ -logpath=/mongodb/logs/mongodb.log -oplogSize=10000 -logappend -master -port=27017 -fork

添加repl用户:

1

2

3

./mongo

>use local

> db.addUser('repl','replication');

启动slaves:

1

2

./mongod -dbpath=/mongodb/data/ -logpath=/mongodb/logs/mongodb.log -slave  -port=27017 -source=test001:27017 --autoresync

-fork

添加repl用户:

1

2

3

./mongo

>use local

> db.addUser('repl','replication');

autoresync 参数会在系统发生意外情况造成主从数据不同步时，自动启动复制操作 (同步复制 10 分钟内仅执行一次)。除此之外，还可以用 –slavedelay 设定更新频率(秒)。

通常我们会使用主从方案实现读写分离，但需要设置 Slave_OK。

slaveOk

When querying a replica pair or replica set, drivers route their requests to the master mongod by default; to perform a query against an (arbitrarily-selected) slave, the query can be run with the slaveOk option. Here’s how to do so in the shell:

db.getMongo().setSlaveOk(); // enable querying a slavedb.users.find(...)

Note: some language drivers permit specifying the slaveOk option on each find(), others make this a connection-wide setting. See your language’s driver for details.

Replica Set模式

Replica Sets 使用 n 个 Mongod 节点，构建具备自动容错转移(auto-failover)、自动恢复(auto-recovery) 的高可用方案。

机器IP角色test001192.168.1.1secondarytest002192.168.1.2secondarytest003192.168.1.3primarytest004192.168.1.4secondarytest005192.168.1.5secondarytest006192.168.1.6secondarytest007192.168.1.7secondary

启动:

1

./mongod -dbpath=/mongodb/data/ -logpath=/mongodb/logs/mongodb.log -oplogSize=10000 -logappend -replSet set1 -port=27017 -fork –rest

添加repl用户:

1

2

3

./mongo

>use local

> db.addUser('repl','replication');

配置:

1

2

3

4

5

6

7

8

9

10

config={_id:'set1',members:[

{_id:0,host:'test001:27017'},

{_id:1,host:'test002:27017'},

{_id:2,host:'test003:27017'},

{_id:3,host:'test004:27017'},

{_id:4,host:'test005:27017'},

{_id:5,host:'test006:27017'},

{_id:6,host:'test007:27017'}]

}

rs.initiate(config);

查看:

访问 http://test001 :28017/_replSet

或者

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

./mongo

> rs.status()

{

"set" : "set1",

"date" : "Fri Dec 03 2010 00:57:44 GMT+0800 (CST)",

"myState" : 2,

"members" : [

{

"_id" : 0,

"name" : "test001:27017",

"health" : 1,

"state" : 2,

"self" : true

},

{

"_id" : 1,

"name" : "test002:27017",

"health" : 1,

"state" : 2,

"uptime" : 194451,

"lastHeartbeat" : "Fri Dec 03 2010 00:57:42 GMT+0800 (CST)"

},

{

"_id" : 2,

"name" : "test003:27017",

"health" : 1,

"state" : 1,

"uptime" : 194689,

"lastHeartbeat" : "Fri Dec 03 2010 00:57:43 GMT+0800 (CST)"

},

{

"_id" : 3,

"name" : "test004:27017",

"health" : 1,

"state" : 2,

"uptime" : 194689,

"lastHeartbeat" : "Fri Dec 03 2010 00:57:42 GMT+0800 (CST)"

},

{

"_id" : 4,

"name" : "test005:27017",

"health" : 1,

"state" : 2,

"uptime" : 194689,

"lastHeartbeat" : "Fri Dec 03 2010 00:57:42 GMT+0800 (CST)"

},

{

"_id" : 5,

"name" : "test006:27017",

"health" : 1,

"state" : 2,

"uptime" : 194689,

"lastHeartbeat" : "Fri Dec 03 2010 00:57:43 GMT+0800 (CST)"

},

{

"_id" : 6,

"name" : "test007:27017",

"health" : 1,

"state" : 2,

"uptime" : 194689,

"lastHeartbeat" : "Fri Dec 03 2010 00:57:42 GMT+0800 (CST)"

}

],

"ok" : 1

}

在Replica Sets上做操作后调用getlasterror使写操作同步到至少3台机器后才返回

db.runCommand( { getlasterror : 1 , w : 3 } )

注：该模式不支持auth功能，需要auth功能请选择m-s模式

Sharding模式

要构建一个 MongoDB Sharding Cluster，需要三种角色：

• Shard Server: mongod 实例，用于存储实际的数据块。
• Config Server: mongod 实例，存储了整个 Cluster Metadata，其中包括 chunk 信息。
• Route Server: mongos 实例，前端路由，客户端由此接入，且让整个集群看上去像单一进程数据库。

