MongDB集群容灾方案步骤

栏目: 数据库 · 发布时间: 5年前

内容简介：MongoDB复制集优/特点支持大数据量、高扩展性、高性能、灵活数据模型、高可用性。

MongoDB复制集

优/特点

支持大数据量、高扩展性、高性能、灵活数据模型、高可用性。

同步机制

数据复制的目的是使数据得到最大的可用性，避免单点故障引起的整站不能访问的情况的发生，Mongodb的副本集在同一时刻只有一台服务器是可以写的，副本集的主从复制也是一个异步同步的过程，是slave端从primary端获取日志，然后在自己身上完全顺序的执行日志所记录的各种操作（该日志是不记录查询操作的），这个日志就是local数据库中的oplog.rs表，默认在64位机器上这个表是比较大的，占磁盘大小的5%，oplog.rs的大小可以在启动参数中设定：--oplogSize 1000,单位是M。

鉴于双机房容灾意外情况可能放生，本方案选择故障时人工介入转移或恢复，不加入仲裁节点。其中A机房为主机房一个primary+2个Secondary节点，B机房作为灾备机房，2个Secondary节点。最坏当主机房挂掉时通过权重来启动B机房某节点为Primary，继续提供服务。

环境规划

A机房角色 B机房角色

192.168.70.214 Primary 192.168.71.214 Secondary 3 复制集节点 3

192.168.70.215 Secondary 1 复制集节点 1 192.168.71.215 Secondary 4 复制集节点 4

192.168.70.216 Secondary 2 复制集节点 2

架构示意图

其中下面是主机房断电断网时的故障转移示意图。

安装配置

这里所有节点目录创建一致，方便管理维护，从配置文件来判断各节点的角色。

创建目录

--为 MongoDB 创建软件、数据、日志目录，默认情况下它将数据存储在/mgdata

[root@test153 /]# mkdir -p /mgdb/mongodbtest/replset/data

[root@test153 /]# mkdir /mgdata

[root@test153 /]# mkdir /mglog

上传介质

sftp> cd /mgdb

sftp> put mongodb-linux-x86_64-2.2.3.tgz.tar

解压

[root@test153 /]# cd /mgdb

$ tar -xvf mongodb-linux-x86_64-2.2.3.tgz.tar

[root@test153 mgdb]# mv mongodb-linux-x86_64-2.2.3 mongodb

服务启动

每个节点都要执行

cd /root/mongodb/bin

192.168.70.214

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27011 --oplogSize 2048 --logpath /root/log27011/log27011.log &

./mongo 192.168.70.214:27017

192.168.70.215

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27012 --oplogSize 2048 --logpath /root/log27012/log27012.log &

./mongo 192.168.70.215:27017

192.168.70.216

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27013 --oplogSize 2048 --logpath /root/log27013/log27013.log &

./mongo 192.168.70.216:27017

192.168.71.214

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27017 --oplogSize 2048 --logpath /root/log27017/log27017.log &

./mongo 192.168.71.214:27017

192.168.71.215

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27018 --oplogSize 2048 --logpath /root/log27018/log27018.log &

./mongo 192.168.71.215:27017

分别通过 tail -f /root/log27011/log27011.log 来观察分析个节点运行情况

复制集配置

在任何一台mongodb实例上登录，进入admin库，执行config命令，配置相应权重

[root@localhost bin]# pwd

/root/mongodb/bin

[root@localhost bin]# ./mongo 192.168.70.214:27017

MongoDB shell version: 2.2.3

connecting to: test

Welcome to the MongoDB shell.

For interactive help, type "help".

For more comprehensive documentation, see

http://docs.mongodb.org/

Questions? Try the support group

http://groups.google.com/group/mongodb-user

> use admin

switched to db admin

> config = { _id:"repset", members:[

... {_id:0,host:"192.168.70.214:27017",priority:10},

... {_id:1,host:"192.168.70.215:27017",priority:7},

... {_id:2,host:"192.168.70.216:27017",priority:6},

... {_id:3,host:"192.168.71.214:27017",priority:9}]

... {_id:4,host:"192.168.71.215:27017",priority:8}]

... }

{

"_id" : "repset",

"members" : [

{

"_id" : 0,

"host" : "192.168.70.214:27017",

"priority" : 10

{

"_id" : 0,

"host" : "192.168.70.215:27017",

"priority" : 7

{

"_id" : 0,

"host" : "192.168.70.216:27017",

"priority" : 6

{

"_id" : 1,

"host" : "192.168.71.214:27017",

"priority" : 9

{

"_id" : 2,

"host" : "192.168.71.215:27017",

"priority" : 8

}

]

