MongoDB副本集搭建和测试

前段时间花了时间看了看mongodb，室友买了mongodb权威指南也借过来瞅瞅，书对于刚刚接触的人还是不错的。废话不多说上代码。

搭建环境

MacBook pro 15款840

VMWare Fusion8

Centos 6.7

mongodb-linux-x86_64-3.0.6.tgz

安装mongodb

这一步其实就是一个解压的过程。很简单。

在安装前把dbpath和logpath路径新建好，然后会报错。

[root@rabbitmq-node2 ~]# mkdir -p /mongodb/data/arbiter
[root@rabbitmq-node2 ~]# mkdir -p /mongodb/logs
[root@rabbitmq-node2 data]# cd /usr/programmer/
[root@rabbitmq-node2 programmer]# tar zxf mongodb-linux-x86_64-3.0.6.tgz
[root@rabbitmq-node2 programmer]# mv mongodb-linux-x86_64-3.0.6 mongodb
[root@rabbitmq-node2 programmer]# mv mongodb /usr/local/
[root@rabbitmq-node2 programmer]# ls
apache-tomcat-8.0.26.tar.gz     nginx-1.9.4.tar.gz
erlang-17.4-1.el6.x86_64.rpm    openssl-1.0.2c.tar.gz
go1.4.2.linux-amd64.tar.gz      rabbitmq-server-3.5.6-1.noarch.rpm
jdk-7u80-linux-x64.tar.gz       redis-3.0.3.tar
mongodb-linux-x86_64-3.0.6.tgz  solr-5.2.1.tgz
mongodb-osx-x86_64-3.0.6.tgz    zookeeper-3.4.6.tar.gz
nginx-1.8.0.tar.gz
[root@rabbitmq-node2 programmer]# cd /usr/local/
[root@rabbitmq-node2 local]# cd mongodb/
[root@rabbitmq-node2 mongodb]# ls
bin  GNU-AGPL-3.0  README  THIRD-PARTY-NOTICES
[root@rabbitmq-node2 mongodb]# mkdir conf
[root@rabbitmq-node2 mongodb]# cd conf/

这里有三台虚拟机都是一样这样安装。我就省略其他的了。

副本集群搭建

进入集群搭建，需要加一些配置：如下

#master.conf
#数据存放目录
dbpath=/mongodb/data/master
#日志存放目录
logpath=/mongodb/logs/master.log
#进程文件
pidfilepath=/mongodb/master.pid
#按照数据库名字建立目录存放数据
directoryperdb=true
#指定日志文件，该文件将保存所有的日志记录、诊断信息。除非另有指定，
#mongod将所有的日志信息输出到标准输出。如果没有指定logappend，重启则日
#志会进行覆盖操作
logappend=true
#副本集名字
replSet=szss
#绑定IP地址
bind_ip=192.168.31.101
#端口号
port=27017
#操作日志大小
oplogSize=10000
#是否后台运行 进程在后台运行的守护进程模式
fork=true
#预分配文件缩短启动时间
noprealloc=true

注意如果之后关闭mongodb后重启报错，可能是你强行关闭造成锁表，或者因为你的磁盘不够大了。需要加上 nojournal = true 关闭操作日志。 但是 非常不建议这样做。因为它可以帮助你备份和故障修复，提高写入的速率。

其他的两个配置也是大差不差。就不一一贴代码了。

启动mongodb服务：

[root@rabbitmq-node1 bin]# ./mongod --fork --rest ../conf/master.config

2015-11-04T15:58:02.803-0500 I CONTROL  ** WARNING: --rest is specified without --httpinterface,

2015-11-04T15:58:02.803-0500 I CONTROL  **          enabling http interface

BadValue --fork has to be used with --logpath or --syslog

try './mongod --help' for more information

[root@rabbitmq-node1 bin]# ./mongod --rest -f  ../conf/master.config

2015-11-04T15:58:41.235-0500 I CONTROL  ** WARNING: --rest is specified without --httpinterface,

2015-11-04T15:58:41.235-0500 I CONTROL  **          enabling http interface

note: noprealloc may hurt performance in many applications

about to fork child process, waiting until server is ready for connections.

forked process: 2494

child process started successfully, parent exiting
[root@rabbitmq-node1 bin]# ps -ef | grep mongod
root       2494      1  0 15:58 ?        00:00:00 ./mongod --rest -f ../conf/master.config
root       2537   2382  0 16:01 pts/0    00:00:00 grep mongod
[root@rabbitmq-node1 bin]# ./mongo 192.168.31.101:27017

MongoDB shell version: 3.0.6

connecting to: 192.168.31.101:27017/test

Server has startup warnings:

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten] ** WARNING: You are running this process as the root user, which is not recommended.