机器IP角色test002192.168.1.2mongod shard11:27017test003192.168.1.3mongod shard21:27017test004192.168.1.4mongod shard31:27017test005192.168.1.5mongod config1:20000
mongs1:30000test006192.168.1.6mongod config2:20000
mongs2:30000test007192.168.1.7mongod config3:20000
mongs3:30000test008192.168.1.8mongod shard12:27017test009192.168.1.9mongod shard22:27017test010192.168.1.10mongod shard32:27017

Shard配置

Shard1

[test002; test008]

test002:

1

./mongod -shardsvr -replSet shard1 -port 27017 -dbpath /mongodb/data/shard11 -oplogSize 10000 -logpath /mongodb/logs/shard11.log -logappend -fork

test008:

1

./mongod -shardsvr -replSet shard1 -port 27017 -dbpath /mongodb/data/shard12 -oplogSize 10000 -logpath /mongodb/logs/shard12.log -logappend -fork

初始化shard1

1

2

3

4

5

config={_id:'shard1',members:[

{_id:0,host:'test002:27017'},

{_id:1,host:'test008:27017'}]

}

rs.initiate(config);

Shard2

[test003; test009]

test003:

1

./mongod -shardsvr -replSet shard2 -port 27017 -dbpath /mongodb/data/shard21 -oplogSize 10000 -logpath /mongodb/logs/shard21.log -logappend -fork

test009:

1

./mongod -shardsvr -replSet shard2 -port 27017 -dbpath /mongodb/data/shard22 -oplogSize 10000 -logpath /mongodb/logs/shard22.log -logappend -fork

初始化shard2

1

2

3

4

5

config={_id:'shard2',members:[

{_id:0,host:'test003:27017'},

{_id:1,host:'test009:27017'}]

}

rs.initiate(config);

Shard3

[test004; test010]

test004:

1

./mongod -shardsvr -replSet shard3 -port 27017 -dbpath /mongodb/data/shard31 -oplogSize 10000 -logpath /mongodb/logs/shard31.log -logappend -fork

test010:

1

./mongod -shardsvr -replSet shard3 -port 27017 -dbpath /mongodb/data/shard32 -oplogSize 10000 -logpath /mongodb/logs/shard32.log -logappend -fork

初始化shard3

1

2

3

4

5

config={_id:'shard3',members:[

{_id:0,host:'test004:27017'},

{_id:1,host:'test010:27017'}]

}

rs.initiate(config);

config server配置

[test005; test006; test007]

1

./mongod -configsvr -dbpath /mongodb/data/config -port 20000 -logpath /mongodb/logs/config.log -logappend -fork

Mongos配置

[test005; test006; test007]

1

./mongos -configdb test005:20000,test006:20000,test007:20000 -port 30000 -chunkSize 5 -logpath /mongodb/logs/mongos.log -logappend -fork

Route 转发请求到实际的目标服务进程，并将多个结果合并回传给客户端。Route 本身并不存储任何数据和状态，仅在启动时从 Config Server 获取信息。Config Server 上的任何变动都会传递给所有的 Route Process。

Configuring the Shard Cluster

1.     连接admin数据库

1

./mongo test005:30000/admin

2.      加入shards

1

2

3

db.runCommand({addshard:"shard1/test002:27017,test008:27017",name:"s1",maxsize:20480});

db.runCommand({addshard:"shard2/test003:27017,test009:27017",name:"s2",maxsize:20480});

db.runCommand({addshard:"shard3/test004:27017,test010:27017",name:"s3",maxsize:20480});

3.      Listing shards

1

db.runCommand({listshards:1})

如果列出了以上3个shards，表示shards已经配置成功

4.      激活数据库和表分片

1

2

db.runCommand({enablesharding:"taobao"});

db.runCommand({shardcollection:"taobao.test0",key:{_id:1}}); db.runCommand({shardcollection:"taobao.test1",key:{_id:1}});

使用

shell操作数据库

超级用户相关：

1)     进入数据库admin

1

use admin

2)     增加或修改用户密码

1

db.addUser('name','pwd')

3)     查看用户列表

1

db.system.users.find()

4)     用户认证

1

db.auth('name','pwd')

5)     删除用户

1

db.removeUser('name')

6)     查看所有用户

1

show users

7)     查看所有数据库

1

show dbs

8)     查看所有的collection

1

show collections

9)     查看各collection的状态

1

db.printCollectionStats()

