MongoDB 主从复制及自动故障转移

【推荐】2019 Java 开发者跳槽指南.pdf(吐血整理) >>>

1、MongoDB 主从复制

MongoDB复制是将数据同步在多个服务器的过程。

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。

官方文档 https://docs.mongodb.com/manual/replication/

1.1 什么是复制?

保障数据的安全性数据高可用性 (24*7) 灾难恢复无需停机维护（如备份，重建索引，压缩）分布式读取数据

1.2 MongoDB复制原理

mongodb的复制至少需要两个节点。其中一个是主节点，负责处理客户端请求，其余的都是从节点，负责复制主节点上的数据。

mongodb各个节点常见的搭配方式为：一主一从、一主多从。

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。

MongoDB复制结构图如下所示：

以上结构图中，客户端从主节点读取数据，在客户端写入数据到主节点时，主节点与从节点进行数据交互保障数据的一致性。

1.3 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

1.4 MongoDB副本集设置

1、关闭正在运行的MongoDB服务器。

service mongod stop

2.节点建点

首先需要去你选择的mongodb数据文件存放的文件夹新建三个数据库，用来模拟三台不通的机器，博主的路径如下

mkdir -p /data/db/node1
mkdir -p /data/db/node2
mkdir -p /data/db/node3

3.启动三个数据库（dbpath），并且端口（－－port 1000x），集群名称（－－replSet gabriel），关闭日志选项（－－nojournal），守护进程方式启动,会自动拉起（－－fork），日志目录（－－logpath）

mongod --dbpath /data/db/node1 --port 10001 --replSet gabriel --nojournal --fork --logpath /data/db/node1.log
mongod --dbpath /data/db/node2 --port 10002 --replSet gabriel --nojournal --fork --logpath /data/db/node2.log
mongod --dbpath /data/db/node3 --port 10003 --replSet gabriel --nojournal --fork --logpath /data/db/node3.log

4.顺便连接一个服务器，做初始化操作，这里博主连入10001端口

终端下进入

mongo localhost:10001

进入后输入初始化方法

MongoDB Enterprise gabriel:OTHER> rs.initiate({_id:"gabriel",members:[
{_id:1,host:"localhost:10001"},
{_id:2,host:"localhost:10002"},
{_id:3,host:"localhost:10003"},
]})

收到如下信息就成功了。

{
	"ok" : 1,
	"operationTime" : Timestamp(1517221411, 1),
	"$clusterTime" : {
		"clusterTime" : Timestamp(1517221411, 1),
		"signature" : {
			"hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
			"keyId" : NumberLong(0)
		}
	}
}
MongoDB Enterprise gabriel:OTHER>

