MongoDB复制原理就是主节点记录在其上的所有操作oplog，从节点定期轮询主节点获取这些操作，然后对自己的数据副本执行这些操作，从而保证从节点的数据与主节点一致。
那什么是副本集呢？主从模式一主一从就是一个副本。MongoDB复制提供了数据的冗余备份，并在多个服务器上存储数据副本，提高了数据的可用性， 并可以保证数据的安全性;复制还允许您从硬件故障和服务中断中恢复数据。副本集具有多个副本保证了容错性，就算一个副本挂掉了还有很多副本存在，并且解决了之前的问题“主节点挂掉了，整个集群内会自动切换”。
副本集特征：
· N 个节点的集群
· 任何节点可作为主节点
· 所有写入操作都在主节点上
· 自动故障转移
· 自动恢复
副本集还有以下几个需要注意的地方：
1. 最小构成是：primary，secondary，arbiter，一般部署是：primary，2 secondary。
2. 成员数应该为奇数，如果为偶数的情况下添加arbiter，arbiter不保存数据，只投票。
3. 最大50 members，但是只能有 7 voting members，其他是non-voting members。
借用网上找的两张架构图说明一下：
由图可以看到客户端连接到整个副本集，不关心具体哪一台机器是否挂掉。主服务器负责整个副本集的读写，副本集定期同步数据备份，一但主节点挂掉，副本节点就会选举一个新的主服务器，这一切对于应用服务器不需要关心。我们看一下主服务器挂掉后的架构：
副本集中的副本节点在主节点挂掉后通过心跳机制检测到后，就会在集群内发起主节点的选举机制，自动选举一位新的主服务器。更加详细的副本机制请参考搭建高可用mongodb集群（三）—— 深入副本集内部机制
官方推荐的副本集机器数量至少为3个，我这里就以三个节点为例介绍。
1. 环境准备
192.168.221.161:27017作为主节点，192.168.221.161:27018,192.168.221.161:27019作为两个副本节点
2. 分别创建三个节点所需要的文件夹
mkdir /data/db{1..3} #存放三个节点数据文件
touch /usr/local/mongodb/mongodb{1..3}.log #存放三个节点的日志
3. 创建三个节点的配置文件
[root@MidApp mongodb]# cat mongodb1.conf
dbpath=/data/db1
logpath=/usr/local/mongodb/logs/mongodb1.log
logappend=true
port=27017
fork=true
auth=false
nohttpinterface=false
bind_ip=192.168.221.161
journal=true
quiet=true
replSet=repset
[root@MidApp mongodb]# cat mongodb2.conf
dbpath=/data/db2
logpath=/usr/local/mongodb/logs/mongodb2.log
logappend=true
port=27018
fork=true
auth=false
nohttpinterface=false
bind_ip=192.168.221.161
journal=true
quiet=true
replSet=repset
[root@MidApp mongodb]# cat mongodb3.conf
dbpath=/data/db3
logpath=/usr/local/mongodb/logs/mongodb3.log
logappend=true
port=27019
fork=true
auth=false
nohttpinterface=false
bind_ip=192.168.221.161
journal=true
quiet=true
replSet=repset
4. 分别启动三个节点
mongod -f mongodb1.conf
mongod -f mongodb2.conf
mongod -f mongodb3.conf
查看日志输出，没有找到replica的配置信息
2017-11-13T16:34:46.212-0800 I CONTROL [initandlisten] options: { config: "mongodb1.conf", net: { bindIp: "192.168.221.160", http: { enabled: true }, port: 27017 }, processManagement: { fork: true }, replication: { replSet: "repset" }, security: { authorization: "disabled" }, storage: { dbPath: "/data/db1", journal: { enabled: true } }, systemLog: { destination: "file", logAppend: true, path: "/usr/local/mongodb/logs/mongodb1.log", quiet: true } }
2017-11-13T16:34:46.214-0800 I INDEX [initandlisten] allocating new ns file /data/db1/local.ns, filling with zeroes...
2017-11-13T16:34:46.216-0800 I JOURNAL [journal writer] Journal writer thread started
2017-11-13T16:34:46.492-0800 I STORAGE [FileAllocator] allocating new datafile /data/db1/local.0, filling with zeroes...
