MongoDB分片搭建

分片（sharding）是指将数据库拆分，将其分散在不同的机器上的过程。将数据分散到不同的机器上，不需要功能强大的服务器就可以存储更多的数据和处理更大的负载。基本思想就是将集合切成小块，这些块分散到若干片里，每个片只负责总数据的一部分，最后通过一个均衡器来对各个分片进行均衡（数据迁移）。通过一个名为mongos的路由进程进行操作，mongos知道数据和片的对应关系（通过配置服务器）。大部分使用场景都是解决磁盘空间的问题，对于写入有可能会变差，查询则尽量避免跨分片查询。使用分片的时机：

1，机器的磁盘不够用了。使用分片解决磁盘空间的问题。
2，单个mongod已经不能满足写数据的性能要求。通过分片让写压力分散到各个分片上面，使用分片服务器自身的资源。
3，想把大量数据放到内存里提高性能。和上面一样，通过分片使用分片服务器自身的资源。

所以简单来说分片就是将数据库进行拆分，将大型集合分隔到不同服务器上，所以组成分片的单元是副本集。比如，本来100G的数据，可以分割成10份存储到10台服务器上，这样每台机器只有10G的数据，一般分片在大型企业或者数据量很大的公司才会使用。

MongoDB通过一个mongos的进程（路由分发）实现分片后的数据存储与访问，也就是说mongos是整个分片架构的核心，是分片的总入口，对客户端而言是不知道是否有分片的，客户端只需要把读写操作转达给mongos即可。

虽然分片会把数据分隔到很多台服务器上，但是每一个节点都是需要有一个备用角色的，这样才能保证数据的高可用。

当系统需要更多空间或者资源的时候，分片可以让我们按照需求方便的横向扩展，只需要把mongodb服务的机器加入到分片集群中即可

MongoDB分片架构图：

MongoDB分片相关概念：

• mongos： 数据库集群请求的入口，所有的请求都通过mongos进行协调，不需要在应用程序添加一个路由选择器，mongos自己就是一个请求分发中心，它负责把对应的数据请求请求转发到对应的shard服务器上。在生产环境通常有多mongos作为请求的入口，防止其中一个挂掉所有的mongodb请求都没有办法操作。

• config server： 配置服务器，存储所有数据库元信息（路由、分片）的配置。mongos本身没有物理存储分片服务器和数据路由信息，只是缓存在内存里，配置服务器则实际存储这些数据。mongos第一次启动或者关掉重启就会从 config server 加载配置信息，以后如果配置服务器信息变化会通知到所有的 mongos 更新自己的状态，这样 mongos 就能继续准确路由。在生产环境通常有多个 config server 配置服务器，因为它存储了分片路由的元数据，防止数据丢失！

• shard： 存储了一个集合部分数据的MongoDB实例，每个分片是单独的mongodb服务或者副本集，在生产环境中，所有的分片都应该是副本集。

21.37/21.38/21.39 mongodb分片搭建

分片搭建 -服务器规划：

资源有限，我这里使用三台机器 A B C 作为演示：

• A搭建：mongos、config server、副本集1主节点、副本集2仲裁、副本集3从节点
• B搭建：mongos、config server、副本集1从节点、副本集2主节点、副本集3仲裁
• C搭建：mongos、config server、副本集1仲裁、副本集2从节点、副本集3主节点
• 端口分配：mongos 20000、config server 21000、副本集1 27001、副本集2 27002、副本集3 27003
• 三台机器全部关闭firewalld服务和selinux，或者增加对应端口的规则

三台机器的IP分别是：
A机器：192.168.77.128
B机器：192.168.77.130
C机器：192.168.77.134

分片搭建 – 创建目录：

分别在三台机器上创建各个角色所需要的目录：

mkdir -p /data/mongodb/mongos/log
mkdir -p /data/mongodb/config/{data,log}
mkdir -p /data/mongodb/shard1/{data,log}
mkdir -p /data/mongodb/shard2/{data,log}
mkdir -p /data/mongodb/shard3/{data,log}

分片搭建–config server配置：

mongodb3.4版本以后需要对config server创建副本集
添加配置文件(三台机器都操作)

[root@localhost ~]# mkdir /etc/mongod/
[root@localhost ~]# vim /etc/mongod/config.conf  # 加入如下内容
pidfilepath = /var/run/mongodb/configsrv.pid
dbpath = /data/mongodb/config/data
logpath = /data/mongodb/config/log/congigsrv.log
logappend = true
bind_ip = 0.0.0.0  # 绑定你的监听ip
port = 21000
fork = true
configsvr = true #declare this is a config db of a cluster;
replSet=configs #副本集名称
maxConns=20000 #设置最大连接数

启动三台机器的config server：

[root@localhost ~]# mongod -f /etc/mongod/config.conf  # 三台机器都要操作
about to fork child process, waiting until server is ready for connections.
forked process: 4183
child process started successfully, parent exiting
[root@localhost ~]# ps aux |grep mongo
mongod     2518  1.1  2.3 1544488 89064 ?       Sl   09:57   0:42 /usr/bin/mongod -f /etc/mongod.conf
root       4183  1.1  1.3 1072404 50992 ?       Sl   10:56   0:00 mongod -f /etc/mongod/config.conf
root       4240  0.0  0.0 112660   964 pts/0    S+   10:57   0:00 grep --color=auto mongo
[root@localhost ~]# netstat -lntp |grep mongod
tcp        0      0 192.168.77.128:21000    0.0.0.0:*               LISTEN      4183/mongod         
tcp        0      0 192.168.77.128:27017    0.0.0.0:*               LISTEN      2518/mongod         
tcp        0      0 127.0.0.1:27017         0.0.0.0:*               LISTEN      2518/mongod         
[root@localhost ~]#

