mongodb学习笔记—mongodb集群搭建（master slave集群）

最新推荐文章于 2023-11-17 10:47:49 发布

chfantv17156

最新推荐文章于 2023-11-17 10:47:49 发布

阅读量394

点赞数

文章标签： git 开发工具数据库

原文链接：https://my.oschina.net/djflying/blog/732100

版权

##集群规划

IP	端口	角色
10.1.200.77	27017	master
10.1.200.78	27017	slave
10.1.200.87	27017	slave
10.1.200.118	27017	slave

master、slave集群具体流程如下图所示：

输入图片说明

流程分析：

1、写操作访问master，读操作访问slave。

2、master和slave之间没有像redis一样主挂从顶的机制。

3、slave根据自己节点中的source来找到当前是和哪个master通信。

4、master和slave之间是通过oplog.$main的操作日志来同步的，默认是每隔10s同步一次。

5、只有master和slave之间有通信，slave和slave之间是没有通信的。

6、master节点挂了之后，slave无法自动切换为写操作，需要手工处理。

##安装和配置

参照单实例。

##各个节点配置文件修改

1、10.1.200.77节点：

[root@centos4 init.d]# vim /etc/mongod.conf 
# 数据库文件位置
dbpath=/usr/local/mongodb/db
# 日志文件位置
logpath=/usr/local/mongodb/logs/mongodb.log
# 以追加方式写入日志
logappend=true
# 绑定服务IP，若绑定127.0.0.1，则只能本机访问，不指定默认本地所有IP  
bind_ip=10.1.200.77
# 默认端口为27017
port=27017

# 是否为master
master=true

2、10.1.200.78节点：

[root@centos4 init.d]# vim /etc/mongod.conf 
# 数据库文件位置
dbpath=/usr/local/mongodb/db
# 日志文件位置
logpath=/usr/local/mongodb/logs/mongodb.log
# 以追加方式写入日志
logappend=true
# 绑定服务IP，若绑定127.0.0.1，则只能本机访问，不指定默认本地所有IP  
bind_ip=10.1.200.78
# 默认端口为27017
port=27017

# 是否为slave
slave=true
# 从哪个master节点复制数据
source=10.1.200.77:27017

3、10.1.200.87节点：

[root@centos4 init.d]# vim /etc/mongod.conf 
# 数据库文件位置
dbpath=/usr/local/mongodb/db
# 日志文件位置
logpath=/usr/local/mongodb/logs/mongodb.log
# 以追加方式写入日志
logappend=true
# 绑定服务IP，若绑定127.0.0.1，则只能本机访问，不指定默认本地所有IP  
bind_ip=10.1.200.87
# 默认端口为27017
port=27017

# 是否为slave
slave=true
# 从哪个master节点复制数据
source=10.1.200.77:27017

4、10.1.200.118节点：

[root@centos4 init.d]# vim /etc/mongod.conf 
# 数据库文件位置
dbpath=/usr/local/mongodb/db
# 日志文件位置
logpath=/usr/local/mongodb/logs/mongodb.log
# 以追加方式写入日志
logappend=true
# 绑定服务IP，若绑定127.0.0.1，则只能本机访问，不指定默认本地所有IP  
bind_ip=10.1.200.118
# 默认端口为27017
port=27017

# 是否为slave
slave=true
# 从哪个master节点复制数据
source=10.1.200.77:27017

##查看各个节点的状态 1、重新启动各个节点的mongodb服务，查看mongodb的日志文件：

1）10.1.200.77节点如下：

2016-08-11T16:29:33.899+0800 I CONTROL  [main] ***** SERVER RESTARTED *****
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten] MongoDB starting : pid=4785 port=27017 dbpath=/usr/local/mongodb/db master=1 64-bit host=centos4
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten] db version v3.2.8
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten] git version: ed70e33130c977bda0024c125b56d159573dbaf0
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten] allocator: tcmalloc
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten] modules: none
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten] build environment:
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten]     distarch: x86_64
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten]     target_arch: x86_64
2016-08-11T16:29:33.906+0800 I CONTROL  [initandlisten] options: { config: "/etc/mongod.conf", master: true, net: { port: 27017 }, storage: { dbPath: "/usr/local/mongodb/db" }, systemLog: { destination: "file", logAppend: true, path: "/usr/local/mongodb/logs/mongodb.log" } }
……