2017-11-13T16:34:46.492-0800 I STORAGE [FileAllocator] creating directory /data/db1/_tmp
2017-11-13T16:34:46.501-0800 I STORAGE [FileAllocator] done allocating datafile /data/db1/local.0, size: 64MB, took 0.001 secs
2017-11-13T16:34:46.520-0800 I REPL [initandlisten] Did not find local replica set configuration document at startup; NoMatchingDocument Did not find replica set configuration document in local.system.replset
5. 初始化副本集
任选一个节点登录进去，可以看到show dbs是不可用，因为还没初始化副本集
[root@MidApp mongodb]# mongo 192.168.221.161
MongoDB shell version: 3.0.6
connecting to: 192.168.221.161/test
Welcome to the MongoDB shell.
For interactive help, type "help".
For more comprehensive documentation, see
Questions? Try the support group
> show dbs
2017-11-13T16:39:33.235-0800 E QUERY Error: listDatabases failed:{ "note" : "from execCommand", "ok" : 0, "errmsg" : "not master" }
at Error (<anonymous>)
at Mongo.getDBs (src/mongo/shell/mongo.js:47:15)
at shellHelper.show (src/mongo/shell/utils.js:630:33)
at shellHelper (src/mongo/shell/utils.js:524:36)
at (shellhelp2):1:1 at src/mongo/shell/mongo.js:47
> use admin
switched to db admin
定义副本集配置变量，这里的 _id:”repset” 和上面配置文件的参数“ –replSet repset” 要保持一样
>config= {_id:"repset",members:[
{_id:0,host:"192.168.221.161:27017"},
{_id:1,host:"192.168.221.161:27018"},
{_id:2,host:"192.168.221.161:27019"}]
}
查看输出信息：
{
"_id" : "repset",
"members" : [
{
"_id" : 0,
"host" : "192.168.221.161:27017"
},
{
"_id" : 1,
"host" : "192.168.221.161:27018"
},
{
"_id" : 2,
"host" : "192.168.221.161:27019"
}
]
}
初始化副本集配置
12 > rs.initiate(config);
{ "ok" : 1 }
查看集群节点状态：
repset:OTHER> rs.status()
{
"set" : "repset",
"date" : ISODate("2017-11-14T00:49:25.782Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "192.168.221.161:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 695,
"optime" : Timestamp(1510620540, 1),
"optimeDate" : ISODate("2017-11-14T00:49:00Z"),
"electionTime" : Timestamp(1510620544, 1),
"electionDate" : ISODate("2017-11-14T00:49:04Z"),
"configVersion" : 1,
"self" : true
},
{
"_id" : 1,
"name" : "192.168.221.161:27018",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 25,
"optime" : Timestamp(1510620540, 1),
"optimeDate" : ISODate("2017-11-14T00:49:00Z"),
"lastHeartbeat" : ISODate("2017-11-14T00:49:24.739Z"),
"lastHeartbeatRecv" : ISODate("2017-11-14T00:49:24.762Z"),
"pingMs" : 0,
"configVersion" : 1
},
{
"_id" : 2,
"name" : "192.168.221.161:27019",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 25,
"optime" : Timestamp(1510620540, 1),
"optimeDate" : ISODate("2017-11-14T00:49:00Z"),
"lastHeartbeat" : ISODate("2017-11-14T00:49:24.739Z"),
"lastHeartbeatRecv" : ISODate("2017-11-14T00:49:24.762Z"),
"pingMs" : 0,
"configVersion" : 1
}
],
"ok" : 1
}
整个副本及已经搭建成功了！
6. 测试副本集数据复制功能
在主节点登录mongodb
[root@MidApp mongodb]# mongo 192.168.221.161:27017
MongoDB shell version: 3.0.6
connecting to: 192.168.221.161:27017/test
repset:PRIMARY> use test;
switched to db test
repset:PRIMARY> db.testdb.insert({"test1":"item1"})
WriteResult({ "nInserted" : 1 })
repset:PRIMARY> exit
bye
在副本节点登录查看数据：