登录任意一台机器的21000端口，初始化副本集：

[root@localhost ~]# mongo --host 192.168.77.128 --port 21000
> config = { _id: "configs", members: [ {_id : 0, host : "192.168.77.128:21000"},{_id : 1, host : "192.168.77.130:21000"},{_id : 2, host : "192.168.77.134:21000"}] }
{
    "_id" : "configs",
    "members" : [
        {
            "_id" : 0,
            "host" : "192.168.77.128:21000"
        },
        {
            "_id" : 1,
            "host" : "192.168.77.130:21000"
        },
        {
            "_id" : 2,
            "host" : "192.168.77.134:21000"
        }
    ]
}
> rs.initiate(config)  # 初始化副本集
{
    "ok" : 1,
    "operationTime" : Timestamp(1515553318, 1),
    "$gleStats" : {
        "lastOpTime" : Timestamp(1515553318, 1),
        "electionId" : ObjectId("000000000000000000000000")
    },
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515553318, 1),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    }
}
configs:SECONDARY> rs.status()  # 确保每台机器都正常
{
    "set" : "configs",
    "date" : ISODate("2018-01-10T03:03:40.244Z"),
    "myState" : 1,
    "term" : NumberLong(1),
    "configsvr" : true,
    "heartbeatIntervalMillis" : NumberLong(2000),
    "optimes" : {
        "lastCommittedOpTime" : {
            "ts" : Timestamp(1515553411, 1),
            "t" : NumberLong(1)
        },
        "readConcernMajorityOpTime" : {
            "ts" : Timestamp(1515553411, 1),
            "t" : NumberLong(1)
        },
        "appliedOpTime" : {
            "ts" : Timestamp(1515553411, 1),
            "t" : NumberLong(1)
        },
        "durableOpTime" : {
            "ts" : Timestamp(1515553411, 1),
            "t" : NumberLong(1)
        }
    },
    "members" : [
        {
            "_id" : 0,
            "name" : "192.168.77.128:21000",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 415,
            "optime" : {
                "ts" : Timestamp(1515553411, 1),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T03:03:31Z"),
            "infoMessage" : "could not find member to sync from",
            "electionTime" : Timestamp(1515553329, 1),
            "electionDate" : ISODate("2018-01-10T03:02:09Z"),
            "configVersion" : 1,
            "self" : true
        },
        {
            "_id" : 1,
            "name" : "192.168.77.130:21000",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 101,
            "optime" : {
                "ts" : Timestamp(1515553411, 1),
                "t" : NumberLong(1)
            },
            "optimeDurable" : {
                "ts" : Timestamp(1515553411, 1),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T03:03:31Z"),
            "optimeDurableDate" : ISODate("2018-01-10T03:03:31Z"),
            "lastHeartbeat" : ISODate("2018-01-10T03:03:39.973Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T03:03:38.804Z"),
            "pingMs" : NumberLong(0),
            "syncingTo" : "192.168.77.134:21000",
            "configVersion" : 1
        },
        {
            "_id" : 2,
            "name" : "192.168.77.134:21000",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 101,
            "optime" : {
                "ts" : Timestamp(1515553411, 1),
                "t" : NumberLong(1)
            },
            "optimeDurable" : {
                "ts" : Timestamp(1515553411, 1),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T03:03:31Z"),
            "optimeDurableDate" : ISODate("2018-01-10T03:03:31Z"),
            "lastHeartbeat" : ISODate("2018-01-10T03:03:39.945Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T03:03:38.726Z"),
            "pingMs" : NumberLong(0),
            "syncingTo" : "192.168.77.128:21000",
            "configVersion" : 1
        }
    ],
    "ok" : 1,
    "operationTime" : Timestamp(1515553411, 1),
    "$gleStats" : {
        "lastOpTime" : Timestamp(1515553318, 1),
        "electionId" : ObjectId("7fffffff0000000000000001")
    },
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515553411, 1),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    }
}
configs:PRIMARY>

分片搭建–分片配置：

添加配置文件(三台机器都需要操作)：

[root@localhost ~]# vim /etc/mongod/shard1.conf
pidfilepath = /var/run/mongodb/shard1.pid
dbpath = /data/mongodb/shard1/data
logpath = /data/mongodb/shard1/log/shard1.log
logappend = true
logRotate=rename
bind_ip = 0.0.0.0  # 绑定你的监听IP
port = 27001
fork = true
replSet=shard1 #副本集名称
shardsvr = true #declare this is a shard db of a cluster;
maxConns=20000 #设置最大连接数

[root@localhost ~]# vim /etc/mongod/shard2.conf //加入如下内容
pidfilepath = /var/run/mongodb/shard2.pid
dbpath = /data/mongodb/shard2/data
logpath = /data/mongodb/shard2/log/shard2.log
logappend = true
logRotate=rename
bind_ip = 0.0.0.0  # 绑定你的监听IP
port = 27002
fork = true
replSet=shard2 #副本集名称
shardsvr = true #declare this is a shard db of a cluster;
maxConns=20000 #设置最大连接数

[root@localhost ~]# vim /etc/mongod/shard3.conf //加入如下内容
pidfilepath = /var/run/mongodb/shard3.pid
dbpath = /data/mongodb/shard3/data
logpath = /data/mongodb/shard3/log/shard3.log
logappend = true
logRotate=rename
bind_ip = 0.0.0.0  # 绑定你的监听IP
port = 27003
fork = true
replSet=shard3 #副本集名称
shardsvr = true #declare this is a shard db of a cluster;
maxConns=20000 #设置最大连接数