此时会发现终端上的输出已经有了变化。

//从单个一个
>
//变成了
gabriel:OTHER>

5.查询状态

MongoDB Enterprise gabriel:OTHER> rs.status()
{
	"set" : "gabriel",
	"date" : ISODate("2018-01-29T10:33:21.227Z"),
	"myState" : 1,
	"term" : NumberLong(1),
	"heartbeatIntervalMillis" : NumberLong(2000),
	"optimes" : {
		"lastCommittedOpTime" : {
			"ts" : Timestamp(1517221984, 1),
			"t" : NumberLong(1)
		},
		"readConcernMajorityOpTime" : {
			"ts" : Timestamp(1517221984, 1),
			"t" : NumberLong(1)
		},
		"appliedOpTime" : {
			"ts" : Timestamp(1517221994, 1),
			"t" : NumberLong(1)
		},
		"durableOpTime" : {
			"ts" : Timestamp(1517221994, 1),
			"t" : NumberLong(1)
		}
	},
	"members" : [
		{
			"_id" : 1,
			"name" : "localhost:10001",
			"health" : 1,
			"state" : 1,
			"stateStr" : "PRIMARY",
			"uptime" : 659,
			"optime" : {
				"ts" : Timestamp(1517221994, 1),
				"t" : NumberLong(1)
			},
			"optimeDate" : ISODate("2018-01-29T10:33:14Z"),
			"electionTime" : Timestamp(1517221422, 1),
			"electionDate" : ISODate("2018-01-29T10:23:42Z"),
			"configVersion" : 1,
			"self" : true
		},
		{
			"_id" : 2,
			"name" : "localhost:10002",
			"health" : 1,
			"state" : 2,
			"stateStr" : "SECONDARY",
			"uptime" : 589,
			"optime" : {
				"ts" : Timestamp(1517221994, 1),
				"t" : NumberLong(1)
			},
			"optimeDurable" : {
				"ts" : Timestamp(1517221984, 1),
				"t" : NumberLong(1)
			},
			"optimeDate" : ISODate("2018-01-29T10:33:14Z"),
			"optimeDurableDate" : ISODate("2018-01-29T10:33:04Z"),
			"lastHeartbeat" : ISODate("2018-01-29T10:33:20.972Z"),
			"lastHeartbeatRecv" : ISODate("2018-01-29T10:33:19.923Z"),
			"pingMs" : NumberLong(0),
			"syncingTo" : "localhost:10001",
			"configVersion" : 1
		},
		{
			"_id" : 3,
			"name" : "localhost:10003",
			"health" : 1,
			"state" : 2,
			"stateStr" : "SECONDARY",
			"uptime" : 589,
			"optime" : {
				"ts" : Timestamp(1517221994, 1),
				"t" : NumberLong(1)
			},
			"optimeDurable" : {
				"ts" : Timestamp(1517221984, 1),
				"t" : NumberLong(1)
			},
			"optimeDate" : ISODate("2018-01-29T10:33:14Z"),
			"optimeDurableDate" : ISODate("2018-01-29T10:33:04Z"),
			"lastHeartbeat" : ISODate("2018-01-29T10:33:20.972Z"),
			"lastHeartbeatRecv" : ISODate("2018-01-29T10:33:19.921Z"),
			"pingMs" : NumberLong(0),
			"syncingTo" : "localhost:10001",
			"configVersion" : 1
		}
	],
	"ok" : 1,
	"operationTime" : Timestamp(1517221994, 1),
	"$clusterTime" : {
		"clusterTime" : Timestamp(1517221994, 1),
		"signature" : {
			"hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
			"keyId" : NumberLong(0)
		}
	}
}

在返回中，参数set后面为集群名称，每个members下面可以看到他们各自的情况，其中stateStr是角色，主节点为（PRIMARY）。

6.进入主节点插入数据，进入从节点查看数据

博主主节点在10001接口

mongo localhost:10001

插入数据

MongoDB Enterprise gabriel:PRIMARY> use test
switched to db test
db.col.insert({title: 'MongoDB 教程', 
    description: 'MongoDB 是一个 Nosql 数据库',
    by: '小罗技术笔记-专注于开发技术的研究与知识分享',
    url: 'http://www.yuwowugua.com',
    tags: ['mongodb', 'database', 'NoSQL'],
    likes: 100
})

MongoDB Enterprise gabriel:PRIMARY> db.col.find()
{ "_id" : ObjectId("5a6ef998525d903d07a00cdf"), "title" : "MongoDB 教程", "description" : "MongoDB 是一个 Nosql 数据库", "by" : "小罗技术笔记-专注于开发技术的研究与知识分享", "url" : "http://www.yuwowugua.com", "tags" : [ "mongodb", "database", "NoSQL" ], "likes" : 100 }
MongoDB Enterprise gabriel:PRIMARY>

博主切换从节点10002

mongo localhost:10002

切换到从节点，你会发现使用show dbs 会报错，是因为还没有开启权限，输入rs.slaveOk();就可以顺利访问了。

MongoDB Enterprise gabriel:SECONDARY> show dbs
2018-01-29T10:40:37.362+0000 E QUERY    [thread1] Error: listDatabases failed:{
	"operationTime" : Timestamp(1517222434, 1),
	"ok" : 0,
	"errmsg" : "not master and slaveOk=false",
	"code" : 13435,
	"codeName" : "NotMasterNoSlaveOk",
	"$clusterTime" : {
		"clusterTime" : Timestamp(1517222434, 1),
		"signature" : {
			"hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
			"keyId" : NumberLong(0)
		}
	}
} :
_getErrorWithCode@src/mongo/shell/utils.js:25:13
Mongo.prototype.getDBs@src/mongo/shell/mongo.js:65:1
shellHelper.show@src/mongo/shell/utils.js:813:19
shellHelper@src/mongo/shell/utils.js:703:15
@(shellhelp2):1:1
MongoDB Enterprise gabriel:SECONDARY>
MongoDB Enterprise gabriel:SECONDARY> rs.slaveOk()