[root@MidApp logs]# mongo 192.168.221.161:27018
MongoDB shell version: 3.0.6
connecting to: 192.168.221.161:27018/test
repset:SECONDARY> use test
switched to db test
repset:SECONDARY> db.getMongo().setSlaveOk();#mongodb默认是从主节点读写数据的，副本节点上不允许读，需要设置副本节点可以读
repset:SECONDARY> db.testdb.find()
{ "_id" : ObjectId("5a0a3e8d40637405ab003b39"), "test1" : "item1" }
7. 测试副本集failover功能
手动把现在的主节点停掉，查看集群状态：
repset:SECONDARY> rs.status()
{
"set" : "repset",
"date" : ISODate("2017-11-14T00:57:14.753Z"),
"myState" : 2,
"members" : [
{
"_id" : 0,
"name" : "192.168.221.161:27017",
"health" : 0,
"state" : 8,
"stateStr" : "(not reachable/healthy)",
"uptime" : 0,
"optime" : Timestamp(0, 0),
"optimeDate" : ISODate("1970-01-01T00:00:00Z"),
"lastHeartbeat" : ISODate("2017-11-14T00:57:13.304Z"),
"lastHeartbeatRecv" : ISODate("2017-11-14T00:57:09.285Z"),
"pingMs" : 0,
"lastHeartbeatMessage" : "Failed attempt to connect to 192.168.221.161:27017; couldn't connect to server 192.168.221.161:27017 (192.168.221.161), connection attempt failed",
"configVersion" : -1
},
{
"_id" : 1,
"name" : "192.168.221.161:27018",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 1167,
"optime" : Timestamp(1510620813, 2),
"optimeDate" : ISODate("2017-11-14T00:53:33Z"),
"infoMessage" : "could not find member to sync from",
"configVersion" : 1,
"self" : true
},
{
"_id" : 2,
"name" : "192.168.221.161:27019",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 17,
"optime" : Timestamp(1510620813, 2),
"optimeDate" : ISODate("2017-11-14T00:53:33Z"),
"lastHeartbeat" : ISODate("2017-11-14T00:57:13.317Z"),
"lastHeartbeatRecv" : ISODate("2017-11-14T00:57:13.330Z"),
"pingMs" : 0,
"electionTime" : Timestamp(1510621032, 1),
"electionDate" : ISODate("2017-11-14T00:57:12Z"),
"configVersion" : 1
}
],
"ok" : 1
}
可以看到原来的主节点状态现在已经变成了不可达，192.168.221.161:27019已经变成新的主节点
8. 再次启动原来的主节点，发现192.168.221.161:27017还是SECONDARY
repset:SECONDARY> rs.status()
{
"set" : "repset",
"date" : ISODate("2017-11-14T17:02:13.837Z"),
"myState" : 2,
"members" : [
{
"_id" : 0,
"name" : "192.168.221.161:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 5,
"optime" : Timestamp(1510620813, 2),
"optimeDate" : ISODate("2017-11-14T00:53:33Z"),
"lastHeartbeat" : ISODate("2017-11-14T17:02:12.642Z"),
"lastHeartbeatRecv" : ISODate("2017-11-14T17:02:11.902Z"),
"pingMs" : 0,
"configVersion" : 1
},
{
"_id" : 1,
"name" : "192.168.221.161:27018",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 59066,
"optime" : Timestamp(1510620813, 2),
"optimeDate" : ISODate("2017-11-14T00:53:33Z"),
"configVersion" : 1,
"self" : true
},
{
"_id" : 2,
"name" : "192.168.221.161:27019",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 57916,
"optime" : Timestamp(1510620813, 2),
"optimeDate" : ISODate("2017-11-14T00:53:33Z"),
"lastHeartbeat" : ISODate("2017-11-14T17:02:12.308Z"),
"lastHeartbeatRecv" : ISODate("2017-11-14T17:02:12.080Z"),
"pingMs" : 0,
"electionTime" : Timestamp(1510621032, 1),
"electionDate" : ISODate("2017-11-14T00:57:12Z"),
"configVersion" : 1
}
],
"ok" : 1
}
现在，mongodb的副本集群已经完成了。
接下来是关于mongodb的读写分离和分片的内容