}

--查看

repset:PRIMARY> rs.conf()

{

"_id" : "repset",

"version" : 38349,

"members" : [

{

"_id" : 4,

"host" : "192.168.71.214:27017",

"priority" : 9

{

"_id" : 5,

"host" : "192.168.71.215:27017",

"priority" : 8

{

"_id" : 6,

"host" : "192.168.70.214:27017",

"priority" : 10

{

"_id" : 7,

"host" : "192.168.70.215:27017",

"priority" : 7

{

"_id" : 8,

"host" : "192.168.70.216:27017",

"priority" : 6

}

]

}

初始化副本集配置

> rs.initiate(config);

{

"info" : "Config now saved locally. Should come online in about a minute.",

"ok" : 1

}

初始需要一点时间同步

查看集群节点状态

repset:PRIMARY> rs.status()

{

"set" : "repset",

"date" : ISODate("2018-11-09T07:55:04Z"),

"myState" : 1,

"members" : [

{

"_id" : 4,

"name" : "192.168.71.214:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 1104,

"optime" : Timestamp(1541749003000, 1),

"optimeDate" : ISODate("2018-11-09T07:36:43Z"),

"lastHeartbeat" : ISODate("2018-11-09T07:55:03Z"),

"pingMs" : 0

{

"_id" : 5,

"name" : "192.168.71.215:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 1104,

"optime" : Timestamp(1541749003000, 1),

"optimeDate" : ISODate("2018-11-09T07:36:43Z"),

"lastHeartbeat" : ISODate("2018-11-09T07:55:03Z"),

"pingMs" : 0

{

"_id" : 6,

"name" : "192.168.70.214:27017",

"health" : 1,

"state" : 1,

"stateStr" : "PRIMARY",

"uptime" : 1680,

"optime" : Timestamp(1541749003000, 1),

"optimeDate" : ISODate("2018-11-09T07:36:43Z"),

"self" : true

{

"_id" : 7,

"name" : "192.168.70.215:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 1104,

"optime" : Timestamp(1541749003000, 1),

"optimeDate" : ISODate("2018-11-09T07:36:43Z"),

"lastHeartbeat" : ISODate("2018-11-09T07:55:03Z"),

"pingMs" : 0

{

"_id" : 8,

"name" : "192.168.70.216:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 1104,

"optime" : Timestamp(1541749003000, 1),

"optimeDate" : ISODate("2018-11-09T07:36:43Z"),

"lastHeartbeat" : ISODate("2018-11-09T07:55:03Z"),

"pingMs" : 0

}

"ok" : 1

}

repset:PRIMARY>

查看后台日志

[root@oracle_master ~]# tail -f /mgdata/mongodb/log27017/mongod.log

验证复制集数据一致性

先进去主库primary的mongodb上，录入数据

如

repset:PRIMARY> use dinpay

switched to db dinpay

repset:PRIMARY> db.dinpay.insert({"test1108":"xiawu1"})

repset:PRIMARY> db.getMongo().setSlaveOk();

去另一个备库上验证数据

repset:SECONDARY> db.dinpay.find()

{ "_id" : ObjectId("5bd676e97e238f7b0dddfb0d"), "MongoDB TEST" : "dinpay" }

{ "_id" : ObjectId("5bd823b65b237ec32e664db2"), "mdbtest" : "zgy20181030" }

{ "_id" : ObjectId("5be53e2c60074628c8509830"), "test1108" : "xiawu1" }

断电断网模拟

断电：直接kill mongdb进程

断网：开启某一机房的防火墙限制机房间通讯

B机房断电断网

192.168.71.214、192.16871.215断电断网后个节点状态

repset:PRIMARY> rs.status()

{

"set" : "repset",

"date" : ISODate("2018-11-09T08:02:36Z"),

"myState" : 1,

"members" : [

{

"_id" : 4,

"name" : "192.168.71.214:27017",

"health" : 0,

"state" : 8,

"stateStr" : "(not reachable/healthy)",

"uptime" : 0,

"optime" : Timestamp(1541750316000, 1),

"optimeDate" : ISODate("2018-11-09T07:58:36Z"),

"lastHeartbeat" : ISODate("2018-11-09T08:01:59Z"),

"pingMs" : 0,

"errmsg" : "socket exception [CONNECT_ERROR] for 192.168.71.214:27017"

{

"_id" : 5,

"name" : "192.168.71.215:27017",

"health" : 0,

"state" : 8,

"stateStr" : "(not reachable/healthy)",

"uptime" : 0,

"optime" : Timestamp(1541750316000, 1),

"optimeDate" : ISODate("2018-11-09T07:58:36Z"),

"lastHeartbeat" : ISODate("2018-11-09T08:01:57Z"),

"pingMs" : 0,

"errmsg" : "socket exception [CONNECT_ERROR] for 192.168.71.215:27017"