10)   查看主从复制状态

1

db.printReplicationInfo()

11)   修复数据库

1

db.repairDatabase()

12)   设置记录profiling，0=off 1=slow 2=all

1

db.setProfilingLevel(1)

13)   查看profiling

1

show profile

14)   拷贝数据库

1

db.copyDatabase('mail_addr','mail_addr_tmp')

15)   删除collection

1

db.mail_addr.drop()

16)   删除当前的数据库

1

db.dropDatabase()

增加删除修改:

1) Insert

1

2

3

db.user.insert({'name':'dump','age':1})

or

db.user.save({'name':'dump','age':1})

嵌套对象:

1

db.foo.save({'name':'dump','address':{'city':'hangzhou','post':310015},'phone':[138888888,13999999999]})

数组对象:

1

db.user_addr.save({'Uid':'dump','Al':['[e-mail]','[e-mail]']})

2) delete

删除name=’dump’的用户信息:

1

db.user.remove({'name':'dump'})

删除foo表所有信息:

1

db.foo.remove()

3) update

//update foo set xx=4 where yy=6

//如果不存在则插入，允许修改多条记录

1

db.foo.update({'yy':6},{'$set':{'xx':4}},upsert=true,multi=true)

查询:

1

2

3

4

5

6

7

8

coll.find() // select * from coll

coll.find().limit(10) // select * from coll limit 10

coll.find().sort({x:1}) // select * from coll order by x asc

coll.find().sort({x:1}).skip(5).limit(10) // select * from coll order by x asc limit 5, 10

coll.find({x:10}) // select * from coll where x = 10

coll.find({x: {$lt:10}}) // select * from coll where x <= 10

coll.find({}, {y:true}) // select y from coll

coll.count() //select count(*) from coll

其他:

1

2

3

4

5

coll.find({"address.city":"gz"}) // 搜索嵌套文档address中city值为gz的记录

coll.find({likes:"math"}) // 搜索数组

coll.find({name: {$exists: true}}); //查询所有存在name字段的记录

coll.find({phone: {$exists: false}}); //查询所有不存在phone字段的记录

coll.find({name: {$type: 2}}); //查询所有name字段是字符类型的coll.find({age: {$type: 16}}); //查询所有age字段是整型的

索引:

1(ascending),-1(descending)

1

2

3

4

5

6

7

coll.ensureIndex({productid:1}) // 在productid上建立普通索引

coll.ensureIndex({district:1, plate:1}) // 多字段索引

coll.ensureIndex({"address.city":1}) // 在嵌套文档的字段上建索引

coll.ensureIndex({productid:1}, {unique:true}) // 唯一索引

coll.ensureIndex({productid:1}, {unique:true, dropDups:true|) // 建索引时，如果遇到索引字段值已经出现过的情况，则删除重复记录

coll.getIndexes() // 查看索引

coll.dropIndex({productid:1}) // 删除单个索引

MongoDB Drivers

C

C#

C++

Haskell

Java

Javascript

Perl

PHP

Python

Ruby

Scala (via Casbah)

Mongodb支持的client 编程api非常多，由于dump中心是建立在hadoop的基础上的，所以着重介绍java api,后面的测试程序采用的也是java api.

MongoDB in Java

下载MongoDB的Java驱动，把jar包(mongo-2.3.jar)扔到项目里去就行了，

Java中，Mongo对象是线程安全的，一个应用中应该只使用一个Mongo对象。Mongo对象会自动维护一个连接池，默认连接数为10。

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

import com.mongodb.*

try{

Mongo mg = new Mongo(server_lists);// List<ServerAddress> server _lists

DB db = mg.getDB("taobao");

if (db.isAuthenticated() == false) {

db.authenticate("name", "pwd".toCharArray());

}

DBCollection coll=db.getCollection("category_property_values");

coll.slaveOk();//repl set模式必须调用，否则所有query将只发到主节点查询

//insert

 

BasicDBObject doc = <strong>new</strong> BasicDBObject();

 

//赋值

doc.put("name", "MongoDB");

doc.put("type", "database");

coll.insert(doc);

……

//select

//查询一条数据

BasicDBObject doc = <strong>new</strong> BasicDBObject();

doc.put("name", "MongoDB");

DBObject query = coll.findOne(doc);

……

//使用游标查询

DBCursor cur = coll.find(doc);

while(cur.hasNext()) {

cur.next();

……

}

……

//update

DBObject dblist = new BasicDBObject();

DBObject qlist = new BasicDBObject();

qlist.put("_id", j);

dblist.put("t1", str);

coll.update(qlist, dblist);

……

 

//delete

DBObject dlist = new BasicDBObject();

dlist.put("_id", j);

coll.remove(dlist);

}catch(MongoException ex){

}

MongoDB 测试

测试版本: 1.6.3

采用单线程分别插入100万，300万,500万,1000万数据和多个线程，每线程插入100万数据.