再次查看

MongoDB Enterprise gabriel:SECONDARY> show dbs
admin   0.000GB
config  0.000GB
local   0.000GB
test    0.000GB
MongoDB Enterprise gabriel:SECONDARY>

切到test 库，查看数据已经同步过来了

MongoDB Enterprise gabriel:SECONDARY> use test
switched to db test
MongoDB Enterprise gabriel:SECONDARY> db.col.find()
{ "_id" : ObjectId("5a6ef998525d903d07a00cdf"), "title" : "MongoDB 教程", "description" : "MongoDB 是一个 Nosql 数据库", "by" : "小罗技术笔记-专注于开发技术的研究与知识分享", "url" : "http://www.yowowugua.com", "tags" : [ "mongodb", "database", "NoSQL" ], "likes" : 100 }
MongoDB Enterprise gabriel:SECONDARY>

以上就是简单的主从复制建立过程，现在已经可以在从服务器看到主服务器插入的数据了。

切换从节点10003 一样的问题

删除从节点

rs.remove('ip:port')

关闭主服务器后，再重新启动，会发现原来的从服务器变为了从服务器，新启动的服务器（原来的从服务器）变为了从服务器

2、 MongoDB 自动故障转移

首先通过 rs.status() 查看，可以看到主节点是10001，主节点"name" :

"localhost:10001", "stateStr" : "PRIMARY" 接下来停止 10001 主节点，测试故障切换

MongoDB Enterprise gabriel:PRIMARY>  rs.status()
{
	"set" : "gabriel",
	"date" : ISODate("2018-01-30T02:39:58.468Z"),
	"myState" : 1,
	"term" : NumberLong(1),
	"heartbeatIntervalMillis" : NumberLong(2000),
	"optimes" : {
		"lastCommittedOpTime" : {
			"ts" : Timestamp(1517279986, 1),
			"t" : NumberLong(1)
		},
		"readConcernMajorityOpTime" : {
			"ts" : Timestamp(1517279986, 1),
			"t" : NumberLong(1)
		},
		"appliedOpTime" : {
			"ts" : Timestamp(1517279996, 1),
			"t" : NumberLong(1)
		},
		"durableOpTime" : {
			"ts" : Timestamp(1517279996, 1),
			"t" : NumberLong(1)
		}
	},
	"members" : [
		{
			"_id" : 1,
			"name" : "localhost:10001",
			"health" : 1,
			"state" : 1,
			"stateStr" : "PRIMARY",
			"uptime" : 58656,
			"optime" : {
				"ts" : Timestamp(1517279996, 1),
				"t" : NumberLong(1)
			},
			"optimeDate" : ISODate("2018-01-30T02:39:56Z"),
			"electionTime" : Timestamp(1517221422, 1),
			"electionDate" : ISODate("2018-01-29T10:23:42Z"),
			"configVersion" : 1,
			"self" : true
		},
		{
			"_id" : 2,
			"name" : "localhost:10002",
			"health" : 1,
			"state" : 2,
			"stateStr" : "SECONDARY",
			"uptime" : 58586,
			"optime" : {
				"ts" : Timestamp(1517279996, 1),
				"t" : NumberLong(1)
			},
			"optimeDurable" : {
				"ts" : Timestamp(1517279986, 1),
				"t" : NumberLong(1)
			},
			"optimeDate" : ISODate("2018-01-30T02:39:56Z"),
			"optimeDurableDate" : ISODate("2018-01-30T02:39:46Z"),
			"lastHeartbeat" : ISODate("2018-01-30T02:39:58.289Z"),
			"lastHeartbeatRecv" : ISODate("2018-01-30T02:39:57.220Z"),
			"pingMs" : NumberLong(0),
			"syncingTo" : "localhost:10001",
			"configVersion" : 1
		},
		{
			"_id" : 3,
			"name" : "localhost:10003",
			"health" : 0,
			"state" : 8,
			"stateStr" : "(not reachable/healthy)",
			"uptime" : 0,
			"optime" : {
				"ts" : Timestamp(0, 0),
				"t" : NumberLong(-1)
			},
			"optimeDurable" : {
				"ts" : Timestamp(0, 0),
				"t" : NumberLong(-1)
			},
			"optimeDate" : ISODate("1970-01-01T00:00:00Z"),
			"optimeDurableDate" : ISODate("1970-01-01T00:00:00Z"),
			"lastHeartbeat" : ISODate("2018-01-30T02:39:58.304Z"),
			"lastHeartbeatRecv" : ISODate("2018-01-30T02:39:21.208Z"),
			"pingMs" : NumberLong(0),
			"lastHeartbeatMessage" : "Connection refused",
			"configVersion" : -1
		}
	],
	"ok" : 1,
	"operationTime" : Timestamp(1517279996, 1),
	"$clusterTime" : {
		"clusterTime" : Timestamp(1517279996, 1),
		"signature" : {
			"hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
			"keyId" : NumberLong(0)
		}
	}
}
MongoDB Enterprise gabriel:PRIMARY>