从日志中可以看出这个节点是master。

2）10.1.200.78、10.1.200.87和10.1.200.118三个节点如下：

2016-08-11T16:30:50.897+0800 I CONTROL  [main] ***** SERVER RESTARTED *****
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten] MongoDB starting : pid=31548 port=27017 dbpath=/usr/local/mongodb/db slave=1 64-bit host=centos4
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten] db version v3.2.8
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten] git version: ed70e33130c977bda0024c125b56d159573dbaf0
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten] allocator: tcmalloc
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten] modules: none
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten] build environment:
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten]     distarch: x86_64
2016-08-11T16:30:50.903+0800 I CONTROL  [initandlisten]     target_arch: x86_64
2016-08-11T16:30:50.904+0800 I CONTROL  [initandlisten] options: { config: "/etc/mongod.conf", net: { port: 27017 }, slave: true, source: "10.1.200.77:27017", storage: { dbPath: "/usr/local/mongodb/db" }, systemLog: { destination: "file", logAppend: true, path: "/usr/local/mongodb/logs/mongodb.log" } }
……

以下为slave复制master的日志信息：
2016-08-11T16:30:52.257+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T16:30:52.264+0800 I REPL     [replslave] resync: dropping database flightswitch
2016-08-11T16:30:52.264+0800 I REPL     [replslave] resync: cloning database flightswitch to get an initial copy
2016-08-11T16:30:52.319+0800 I INDEX    [replslave] build index on: flightswitch.flight properties: { v: 1, key: { _id: 1 }, name: "_id_", ns: "flightswitch.flight" }
2016-08-11T16:30:52.319+0800 I INDEX    [replslave]      building index using bulk method
2016-08-11T16:30:52.323+0800 I INDEX    [replslave] build index done.  scanned 1 total records. 0 secs
2016-08-11T16:30:52.323+0800 I STORAGE  [replslave] copying indexes for: { name: "flight", options: { autoIndexId: true } }
2016-08-11T16:30:52.323+0800 I REPL     [replslave] resync: done with initial clone for db: flightswitch
2016-08-11T16:30:53.328+0800 I REPL     [replslave] sleep 2 sec before next pass
2016-08-11T16:30:55.329+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T16:30:56.331+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T16:30:57.334+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T16:30:58.336+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T16:30:59.339+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T16:31:00.341+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T16:31:01.444+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017

从日志中可以看出3个节点都是slave，而且slave都在同步master的信息。

2、也可以使用db.isMaster()命令来查看各个节点是master还是slave：

1）10.1.200.77节点：

> db.isMaster()
{
        "ismaster" : true,
        "maxBsonObjectSize" : 16777216,
        "maxMessageSizeBytes" : 48000000,
        "maxWriteBatchSize" : 1000,
        "localTime" : ISODate("2016-08-11T08:46:54.433Z"),
        "maxWireVersion" : 4,
        "minWireVersion" : 0,
        "ok" : 1
}
>

2）10.1.200.78节点：

> db.isMaster()
{
        "ismaster" : false,
        "maxBsonObjectSize" : 16777216,
        "maxMessageSizeBytes" : 48000000,
        "maxWriteBatchSize" : 1000,
        "localTime" : ISODate("2016-08-11T08:51:25.190Z"),
        "maxWireVersion" : 4,
        "minWireVersion" : 0,
        "ok" : 1
}
>

3）10.1.200.87节点：

> db.isMaster()
{
        "ismaster" : false,
        "maxBsonObjectSize" : 16777216,
        "maxMessageSizeBytes" : 48000000,
        "maxWriteBatchSize" : 1000,
        "localTime" : ISODate("2016-08-11T08:53:15.192Z"),
        "maxWireVersion" : 4,
        "minWireVersion" : 0,
        "ok" : 1
}
>

4）10.1.200.118节点：

> db.isMaster()
{
        "ismaster" : false,
        "maxBsonObjectSize" : 16777216,
        "maxMessageSizeBytes" : 48000000,
        "maxWriteBatchSize" : 1000,
        "localTime" : ISODate("2016-08-11T08:54:05.810Z"),
        "maxWireVersion" : 4,
        "minWireVersion" : 0,
        "ok" : 1
}
>

##集群测试 1、在10.1.200.77节点写入数据：

> use flightswitch
switched to db flightswitch
> db.flight.insert({"dai":"daijiong"})
WriteResult({ "nInserted" : 1 })
>

2、在10.1.200.78节点读这个数据：

> use flightswitch
switched to db flightswitch
> db.flight.find()
Error: error: { "ok" : 0, "errmsg" : "not master and slaveOk=false", "code" : 13435 }

此时发现读取的时候报错了，这是正常的，因为slave是不允许读写的，需要在slave中指定slaveOk：

> db.getMongo().setSlaveOk()