都配置完成之后逐个进行启动，三台机器都需要启动：

1.先启动shard1：

[root@localhost ~]# mongod -f /etc/mongod/shard1.conf  # 三台机器都要操作
about to fork child process, waiting until server is ready for connections.
forked process: 13615
child process started successfully, parent exiting
[root@localhost ~]# ps aux |grep shard1
root      13615  0.7  1.3 1023224 52660 ?       Sl   17:16   0:00 mongod -f /etc/mongod/shard1.conf
root      13670  0.0  0.0 112660   964 pts/0    R+   17:17   0:00 grep --color=auto shard1
[root@localhost ~]#

然后登录128或者130机器的27001端口初始化副本集，134之所以不行，是因为shard1我们把134这台机器的27001端口作为了仲裁节点：

[root@localhost ~]# mongo --host 192.168.77.128 --port 27001
> use admin
switched to db admin
> config = { _id: "shard1", members: [ {_id : 0, host : "192.168.77.128:27001"}, {_id: 1,host : "192.168.77.130:27001"},{_id : 2, host : "192.168.77.134:27001",arbiterOnly:true}] }
{
    "_id" : "shard1",
    "members" : [
        {
            "_id" : 0,
            "host" : "192.168.77.128:27001"
        },
        {
            "_id" : 1,
            "host" : "192.168.77.130:27001"
        },
        {
            "_id" : 2,
            "host" : "192.168.77.134:27001",
            "arbiterOnly" : true
        }
    ]
}
> rs.initiate(config)  # 初始化副本集
{ "ok" : 1 }
shard1:SECONDARY> rs.status()  # 查看状态
{
    "set" : "shard1",
    "date" : ISODate("2018-01-10T09:21:37.682Z"),
    "myState" : 1,
    "term" : NumberLong(1),
    "heartbeatIntervalMillis" : NumberLong(2000),
    "optimes" : {
        "lastCommittedOpTime" : {
            "ts" : Timestamp(1515576097, 1),
            "t" : NumberLong(1)
        },
        "readConcernMajorityOpTime" : {
            "ts" : Timestamp(1515576097, 1),
            "t" : NumberLong(1)
        },
        "appliedOpTime" : {
            "ts" : Timestamp(1515576097, 1),
            "t" : NumberLong(1)
        },
        "durableOpTime" : {
            "ts" : Timestamp(1515576097, 1),
            "t" : NumberLong(1)
        }
    },
    "members" : [
        {
            "_id" : 0,
            "name" : "192.168.77.128:27001",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 317,
            "optime" : {
                "ts" : Timestamp(1515576097, 1),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T09:21:37Z"),
            "infoMessage" : "could not find member to sync from",
            "electionTime" : Timestamp(1515576075, 1),
            "electionDate" : ISODate("2018-01-10T09:21:15Z"),
            "configVersion" : 1,
            "self" : true
        },
        {
            "_id" : 1,
            "name" : "192.168.77.130:27001",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 33,
            "optime" : {
                "ts" : Timestamp(1515576097, 1),
                "t" : NumberLong(1)
            },
            "optimeDurable" : {
                "ts" : Timestamp(1515576097, 1),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T09:21:37Z"),
            "optimeDurableDate" : ISODate("2018-01-10T09:21:37Z"),
            "lastHeartbeat" : ISODate("2018-01-10T09:21:37.262Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T09:21:36.213Z"),
            "pingMs" : NumberLong(0),
            "syncingTo" : "192.168.77.128:27001",
            "configVersion" : 1
        },
        {
            "_id" : 2,
            "name" : "192.168.77.134:27001",
            "health" : 1,
            "state" : 7,
            "stateStr" : "ARBITER",  # 可以看到134是仲裁节点
            "uptime" : 33,
            "lastHeartbeat" : ISODate("2018-01-10T09:21:37.256Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T09:21:36.024Z"),
            "pingMs" : NumberLong(0),
            "configVersion" : 1
        }
    ],
    "ok" : 1
}
shard1:PRIMARY>

2.shard1配置完毕之后启动shard2：

[root@localhost ~]# mongod -f /etc/mongod/shard2.conf   # 三台机器都要进行启动操作
about to fork child process, waiting until server is ready for connections.
forked process: 13910
child process started successfully, parent exiting
[root@localhost ~]# ps aux |grep shard2
root      13910  1.9  1.2 1023224 50096 ?       Sl   17:25   0:00 mongod -f /etc/mongod/shard2.conf
root      13943  0.0  0.0 112660   964 pts/0    S+   17:25   0:00 grep --color=auto shard2
[root@localhost ~]#