连接到主节点

mongo localhost:10001

显示所有数据库

MongoDB Enterprise gabriel:PRIMARY> show dbs
admin   0.000GB
config  0.000GB
local   0.000GB
test    0.000GB

切换到admin

MongoDB Enterprise gabriel:PRIMARY> use admin
switched to db admin

停止数据库，必须进入 admin 库

MongoDB Enterprise gabriel:PRIMARY> db.shutdownServer()

响应

2018-01-30T02:51:34.503+0000 I NETWORK  [thread1] trying reconnect to localhost:10001 (127.0.0.1) failed
2018-01-30T02:51:35.398+0000 I NETWORK  [thread1] Socket recv() Connection reset by peer 127.0.0.1:10001
2018-01-30T02:51:35.398+0000 I NETWORK  [thread1] SocketException: remote: (NONE):0 error: SocketException socket exception [RECV_ERROR] server [127.0.0.1:10001] 
2018-01-30T02:51:35.399+0000 I NETWORK  [thread1] reconnect localhost:10001 (127.0.0.1) failed failed 
2018-01-30T02:51:35.404+0000 I NETWORK  [thread1] trying reconnect to localhost:10001 (127.0.0.1) failed
2018-01-30T02:51:35.404+0000 W NETWORK  [thread1] Failed to connect to 127.0.0.1:10001, in(checking socket for error after poll), reason: Connection refused
2018-01-30T02:51:35.404+0000 I NETWORK  [thread1] reconnect localhost:10001 (127.0.0.1) failed failed 
MongoDB Enterprise >

查看是否真的停止了,发现已经没有10001 节点进程了

root@admin-2:# ps -ef | grep mongo
root      5554     1  0 Jan29 ?        00:03:34 mongod --dbpath /data/db/node2 --port 10002 --replSet gabriel --nojournal --fork --logpath /data/db/node2.log
root     12284     1  0 02:43 ?        00:00:02 mongod --dbpath /data/db/node3 --port 10003 --replSet gabriel --nojournal --fork --logpath /data/db/node3.log
root     12436  5132  0 02:53 pts/1    00:00:00 grep --color=auto mongo
root@admin-2:# netstat -nltp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address           Foreign Address         State       PID/Program name
tcp        0      0 127.0.0.1:10002         0.0.0.0:*               LISTEN      5554/mongod     
tcp        0      0 127.0.0.1:10003         0.0.0.0:*               LISTEN      12284/mongod          
root@admin-2:/data/db#