插入数据格式:

1

{ "_id" : NumberLong(16), "nid" : NumberLong(16), "t1" : "search_engine_insert", "t2" : "search_engine_insert", "t3" : "search_engine_insert", "t4" : "search_engine_insert" }

1) Master slaves模式

Insert

Per-thread rowsrun timePer-thread insertTotal-insertTotal rowsthreads1000000205000050000100000013000000605000050000300000015000000995050550505500000018000000159503145031480000001100000002084807648076100000001100000064156253125020000002

Mongodb只有主节点才能进行插入和更新操作.

Update

数据格式:

1

{ "_id" : NumberLong(16), "nid" : NumberLong(16), "t1" : "search_engine_update", "t2" : "search_engine_update", "t3" : "search_engine_update", "t4" : "search_engine_update" }

Per-thread rowsrun timePer-thread updateTotal-updateTotal rowsthreads1000000961041610416100000013000000287104521045230000001100000018853191595730000003100000035128491424550000005

Select

以”_id”字段为key，返回整条记录

a)      客户端:单机多线程

Per-thread rowsrun timePer-thread selectTotal-selectTotal rowsthreads100000072138881388810000001100000012977517751910000000101000000554180590252500000005010000001121892892061000000001001000000225644388652200000000200

b)      客户端:分布式多线程

程序部署在39台机器上

Per-thread rowsrun timePer-thread selectTotal-selectTotal rowsthreads100000017357805780*39=2234701000000*391100000014027137132*39=27814810000000*391050000014063557112*39=27736810000000*392020000014331396978*39=27214210000000*3950

2) Replica Set 模式

Insert

Per-thread rowsrun timePer-thread insertTotal-insertTotal rowsthreads100000040250002500010000001300000011725641256413000000150000002112369623696500000018000000289276812768180000001100000003882577325773100000001100000083120482409620000002100000021047622380950000005

Update

Per-thread rowsrun timePer-thread updateTotal-updateTotal rowsthreads100000028357143571410000001300000083361443614430000001100000014668492054730000003100000026238161908350000005

Select

以”_id”字段为key，返回整条记录

a)      客户端:单机多线程

Per-thread rowsrun timePer-thread selectTotal-selectTotal rowsthreads10000001985050505010000001100000026437873787810000000101000000436229311467850000000501000000754132613262510000000010010000001526655131061200000000200

b)      客户端:分布式多线程

程序部署在39台机器上

Per-thread rowsrun timePer-thread selectTotal-selectTotal rowsthreads100000021646294629*39=1805311000000*391100000013757297293*39=28442710000000*391050000014693406807*39=26547310000000*392020000015611286406*39=24983410000000*3950

3) Sharding 模式

Insert

Per-thread rowsrun timePer-thread insertTotal-insertTotal rowsthreads100000058172411724110000001300000018016666166663000000150000003731340413404500000012000000234854717094400000022000000447447422371100000005

Update

Per-thread rowsrun timePer-thread updateTotal-updateTotal rowsthreads1000000382631526315100000013000000115260862608630000001100000064156254687530000003100000093107525376350000005

Select

以”_id”字段为key，返回整条记录

a)      客户端:单机多线程

Per-thread rowsrun timePer-thread selectTotal-selectTotal rowsthreads1000000277361036101000000110000004562192219291000000010100000011588634317750000000501000000229943443497100000000100

b)      客户端:分布式多线程

程序部署在39台机器上

Per-thread rowsrun timePer-thread selectTotal-selectTotal rowsthreads100000065915171517*39= 591631000000*391100000085401171170*39=4563010000000*3910

小结:

Mongodb在M-S和Repl-Set模式下查询效率还是不错的，区别在于Repl-Set模式如果有primary节点挂掉，系统自己会选举出另一个primary节点，不会影响后续的使用，原来的主节点恢复后自动成为secondary节点,而M-S模式一旦master 节点挂掉需要手工将别的slaves 节点修改成master,另外Repl-Set模式最多只能有7个节点.

由于sharding模式查询速度下降明显，耗时太长,所以只测试了2轮,估计他的威力应该在数据量非常大的环境下才能体现出来吧,以上数据仅供参考，现在只是简单的进行了测试，接下来会对源码进行一下研究，欢迎和感兴趣的同学多多交流！

http://www.searchtb.com/2010/12/a-probe-into-the-mongodb.html