登录130或者134任何一台机器的27002端口初始化副本集，128之所以不行，是因为shard2我们把128这台机器的27002端口作为了仲裁节点：

[root@localhost ~]# mongo --host 192.168.77.130 --port 27002
> use admin
switched to db admin
> config = { _id: "shard2", members: [ {_id : 0, host : "192.168.77.128:27002" ,arbiterOnly:true},{_id : 1, host : "192.168.77.130:27002"},{_id : 2, host : "192.168.77.134:27002"}] }
{
    "_id" : "shard2",
    "members" : [
        {
            "_id" : 0,
            "host" : "192.168.77.128:27002",
            "arbiterOnly" : true
        },
        {
            "_id" : 1,
            "host" : "192.168.77.130:27002"
        },
        {
            "_id" : 2,
            "host" : "192.168.77.134:27002"
        }
    ]
}
> rs.initiate(config)
{ "ok" : 1 }
shard2:SECONDARY> rs.status()
{
    "set" : "shard2",
    "date" : ISODate("2018-01-10T17:26:12.250Z"),
    "myState" : 1,
    "term" : NumberLong(1),
    "heartbeatIntervalMillis" : NumberLong(2000),
    "optimes" : {
        "lastCommittedOpTime" : {
            "ts" : Timestamp(1515605171, 1),
            "t" : NumberLong(1)
        },
        "readConcernMajorityOpTime" : {
            "ts" : Timestamp(1515605171, 1),
            "t" : NumberLong(1)
        },
        "appliedOpTime" : {
            "ts" : Timestamp(1515605171, 1),
            "t" : NumberLong(1)
        },
        "durableOpTime" : {
            "ts" : Timestamp(1515605171, 1),
            "t" : NumberLong(1)
        }
    },
    "members" : [
        {
            "_id" : 0,
            "name" : "192.168.77.128:27002",
            "health" : 1,
            "state" : 7,
            "stateStr" : "ARBITER",  # 仲裁节点
            "uptime" : 42,
            "lastHeartbeat" : ISODate("2018-01-10T17:26:10.792Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T17:26:11.607Z"),
            "pingMs" : NumberLong(0),
            "configVersion" : 1
        },
        {
            "_id" : 1,
            "name" : "192.168.77.130:27002",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",  # 主节点
            "uptime" : 546,
            "optime" : {
                "ts" : Timestamp(1515605171, 1),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T17:26:11Z"),
            "infoMessage" : "could not find member to sync from",
            "electionTime" : Timestamp(1515605140, 1),
            "electionDate" : ISODate("2018-01-10T17:25:40Z"),
            "configVersion" : 1,
            "self" : true
        },
        {
            "_id" : 2,
            "name" : "192.168.77.134:27002",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",  # 从节点
            "uptime" : 42,
            "optime" : {
                "ts" : Timestamp(1515605161, 1),
                "t" : NumberLong(1)
            },
            "optimeDurable" : {
                "ts" : Timestamp(1515605161, 1),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T17:26:01Z"),
            "optimeDurableDate" : ISODate("2018-01-10T17:26:01Z"),
            "lastHeartbeat" : ISODate("2018-01-10T17:26:10.776Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T17:26:10.823Z"),
            "pingMs" : NumberLong(0),
            "syncingTo" : "192.168.77.130:27002",
            "configVersion" : 1
        }
    ],
    "ok" : 1
}
shard2:PRIMARY>

3.接着启动shard3：

[root@localhost ~]# mongod -f /etc/mongod/shard3.conf   # 三台机器都要操作
about to fork child process, waiting until server is ready for connections.
forked process: 14204
child process started successfully, parent exiting
[root@localhost ~]# ps aux |grep shard3
root      14204  2.2  1.2 1023228 50096 ?       Sl   17:36   0:00 mongod -f /etc/mongod/shard3.conf
root      14237  0.0  0.0 112660   960 pts/0    S+   17:36   0:00 grep --color=auto shard3
[root@localhost ~]#

然后登录128或者134任何一台机器的27003端口初始化副本集，130之所以不行，是因为shard3我们把130这台机器的27003端口作为了仲裁节点：

[root@localhost ~]# mongo --host 192.168.77.128 --port 27003
> use admin
switched to db admin
> config = { _id: "shard3", members: [ {_id : 0, host : "192.168.77.128:27003"},  {_id : 1, host : "192.168.77.130:27003", arbiterOnly:true}, {_id : 2, host : "192.168.77.134:27003"}] }
{
    "_id" : "shard3",
    "members" : [
        {
            "_id" : 0,
            "host" : "192.168.77.128:27003"
        },
        {
            "_id" : 1,
            "host" : "192.168.77.130:27003",
            "arbiterOnly" : true
        },
        {
            "_id" : 2,
            "host" : "192.168.77.134:27003"
        }
    ]
}
> rs.initiate(config)
{ "ok" : 1 }
shard3:SECONDARY> rs.status()
{
    "set" : "shard3",
    "date" : ISODate("2018-01-10T09:39:47.530Z"),
    "myState" : 1,
    "term" : NumberLong(1),
    "heartbeatIntervalMillis" : NumberLong(2000),
    "optimes" : {
        "lastCommittedOpTime" : {
            "ts" : Timestamp(1515577180, 2),
            "t" : NumberLong(1)
        },
        "readConcernMajorityOpTime" : {
            "ts" : Timestamp(1515577180, 2),
            "t" : NumberLong(1)
        },
        "appliedOpTime" : {
            "ts" : Timestamp(1515577180, 2),
            "t" : NumberLong(1)
        },
        "durableOpTime" : {
            "ts" : Timestamp(1515577180, 2),
            "t" : NumberLong(1)
        }
    },
    "members" : [
        {
            "_id" : 0,
            "name" : "192.168.77.128:27003",
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",  # 主节点
            "uptime" : 221,
            "optime" : {
                "ts" : Timestamp(1515577180, 2),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T09:39:40Z"),
            "infoMessage" : "could not find member to sync from",
            "electionTime" : Timestamp(1515577179, 1),
            "electionDate" : ISODate("2018-01-10T09:39:39Z"),
            "configVersion" : 1,
            "self" : true
        },
        {
            "_id" : 1,
            "name" : "192.168.77.130:27003",
            "health" : 1,
            "state" : 7,
            "stateStr" : "ARBITER",  # 仲裁节点
            "uptime" : 18,
            "lastHeartbeat" : ISODate("2018-01-10T09:39:47.477Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T09:39:45.715Z"),
            "pingMs" : NumberLong(0),
            "configVersion" : 1
        },
        {
            "_id" : 2,
            "name" : "192.168.77.134:27003",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",  # 从节点
            "uptime" : 18,
            "optime" : {
                "ts" : Timestamp(1515577180, 2),
                "t" : NumberLong(1)
            },
            "optimeDurable" : {
                "ts" : Timestamp(1515577180, 2),
                "t" : NumberLong(1)
            },
            "optimeDate" : ISODate("2018-01-10T09:39:40Z"),
            "optimeDurableDate" : ISODate("2018-01-10T09:39:40Z"),
            "lastHeartbeat" : ISODate("2018-01-10T09:39:47.477Z"),
            "lastHeartbeatRecv" : ISODate("2018-01-10T09:39:45.779Z"),
            "pingMs" : NumberLong(0),
            "syncingTo" : "192.168.77.128:27003",
            "configVersion" : 1
        }
    ],
    "ok" : 1
}
shard3:PRIMARY>