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten]

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten]

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten] ** WARNING: /sys/kernel/mm/transparent_hugepage/enabled is 'always'.

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten] **        We suggest setting it to 'never'

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten]

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten] ** WARNING: /sys/kernel/mm/transparent_hugepage/defrag is 'always'.

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten] **        We suggest setting it to 'never'

2015-11-04T15:49:55.987-0500 I CONTROL  [initandlisten]

> use admin

switched to db admin

> db.shutdownServer();

assert failed : unexpected error: Error: shutdownServer failed: shutdown must run from localhost when running db without auth

Error: assert failed : unexpected error: Error: shutdownServer failed: shutdown must run from localhost when running db without auth

at Error (<anonymous>)

at doassert (src/mongo/shell/assert.js:11:14)

at assert (src/mongo/shell/assert.js:20:5)

at DB.shutdownServer (src/mongo/shell/db.js:212:9)

at (shell):1:4

2015-11-04T15:54:48.911-0500 E QUERY    Error: assert failed : unexpected error: Error: shutdownServer failed: shutdown must run from localhost when running db without auth

at Error (<anonymous>)

at doassert (src/mongo/shell/assert.js:11:14)

at assert (src/mongo/shell/assert.js:20:5)

at DB.shutdownServer (src/mongo/shell/db.js:212:9)

at (shell):1:4 at src/mongo/shell/assert.js:13

> ^C

bye

–rest是开启webUI监考页面



这里说 使用使用这种方式关闭服务 必须在有登录验证情况下远程ssh才拥有这个权限的，所以

[root@rabbitmq-node1 bin]# ./mongod --shutdown --dbpath /mongodb/data/master/

killing process with pid: 2434

使用客户端命令来关闭。

启动mongodb服务：

[root@rabbitmq-node1 bin]# ./mongo 127.0.0.1

MongoDB shell version: 3.0.6

connecting to: 127.0.0.1/test

2015-11-04T16:01:53.200-0500 W NETWORK  Failed to connect to 127.0.0.1:27017, reason: errno:111 Connection refused

2015-11-04T16:01:53.201-0500 E QUERY    Error: couldn't connect to server 127.0.0.1:27017 (127.0.0.1), connection attempt failed

at connect (src/mongo/shell/mongo.js:181:14)

at (connect):1:6 at src/mongo/shell/mongo.js:181

exception: connect failed

可以看出是不识别localhost 127.0.0.1的

[root@rabbitmq-node1 bin]# ./mongo 192.168.31.101
MongoDB shell version: 3.0.6
connecting to: 192.168.31.101/test
Server has startup warnings:
2015-11-04T15:58:41.235-0500 I CONTROL  ** WARNING: --rest is specified without --httpinterface,
2015-11-04T15:58:41.235-0500 I CONTROL  **          enabling http interface
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten] ** WARNING: You are running this process as the root user, which is not recommended.
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten]
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten]
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten] ** WARNING: /sys/kernel/mm/transparent_hugepage/enabled is 'always'.
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten] **        We suggest setting it to 'never'
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten]
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten] ** WARNING: /sys/kernel/mm/transparent_hugepage/defrag is 'always'.
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten] **        We suggest setting it to 'never'
2015-11-04T15:58:41.304-0500 I CONTROL  [initandlisten]
> use admin
switched to db admin
> cfg={ "_id" : "szss", "members" : [ {  "_id" : 0,  "host" : "192.168.31.101:27017"  }  ]  }
{
"_id": "szss",
"members": [
{
"_id": 0,
"host": "192.168.31.101:27017"
}
]
}
> rs.initiate(cfg)
{ "ok" : 1 }
szss:OTHER> rs.add("192.168.31.154:27017")
{ "ok" : 1 }
szss:PRIMARY> rs.add("192.168.31.159:27017")
{ "ok" : 1 }

加入集群配置 并初始化 并且加上其他虚机的服务 这里也可以一步到位。 我也比较推荐这种方式。

{
_id: "szss",
members: [
{
_id: 0,
host: '192.168.31.101: 27017',
priority: 2
},
{
_id: 1,
host: '192.168.31.154: 27017',
priority: 1
},
{
_id: 2,
host: '192.168.31.159: 27017',
arbiterOnly: true
}
]
};