查看是否故障专业

root@admin-2:# mongo localhost:10001

查看主从状态

MongoDB Enterprise gabriel:SECONDARY> rs.status()
{
	"set" : "gabriel",
	"date" : ISODate("2018-01-30T02:56:48.074Z"),
	"myState" : 2,
	"term" : NumberLong(2),
	"syncingTo" : "localhost:10003",
	"heartbeatIntervalMillis" : NumberLong(2000),
	"optimes" : {
		"lastCommittedOpTime" : {
			"ts" : Timestamp(1517280995, 1),
			"t" : NumberLong(2)
		},
		"readConcernMajorityOpTime" : {
			"ts" : Timestamp(1517280995, 1),
			"t" : NumberLong(2)
		},
		"appliedOpTime" : {
			"ts" : Timestamp(1517281005, 1),
			"t" : NumberLong(2)
		},
		"durableOpTime" : {
			"ts" : Timestamp(1517280995, 1),
			"t" : NumberLong(2)
		}
	},
	"members" : [
		{
			"_id" : 1,
			"name" : "localhost:10001",
			"health" : 0,
			"state" : 8,
			"stateStr" : "(not reachable/healthy)",
			"uptime" : 0,
			"optime" : {
				"ts" : Timestamp(0, 0),
				"t" : NumberLong(-1)
			},
			"optimeDurable" : {
				"ts" : Timestamp(0, 0),
				"t" : NumberLong(-1)
			},
			"optimeDate" : ISODate("1970-01-01T00:00:00Z"),
			"optimeDurableDate" : ISODate("1970-01-01T00:00:00Z"),
			"lastHeartbeat" : ISODate("2018-01-30T02:56:47.605Z"),
			"lastHeartbeatRecv" : ISODate("2018-01-30T02:51:34.519Z"),
			"pingMs" : NumberLong(0),
			"lastHeartbeatMessage" : "Connection refused",
			"configVersion" : -1
		},
		{
			"_id" : 2,
			"name" : "localhost:10002",
			"health" : 1,
			"state" : 2,
			"stateStr" : "SECONDARY",
			"uptime" : 59660,
			"optime" : {
				"ts" : Timestamp(1517281005, 1),
				"t" : NumberLong(2)
			},
			"optimeDate" : ISODate("2018-01-30T02:56:45Z"),
			"syncingTo" : "localhost:10003",
			"configVersion" : 1,
			"self" : true
		},
		{
			"_id" : 3,
			"name" : "localhost:10003",
			"health" : 1,
			"state" : 1,
			"stateStr" : "PRIMARY",
			"uptime" : 784,
			"optime" : {
				"ts" : Timestamp(1517281005, 1),
				"t" : NumberLong(2)
			},
			"optimeDurable" : {
				"ts" : Timestamp(1517281005, 1),
				"t" : NumberLong(2)
			},
			"optimeDate" : ISODate("2018-01-30T02:56:45Z"),
			"optimeDurableDate" : ISODate("2018-01-30T02:56:45Z"),
			"lastHeartbeat" : ISODate("2018-01-30T02:56:46.486Z"),
			"lastHeartbeatRecv" : ISODate("2018-01-30T02:56:47.147Z"),
			"pingMs" : NumberLong(0),
			"electionTime" : Timestamp(1517280703, 1),
			"electionDate" : ISODate("2018-01-30T02:51:43Z"),
			"configVersion" : 1
		}
	],
	"ok" : 1,
	"operationTime" : Timestamp(1517281005, 1),
	"$clusterTime" : {
		"clusterTime" : Timestamp(1517281005, 1),
		"signature" : {
			"hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
			"keyId" : NumberLong(0)
		}
	}
}

MongoDB Enterprise gabriel:SECONDARY>

发现 “name” : “localhost:10001”,”stateStr” : “(not reachable/healthy)”, 健康状态已经是“无法访问状态了”

主节点已经切换成 10003 节点了

"_id" : 3,
"name" : "localhost:10003",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",

重启节点10001

mongod --dbpath /data/db/node1 --port 10001 --replSet gabriel --nojournal --fork --logpath /data/db/node1.log

来源：oschina

链接：https://my.oschina.net/u/4301644/blog/3145763

标签

Gabriel

mongodb

MongoDB 主从复制及 自动故障转移

MongoDB 主从复制及自动故障转移