分片搭建–配置路由服务器

mongos放在最后面配置是因为它需要知道作为config server的是哪个机器，以及作为shard副本集的机器。

1添加配置文件(三台机器都操作)：

[root@localhost ~]# vim /etc/mongod/mongos.conf  # 加入如下内容
pidfilepath = /var/run/mongodb/mongos.pid
logpath = /data/mongodb/mongos/log/mongos.log
logappend = true
bind_ip = 0.0.0.0  # 绑定你的监听ip
port = 20000
fork = true

#监听的配置服务器,只能有1个或者3个，configs为配置服务器的副本集名字
configdb = configs/192.168.77.128:21000, 192.168.77.130:21000, 192.168.77.134:21000 
maxConns=20000 #设置最大连接数

2.然后三台机器上都启动mongos服务，注意命令，前面都是mongod，这里是mongos：

[root@localhost ~]# mongos -f /etc/mongod/mongos.conf   # 三台机器上都需要执行
2018-01-10T18:26:02.566+0800 I NETWORK  [main] getaddrinfo(" 192.168.77.130") failed: Name or service not known
2018-01-10T18:26:22.583+0800 I NETWORK  [main] getaddrinfo(" 192.168.77.134") failed: Name or service not known
about to fork child process, waiting until server is ready for connections.
forked process: 15552
child process started successfully, parent exiting
[root@localhost ~]# ps aux |grep mongos  # 三台机器上都需要检查进程是否已启动
root      15552  0.2  0.3 279940 15380 ?        Sl   18:26   0:00 mongos -f /etc/mongod/mongos.conf
root      15597  0.0  0.0 112660   964 pts/0    S+   18:27   0:00 grep --color=auto mongos
[root@localhost ~]# netstat -lntp |grep mongos  # 三台机器上都需要检查端口是否已监听
tcp        0      0 0.0.0.0:20000           0.0.0.0:*               LISTEN      15552/mongos        
[root@localhost ~]#

分片搭建–启用分片

1.登录任意一台机器的20000端口，然后把所有分片和路由器串联：

[root@localhost ~]# mongo --host 192.168.77.128 --port 20000
# 串联shard1
mongos> sh.addShard("shard1/192.168.77.128:27001,192.168.77.130:27001,192.168.77.134:27001")
{
    "shardAdded" : "shard1",  # 这里得对应的是shard1才行
    "ok" : 1,  # 注意，这里得是1才是成功
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515580345, 6),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515580345, 6)
}

# 串联shard2
mongos> sh.addShard("shard2/192.168.77.128:27002,192.168.77.130:27002,192.168.77.134:27002")
{
    "shardAdded" : "shard2",   # 这里得对应的是shard2才行
    "ok" : 1,   # 注意，这里得是1才是成功
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515608789, 6),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515608789, 6)
}

# 串联shard3
mongos> sh.addShard("shard3/192.168.77.128:27003,192.168.77.130:27003,192.168.77.134:27003")
{
    "shardAdded" : "shard3",  # 这里得对应的是shard3才行
    "ok" : 1,   # 注意，这里得是1才是成功
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515608789, 14),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515608789, 14)
}
mongos>

使用sh.status()命令查询分片状态，要确认状态正常：

mongos> sh.status()
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "minCompatibleVersion" : 5,
    "currentVersion" : 6,
    "clusterId" : ObjectId("5a55823348aee75ba3928fea")
  }
  shards:  # 成功的情况下，这里会列出分片信息和状态，state的值要为1
        {  "_id" : "shard1",  "host" : "shard1/192.168.77.128:27001,192.168.77.130:27001",  "state" : 1 }
        {  "_id" : "shard2",  "host" : "shard2/192.168.77.130:27002,192.168.77.134:27002",  "state" : 1 }
        {  "_id" : "shard3",  "host" : "shard3/192.168.77.128:27003,192.168.77.134:27003",  "state" : 1 }
  active mongoses:
        "3.6.1" : 1
  autosplit:
        Currently enabled: yes  # 成功的情况下，这里是yes
  balancer:
        Currently enabled:  yes  # 成功的情况下，这里是yes
        Currently running:  no   # 没有创建库和表的情况下，这里是no，反之则得是yes
        Failed balancer rounds in last 5 attempts:  0
        Migration Results for the last 24 hours: 
                No recent migrations
  databases:
        {  "_id" : "config",  "primary" : "config",  "partitioned" : true }
                config.system.sessions
                        shard key: { "_id" : 1 }
                        unique: false
                        balancing: true
                        chunks:
                                shard1  1
                        { "_id" : { "$minKey" : 1 } } -->> { "_id" : { "$maxKey" : 1 } } on : shard1 Timestamp(1, 0) 

mongos>

mongodb分片测试

1.登录任意一台20000端口：

[root@localhost ~]# mongo --host 192.168.77.128 --port 20000

2.进入admin库，使用以下任意一条命令指定要分片的数据库：

db.runCommand({ enablesharding : "testdb"})
sh.enableSharding("testdb")