此时再查询就成功了：

> db.flight.find()
{ "_id" : ObjectId("57a9607cec764a2c6cecb687"), "originCity" : "SHA" }
{ "_id" : ObjectId("57ac3fa39ba331615f2474b7"), "dai" : "daijiong" }
{ "_id" : ObjectId("57ac43e3bc145f0d25ae7186"), "jiong" : "daijiong" }
{ "_id" : ObjectId("57ac489dbc145f0d25ae7187"), "originCity" : "SHA" }
>

##集群故障转移测试模拟当master节点挂了的情况下，slave是否可以当成master。

1）在10.1.200.77节点关闭mongodb服务，

[root@centos4 logs]# service mongodb stop
/etc/init.d/mongodb: line 1: tartup: command not found
db path is: /usr/local/mongodb/db
/usr/local/mongodb/bin/mongod
Stopping mongod:                                           [确定]
[root@centos4 logs]#

在10.1.200.78节点尝试写入操作

> use flightswitch
switched to db flightswitch
> db.flight.insert({"dai1":"daijiong"})
WriteResult({ "writeError" : { "code" : 10107, "errmsg" : "not master" } })
>

此时发现无法写入，说明当master挂了，slave无法当成maser。

2）那此时整个集群中就没有了master，只能从集群中读取数据，不能往集群里写入数据，这时候只能手动的将其中一个slave变成master。比如此时想要把10.1.200.78这个slave节点变成master节点，首先需要停止10.1.200.78节点上的mongodb服务

[root@centos4 logs]# service mongodb stop
db path is: /usr/local/mongodb/db
/usr/local/mongodb/bin/mongod
Stopping mongod:                                           [确定]

然后修改配置文件将slave=true修改为master=true。

[root@centos4 logs]# vim /etc/mongod.conf 
# 数据库文件位置
dbpath=/usr/local/mongodb/db
# 日志文件位置
logpath=/usr/local/mongodb/logs/mongodb.log
# 以追加方式写入日志
logappend=true
# 默认端口为27017
port=27017
# 是否为master
master=true

然后再启动mongodb服务：

[root@centos4 logs]# service mongodb start
db path is: /usr/local/mongodb/db
/usr/local/mongodb/bin/mongod
Starting mongod: /usr/local/mongodb/bin/mongod --config /etc/mongod.conf

[root@centos4 logs]#

然后再尝试写入操作：

> use flightswitch
switched to db flightswitch
> db.flight.insert({"dai1":"daijiong"})
WriteResult({ "nInserted" : 1 })
>

发现此时是正常的。但是此时另外两个slave里面配置连接的source需要修改成这个新的master的ip和端口，否则会一直报连接失败：

2016-08-11T18:14:12.342+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T18:14:12.343+0800 W NETWORK  [replslave] Failed to connect to 10.1.200.77:27017, reason: errno:111 Connection refused
2016-08-11T18:14:12.343+0800 E REPL     [replslave] couldn't connect to server 10.1.200.77:27017, connection attempt failed
2016-08-11T18:14:12.343+0800 I REPL     [replslave] sleep 3 sec before next pass
2016-08-11T18:14:15.345+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T18:14:15.345+0800 W NETWORK  [replslave] Failed to connect to 10.1.200.77:27017, reason: errno:111 Connection refused
2016-08-11T18:14:15.345+0800 E REPL     [replslave] couldn't connect to server 10.1.200.77:27017, connection attempt failed
2016-08-11T18:14:15.345+0800 I REPL     [replslave] sleep 3 sec before next pass
2016-08-11T18:14:18.347+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T18:14:18.348+0800 W NETWORK  [replslave] Failed to connect to 10.1.200.77:27017, reason: errno:111 Connection refused
2016-08-11T18:14:18.348+0800 E REPL     [replslave] couldn't connect to server 10.1.200.77:27017, connection attempt failed
2016-08-11T18:14:18.348+0800 I REPL     [replslave] sleep 3 sec before next pass
2016-08-11T18:14:21.350+0800 I REPL     [replslave] syncing from host:10.1.200.77:27017
2016-08-11T18:14:21.350+0800 W NETWORK  [replslave] Failed to connect to 10.1.200.77:27017, reason: errno:111 Connection refused
2016-08-11T18:14:21.351+0800 E REPL     [replslave] couldn't connect to server 10.1.200.77:27017, connection attempt failed
2016-08-11T18:14:21.351+0800 I REPL     [replslave] sleep 3 sec before next pass

由此可见这种集群一旦master出现问题，需要手工操作才能恢复这个集群。

转载于:https://my.oschina.net/djflying/blog/732100

chfantv17156

关注

0
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
mongodb学习笔记—mongodb集群搭建（master slave集群）

##集群规划IP端口角色10.1.200.7727017master10.1.200.7827017slave10.1.200.8727017slave10.1.200.11827017slavemaster、slave集群具体流...
复制链接

扫一扫