这种方式是 主节点，备节点，仲裁节点的副本集集群，但是在实际中，仲裁节点比较鸡肋，他要占用一个服务，但是只是为了投票而已，只要将备节点的个数控制在奇数就足够了。

回到现在的这种副本集集群搭建的方式，现在要赋予节点权重。权重是选举中优先级最高的。

szss:PRIMARY> rs.conf()
{
"_id": "szss",
"version": 3,
"members": [
{
"_id": 0,
"host": "192.168.31.101:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {

},
"slaveDelay": 0,
"votes": 1
},
{
"_id": 1,
"host": "192.168.31.154:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {

},
"slaveDelay": 0,
"votes": 1
},
{
"_id": 2,
"host": "192.168.31.159:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {

},
"slaveDelay": 0,
"votes": 1
}
],
"settings": {
"chainingAllowed": true,
"heartbeatTimeoutSecs": 10,
"getLastErrorModes": {

},
"getLastErrorDefaults": {
"w": 1,
"wtimeout": 0
}
}
}
szss:PRIMARY> cfgs=rs.conf()
{
"_id": "szss",
"version": 3,
"members": [
{
"_id": 0,
"host": "192.168.31.101:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {},
"slaveDelay": 0,
"votes": 1
},
{
"_id": 1,
"host": "192.168.31.154:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {},
"slaveDelay": 0,
"votes": 1
},
{
"_id": 2,
"host": "192.168.31.159:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {},
"slaveDelay": 0,
"votes": 1
}
],
"settings": {
"chainingAllowed": true,
"heartbeatTimeoutSecs": 10,
"getLastErrorModes": {},
"getLastErrorDefaults": {
"w": 1,
"wtimeout": 0
}
}
}
#赋予权重
szss:PRIMARY> cfgs.members[0].priority=2
2
szss:PRIMARY> rs.conf()
{
"_id": "szss",
"version": 3,
"members": [
{
"_id": 0,
"host": "192.168.31.101:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {},
"slaveDelay": 0,
"votes": 1
},
{
"_id": 1,
"host": "192.168.31.154:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {},
"slaveDelay": 0,
"votes": 1
},
{
"_id": 2,
"host": "192.168.31.159:27017",
"arbiterOnly": false,
"buildIndexes": true,
"hidden": false,
"priority": 1,
"tags": {},
"slaveDelay": 0,
"votes": 1
}
],
"settings": {
"chainingAllowed": true,
"heartbeatTimeoutSecs": 10,
"getLastErrorModes": {},
"getLastErrorDefaults": {
"w": 1,
"wtimeout": 0
}
}
}
#重新读取 初始化
szss:PRIMARY> rs.reconfig(cfgs)
{ "ok" : 1 }

进入到浏览器 输入 192.168.31.101:28017 如果是27017将访问不了，需要将启动端口号加上1000



测试

测试一下集群是否可用：

关掉主节点，模拟主节点宕机的情景

[root@rabbitmq-node1 bin]# ps -ef | grep mongod
root       2494      1  0 15:58 ?        00:00:17 ./mongod --rest -f ../conf/master.config
root       3344   2382  0 16:30 pts/0    00:00:00 grep mongod
[root@rabbitmq-node1 bin]# kill -9 2494

看一下备节点的状态：



这里可以看到备节点已经变为主节点。

szss:PRIMARY>
for(var i=0;i<10000;i++)db.test.insert(
{
"name": "lxc"+i,
"age": 22,
"sex": "man"
}
)
WriteResult({ "nInserted" : 1 })

在主节点插入1w条数据。再看看备节点。

szss:SECONDARY> use admin
switched to db admin
szss:SECONDARY> use test
switched to db test
szss:SECONDARY> db.getMongo().setSlaveOk()
szss:SECONDARY> db.test.count()
10000
szss:SECONDARY>

szss:SECONDARY> use test
switched to db test
szss:SECONDARY> db.getMongo().setSlaveOk()
szss:SECONDARY> db.test.count()
10000

均已经同步过来了。

到这里就基本完成搭建了。mongodb的中文社区也很不错。而且它的官方文档写的也很不错。一起学习吧~

借鉴大神的文章：

http://www.cnblogs.com/zhanjindong/category/507349.html

http://www.mongoing.com/

http://blog.csdn.net/luonanqin/article/details/8497860