示例：

mongos> use admin
switched to db admin
mongos> sh.enableSharding("testdb")
{
    "ok" : 1,
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515609562, 6),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515609562, 6)
}
mongos>

3.使用以下任意一条命令指定数据库里需要分片的集合和片键：

db.runCommand( { shardcollection : "testdb.table1",key : {id: 1} } )
sh.shardCollection("testdb.table1",{"id":1} )

示例：

mongos> sh.shardCollection("testdb.table1",{"id":1} )
{
    "collectionsharded" : "testdb.table1",
    "collectionUUID" : UUID("f98762a6-8b2b-4ae5-9142-3d8acc589255"),
    "ok" : 1,
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515609671, 12),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515609671, 12)
}
mongos>

4.进入刚刚创建的testdb库里插入测试数据：

mongos> use testdb
switched to db testdb
mongos> for (var i = 1; i <= 10000; i++) db.table1.save({id:i,"test1":"testval1"})
WriteResult({ "nInserted" : 1 })
mongos>

5.然后创建多几个库和集合：

mongos> sh.enableSharding("db1")
mongos> sh.shardCollection("db1.table1",{"id":1} )
mongos> sh.enableSharding("db2")
mongos> sh.shardCollection("db2.table1",{"id":1} )
mongos> sh.enableSharding("db3")
mongos> sh.shardCollection("db3.table1",{"id":1} )

6.查看状态：

mongos> sh.status()
--- Sharding Status --- 
  sharding version: {
    "_id" : 1,
    "minCompatibleVersion" : 5,
    "currentVersion" : 6,
    "clusterId" : ObjectId("5a55823348aee75ba3928fea")
  }
  shards:
        {  "_id" : "shard1",  "host" : "shard1/192.168.77.128:27001,192.168.77.130:27001",  "state" : 1 }
        {  "_id" : "shard2",  "host" : "shard2/192.168.77.130:27002,192.168.77.134:27002",  "state" : 1 }
        {  "_id" : "shard3",  "host" : "shard3/192.168.77.128:27003,192.168.77.134:27003",  "state" : 1 }
  active mongoses:
        "3.6.1" : 1
  autosplit:
        Currently enabled: yes
  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" : "config",  "primary" : "config",  "partitioned" : true }
                config.system.sessions
                        shard key: { "_id" : 1 }
                        unique: false
                        balancing: true
                        chunks:
                                shard1  1  
                        { "_id" : { "$minKey" : 1 } } -->> { "_id" : { "$maxKey" : 1 } } on : shard1 Timestamp(1, 0) 
        {  "_id" : "db1",  "primary" : "shard3",  "partitioned" : true }
                db1.table1
                        shard key: { "id" : 1 }
                        unique: false
                        balancing: true
                        chunks:
                                shard3  1  # db1存储到了shard3中
                        { "id" : { "$minKey" : 1 } } -->> { "id" : { "$maxKey" : 1 } } on : shard3 Timestamp(1, 0) 
        {  "_id" : "db2",  "primary" : "shard1",  "partitioned" : true }
                db2.table1
                        shard key: { "id" : 1 }
                        unique: false
                        balancing: true
                        chunks:
                                shard1  1  # db2存储到了shard1中
                        { "id" : { "$minKey" : 1 } } -->> { "id" : { "$maxKey" : 1 } } on : shard1 Timestamp(1, 0) 
        {  "_id" : "db3",  "primary" : "shard3",  "partitioned" : true }
                db3.table1
                        shard key: { "id" : 1 }
                        unique: false
                        balancing: true
                        chunks:
                                shard3  1  # db3存储到了shard3中
                        { "id" : { "$minKey" : 1 } } -->> { "id" : { "$maxKey" : 1 } } on : shard3 Timestamp(1, 0) 
        {  "_id" : "testdb",  "primary" : "shard2",  "partitioned" : true }
                testdb.table1
                        shard key: { "id" : 1 }
                        unique: false
                        balancing: true
                        chunks:
                                shard2  1  # testdb存储到了shard2中
                        { "id" : { "$minKey" : 1 } } -->> { "id" : { "$maxKey" : 1 } } on : shard2 Timestamp(1, 0) 

mongos> 

如上，可以看到，刚刚创建的库都存储在了各个分片上，证明分片已经搭建成功。

使用以下命令可以查看某个集合的状态：

db.集合名称.stats()

mongodb备份/恢复

mongodb备份数据

1.首先演示备份某个指定库：

[root@localhost ~]# mkdir /tmp/mongobak  # 先创建一个目录用来存放备份文件
[root@localhost ~]# mongodump --host 192.168.77.128 --port 20000  -d testdb -o /tmp/mongobak
2018-01-10T20:47:51.893+0800    writing testdb.table1 to 
2018-01-10T20:47:51.968+0800    done dumping testdb.table1 (10000 documents)
[root@localhost ~]# ls /tmp/mongobak/  # 备份成功后会生成一个目录
testdb
[root@localhost ~]# ls /tmp/mongobak/testdb/  # 目录里则会生成相应的数据文件
table1.bson  table1.metadata.json
[root@localhost /tmp/mongobak/testdb]# du -sh *  # 可以看到，存放数据的是.bson文件
528K    table1.bson
4.0K    table1.metadata.json
[root@localhost /tmp/mongobak/testdb]#