{

"_id" : 6,

"name" : "192.168.70.214:27017",

"health" : 1,

"state" : 1,

"stateStr" : "PRIMARY",

"uptime" : 2132,

"optime" : Timestamp(1541750316000, 1),

"optimeDate" : ISODate("2018-11-09T07:58:36Z"),

"self" : true

{

"_id" : 7,

"name" : "192.168.70.215:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 1556,

"optime" : Timestamp(1541750316000, 1),

"optimeDate" : ISODate("2018-11-09T07:58:36Z"),

"lastHeartbeat" : ISODate("2018-11-09T08:02:35Z"),

"pingMs" : 0

{

"_id" : 8,

"name" : "192.168.70.216:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 1556,

"optime" : Timestamp(1541750316000, 1),

"optimeDate" : ISODate("2018-11-09T07:58:36Z"),

"lastHeartbeat" : ISODate("2018-11-09T08:02:35Z"),

"pingMs" : 0

}

"ok" : 1

}

repset:PRIMARY>

结论：A机房运行正常。

A机房断电断网

192.168.70.214(PRI)、192.168.70.215、192.168.70.216

登录B机房任一台节点强制reconfig恢复副本集，仅保留活动着的节点

repset:SECONDARY> use admin

switched to db admin

--查看现有配置，其中70网段3个节点都已死掉了

repset:SECONDARY> cfg=rs.conf()

{

"_id" : "repset",

"version" : 79,

"members" : [

{

"_id" : 4,

"host" : "192.168.71.214:27017",

"priority" : 10

{

"_id" : 5,

"host" : "192.168.71.215:27017",

"priority" : 9

{

"_id" : 7,

"host" : "192.168.70.214:27017",

"priority" : 11

{

"_id" : 8,

"host" : "192.168.70.215:27017",

"priority" : 6

{

"_id" : 13,

"host" : "192.168.70.216:27017",

"priority" : 5

}

]

}

--只保留活着的节点

repset:SECONDARY> cfg.members = [cfg.members[0], cfg.members[1]]

[

{

"_id" : 4,

"host" : "192.168.71.214:27017",

"priority" : 10

{

"_id" : 5,

"host" : "192.168.71.215:27017",

"priority" : 9

}

]

--强制启动并新产生一个PRIMARY组成2节点的备份集

repset:SECONDARY> rs.reconfig(cfg, {force :true })

{ "ok" : 1 }

repset:SECONDARY> rs.status()

{

"set" : "repset",

"date" : ISODate("2018-11-09T03:45:29Z"),

"myState" : 1,

"members" : [

{

"_id" : 4,

"name" : "192.168.71.214:27017",

"health" : 1,

"state" : 1,

"stateStr" : "PRIMARY",

"uptime" : 69133,

"optime" : Timestamp(1541663971000, 1),

"optimeDate" : ISODate("2018-11-08T07:59:31Z"),

"self" : true

{

"_id" : 5,

"name" : "192.168.71.215:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 8,

"optime" : Timestamp(1541663971000, 1),

"optimeDate" : ISODate("2018-11-08T07:59:31Z"),

"lastHeartbeat" : ISODate("2018-11-09T03:45:29Z"),

"pingMs" : 0

}

"ok" : 1

}

repset:PRIMARY>

--检查数据，先前数据仍存在

repset:PRIMARY> use test

switched to db test

repset:PRIMARY> show collections

system.indexes

test

Testdb

--状态查询

repset:PRIMARY> rs.conf()

{

"_id" : "repset",

"version" : 38342,

"members" : [

{

"_id" : 4,

"host" : "192.168.71.214:27017",

"priority" : 10

{

"_id" : 5,

"host" : "192.168.71.215:27017",

"priority" : 9

}

]

}

结论：B机房强制启动，变成新的备份集

尽管已产生新的备份集，但不能保证断电断网瞬间存在旧PRIMARY未同步至各从节点而导致数据丢失的可能性。

恢复初始状态

Kill并重启A机房个节点mongdb

192.168.70.214

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27011 --oplogSize 2048 --logpath /root/log27011/log27011.log &

192.168.70.215

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27012 --oplogSize 2048 --logpath /root/log27012/log27012.log &

192.168.70.216

/root/mongodb/bin/mongod --replSet repset --port 27017 --dbpath /root/data27013 --oplogSize 2048 --logpath /root/log27013/log27013.log &