mongodump 命令中，-d指定需要备份的库，-o指定备份路径

2.备份所有库示例：

[root@localhost ~]# mongodump --host 192.168.77.128 --port 20000 -o /tmp/mongobak
2018-01-10T20:52:28.231+0800    writing admin.system.version to 
2018-01-10T20:52:28.233+0800    done dumping admin.system.version (1 document)
2018-01-10T20:52:28.233+0800    writing testdb.table1 to 
2018-01-10T20:52:28.234+0800    writing config.locks to 
2018-01-10T20:52:28.234+0800    writing config.changelog to 
2018-01-10T20:52:28.234+0800    writing config.lockpings to 
2018-01-10T20:52:28.235+0800    done dumping config.locks (15 documents)
2018-01-10T20:52:28.236+0800    writing config.chunks to 
2018-01-10T20:52:28.236+0800    done dumping config.lockpings (10 documents)
2018-01-10T20:52:28.236+0800    writing config.collections to 
2018-01-10T20:52:28.236+0800    done dumping config.changelog (13 documents)
2018-01-10T20:52:28.236+0800    writing config.databases to 
2018-01-10T20:52:28.237+0800    done dumping config.collections (5 documents)
2018-01-10T20:52:28.237+0800    writing config.shards to 
2018-01-10T20:52:28.237+0800    done dumping config.chunks (5 documents)
2018-01-10T20:52:28.237+0800    writing config.version to 
2018-01-10T20:52:28.238+0800    done dumping config.databases (4 documents)
2018-01-10T20:52:28.238+0800    writing config.mongos to 
2018-01-10T20:52:28.238+0800    done dumping config.version (1 document)
2018-01-10T20:52:28.238+0800    writing config.migrations to 
2018-01-10T20:52:28.239+0800    done dumping config.mongos (1 document)
2018-01-10T20:52:28.239+0800    writing db1.table1 to 
2018-01-10T20:52:28.239+0800    done dumping config.shards (3 documents)
2018-01-10T20:52:28.239+0800    writing db2.table1 to 
2018-01-10T20:52:28.239+0800    done dumping config.migrations (0 documents)
2018-01-10T20:52:28.239+0800    writing db3.table1 to 
2018-01-10T20:52:28.241+0800    done dumping db2.table1 (0 documents)
2018-01-10T20:52:28.241+0800    writing config.tags to 
2018-01-10T20:52:28.241+0800    done dumping db1.table1 (0 documents)
2018-01-10T20:52:28.242+0800    done dumping db3.table1 (0 documents)
2018-01-10T20:52:28.243+0800    done dumping config.tags (0 documents)
2018-01-10T20:52:28.272+0800    done dumping testdb.table1 (10000 documents)
[root@localhost ~]# ls /tmp/mongobak/
admin  config  db1  db2  db3  testdb
[root@localhost ~]#

没有指定-d选项就会备份所有的库。

3.除了备份库之外，还可以备份某个指定的集合：

[root@localhost ~]# mongodump --host 192.168.77.128 --port 20000 -d testdb -c table1 -o /tmp/collectionbak
2018-01-10T20:56:55.300+0800    writing testdb.table1 to 
2018-01-10T20:56:55.335+0800    done dumping testdb.table1 (10000 documents)
[root@localhost ~]# ls !$
ls /tmp/collectionbak
testdb
[root@localhost ~]# ls /tmp/collectionbak/testdb/
table1.bson  table1.metadata.json
[root@localhost ~]#

-c选项指定需要备份的集合，如果没有指定-c选项，则会备份该库的所有集合。

4.mongoexport 命令可以将集合导出为json文件：

[root@localhost ~]# mongoexport --host 192.168.77.128 --port 20000 -d testdb -c table1 -o /tmp/table1.json  # 导出来的是一个json文件
2018-01-10T21:00:48.098+0800    connected to: 192.168.77.128:20000
2018-01-10T21:00:48.236+0800    exported 10000 records
[root@localhost ~]# ls !$
ls /tmp/table1.json
/tmp/table1.json
[root@localhost ~]# tail -n5 !$  # 可以看到文件中都是json格式的数据
tail -n5 /tmp/table1.json
{"_id":{"$oid":"5a55f036f6179723bfb03611"},"id":9996.0,"test1":"testval1"}
{"_id":{"$oid":"5a55f036f6179723bfb03612"},"id":9997.0,"test1":"testval1"}
{"_id":{"$oid":"5a55f036f6179723bfb03613"},"id":9998.0,"test1":"testval1"}
{"_id":{"$oid":"5a55f036f6179723bfb03614"},"id":9999.0,"test1":"testval1"}
{"_id":{"$oid":"5a55f036f6179723bfb03615"},"id":10000.0,"test1":"testval1"}
[root@localhost ~]#

mongodb恢复数据

1.上面我们已经备份好了数据，现在我们先把MongoDB中的数据都删除：

[root@localhost ~]# mongo --host 192.168.77.128 --port 20000
mongos> use testdb
switched to db testdb
mongos> db.dropDatabase()
{
    "dropped" : "testdb",
    "ok" : 1,
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515617938, 13),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515617938, 13)
}
mongos> use db1
switched to db db1
mongos> db.dropDatabase()
{
    "dropped" : "db1",
    "ok" : 1,
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515617993, 19),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515617993, 19)
}
mongos> use db2
switched to db db2
mongos> db.dropDatabase()
{
    "dropped" : "db2",
    "ok" : 1,
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515618003, 13),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515618003, 13)
}
mongos> use db3
switched to db db3
mongos> db.dropDatabase()
{
    "dropped" : "db3",
    "ok" : 1,
    "$clusterTime" : {
        "clusterTime" : Timestamp(1515618003, 34),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    },
    "operationTime" : Timestamp(1515618003, 34)
}
mongos> show databases
admin   0.000GB
config  0.001GB
mongos>

2.恢复所有的库：

[root@localhost ~]# rm -rf /tmp/mongobak/config/  # 因为不需要恢复config和admin库，所以先把备份文件删掉
[root@localhost ~]# rm -rf /tmp/mongobak/admin/
[root@localhost ~]# mongorestore --host 192.168.77.128 --port 20000 --drop /tmp/mongobak/
2018-01-10T21:11:40.031+0800    preparing collections to restore from
2018-01-10T21:11:40.033+0800    reading metadata for testdb.table1 from /tmp/mongobak/testdb/table1.metadata.json
2018-01-10T21:11:40.035+0800    reading metadata for db2.table1 from /tmp/mongobak/db2/table1.metadata.json
2018-01-10T21:11:40.040+0800    reading metadata for db3.table1 from /tmp/mongobak/db3/table1.metadata.json
2018-01-10T21:11:40.050+0800    reading metadata for db1.table1 from /tmp/mongobak/db1/table1.metadata.json
2018-01-10T21:11:40.086+0800    restoring testdb.table1 from /tmp/mongobak/testdb/table1.bson
2018-01-10T21:11:40.100+0800    restoring db2.table1 from /tmp/mongobak/db2/table1.bson
2018-01-10T21:11:40.102+0800    restoring indexes for collection db2.table1 from metadata
2018-01-10T21:11:40.118+0800    finished restoring db2.table1 (0 documents)
2018-01-10T21:11:40.123+0800    restoring db3.table1 from /tmp/mongobak/db3/table1.bson
2018-01-10T21:11:40.124+0800    restoring indexes for collection db3.table1 from metadata
2018-01-10T21:11:40.126+0800    restoring db1.table1 from /tmp/mongobak/db1/table1.bson
2018-01-10T21:11:40.172+0800    finished restoring db3.table1 (0 documents)
2018-01-10T21:11:40.173+0800    restoring indexes for collection db1.table1 from metadata
2018-01-10T21:11:40.185+0800    finished restoring db1.table1 (0 documents)
2018-01-10T21:11:40.417+0800    restoring indexes for collection testdb.table1 from metadata
2018-01-10T21:11:40.437+0800    finished restoring testdb.table1 (10000 documents)
2018-01-10T21:11:40.437+0800    done
[root@localhost ~]# mongo --host 192.168.77.128 --port 20000
mongos> show databases;  # 可以看到，所有的库都恢复了
admin   0.000GB
config  0.001GB
db1     0.000GB
db2     0.000GB
db3     0.000GB
testdb  0.000GB
mongos>

mongorestore 命令中的--drop可选，意思是当恢复之前先把之前的数据删除，生产环境不建议使用

3.恢复指定的库：

[root@localhost ~]# mongorestore --host 192.168.77.128 --port 20000 -d testdb --drop /tmp/mongobak/testdb/
2018-01-10T21:15:40.185+0800    the --db and --collection args should only be used when restoring from a BSON file. Other uses are deprecated and will not exist in the future; use --nsInclude instead
2018-01-10T21:15:40.185+0800    building a list of collections to restore from /tmp/mongobak/testdb dir
2018-01-10T21:15:40.232+0800    reading metadata for testdb.table1 from /tmp/mongobak/testdb/table1.metadata.json
2018-01-10T21:15:40.241+0800    restoring testdb.table1 from /tmp/mongobak/testdb/table1.bson
2018-01-10T21:15:40.507+0800    restoring indexes for collection testdb.table1 from metadata
2018-01-10T21:15:40.529+0800    finished restoring testdb.table1 (10000 documents)
2018-01-10T21:15:40.529+0800    done
[root@localhost ~]#

恢复某个指定库的时候要指定到具体的备份该库的目录。

4.恢复指定的集合：

[root@localhost ~]# mongorestore --host 192.168.77.128 --port 20000 -d testdb -c table1 --drop /tmp/mongobak/testdb/table1.bson 
2018-01-10T21:18:14.097+0800    checking for collection data in /tmp/mongobak/testdb/table1.bson
2018-01-10T21:18:14.139+0800    reading metadata for testdb.table1 from /tmp/mongobak/testdb/table1.metadata.json
2018-01-10T21:18:14.149+0800    restoring testdb.table1 from /tmp/mongobak/testdb/table1.bson
2018-01-10T21:18:14.331+0800    restoring indexes for collection testdb.table1 from metadata
2018-01-10T21:18:14.353+0800    finished restoring testdb.table1 (10000 documents)
2018-01-10T21:18:14.353+0800    done
[root@localhost ~]#

同样的恢复某个指定集合的时候要指定到具体的备份该集合的.bson文件。

5.恢复json文件中的集合数据：

[root@localhost ~]# mongoimport --host 192.168.77.128 --port 20000 -d testdb -c table1 --file /tmp/table1.json

恢复json文件中的集合数据使用的是mongoimport 命令，--file指定json文件所在路径。

本文转自 ZeroOne01 51CTO博客，原文链接：http://blog.51cto.com/zero01/2059598，如需转载请自行联系原作者