将A机房个节点加入新备份集（B机房），并对A机房某节点提权升为新的PRIMARY，恢复至断电断网前的状态

epset:PRIMARY> use admin

switched to db admin

repset:PRIMARY> cfg=rs.conf()

repset:PRIMARY> cfg.members[XX].priority = 8

repset:PRIMARY> rs.reconfig(cfg)

////////////////////////////////////////////////////////////////////////////////

主机房挂了再恢复测试。。。比上面恢复详细

--A机房(主机房) 192.168.70.214（主）/192.168.70.215/192.168.70.216断电断网，B机房强制重启后成为了新的集群，现在将2机房重新回到初始状态，首先要确认之前各节点都是什么角色

--加节点

--设权重

repset:PRIMARY> use admin

switched to db admin

repset:PRIMARY> rs.add("192.168.70.214:27017")

{ "ok" : 1 }

repset:PRIMARY> rs.add("192.168.70.215:27017")

{ "ok" : 1 }

repset:PRIMARY> rs.add("192.168.70.216:27017")

{ "ok" : 1 }

repset:PRIMARY>

repset:PRIMARY> cfg=rs.conf()

{

"_id" : "repset",

"version" : 63044,

"members" : [

{

"_id" : 4,

"host" : "192.168.71.214:27017",

"priority" : 9

{

"_id" : 5,

"host" : "192.168.71.215:27017",

"priority" : 8

{

"_id" : 6,

"host" : "192.168.70.214:27017"

{

"_id" : 7,

"host" : "192.168.70.215:27017"

{

"_id" : 8,

"host" : "192.168.70.216:27017"

}

]

}

repset:PRIMARY> cfg.members[2].priority = 11

repset:PRIMARY> cfg.members[3].priority = 6

repset:PRIMARY> cfg.members[4].priority = 5

repset:PRIMARY> rs.reconfig(cfg)

Mon Nov 12 16:04:10 DBClientCursor::init call() failed

Mon Nov 12 16:04:10 query failed : admin.$cmd { replSetReconfig: { _id: "repset", version: 63045, members: [ { _id: 4, host: "192.168.71.214:27017", priority: 9.0 }, { _id: 5, host: "192.168.71.215:27017", priority: 8.0 }, { _id: 6, host: "192.168.70.214:27017", priority: 11.0 }, { _id: 7, host: "192.168.70.215:27017", priority: 6.0 }, { _id: 8, host: "192.168.70.216:27017", priority: 5.0 } ] } } to: 192.168.71.214:27017

Mon Nov 12 16:04:10 trying reconnect to 192.168.71.214:27017

Mon Nov 12 16:04:10 reconnect 192.168.71.214:27017 ok

reconnected to server after rs command (which is normal)

repset:PRIMARY>

Mon Nov 12 16:04:29 Socket recv() errno:104 Connection reset by peer 192.168.71.214:27017

Mon Nov 12 16:04:29 SocketException: remote: 192.168.71.214:27017 error: 9001 socket exception [1] server [192.168.71.214:27017]

Mon Nov 12 16:04:29 DBClientCursor::init call() failed

Mon Nov 12 16:04:29 query failed : admin.$cmd { replSetGetStatus: 1.0, forShell: 1.0 } to: 192.168.71.214:27017

Mon Nov 12 16:04:37 trying reconnect to 192.168.71.214:27017

Mon Nov 12 16:04:37 reconnect 192.168.71.214:27017 ok

repset:SECONDARY>

repset:SECONDARY> rs.conf()

{

"_id" : "repset",

"version" : 63045,

"members" : [

{

"_id" : 4,

"host" : "192.168.71.214:27017",

"priority" : 9

{

"_id" : 5,

"host" : "192.168.71.215:27017",

"priority" : 8

{

"_id" : 6,

"host" : "192.168.70.214:27017",

"priority" : 11

{

"_id" : 7,

"host" : "192.168.70.215:27017",

"priority" : 6

{

"_id" : 8,

"host" : "192.168.70.216:27017",

"priority" : 5

}

]

}

repset:SECONDARY>

repset:SECONDARY> rs.status()

{

"set" : "repset",

"date" : ISODate("2018-11-12T08:05:16Z"),

"myState" : 2,

"syncingTo" : "192.168.70.214:27017",

"members" : [

{

"_id" : 4,

"name" : "192.168.71.214:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 19426,

"optime" : Timestamp(1542009850000, 1),

"optimeDate" : ISODate("2018-11-12T08:04:10Z"),

"errmsg" : "syncing to: 192.168.70.214:27017",

"self" : true

{

"_id" : 5,

"name" : "192.168.71.215:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 50,

"optime" : Timestamp(1542009850000, 1),

"optimeDate" : ISODate("2018-11-12T08:04:10Z"),

"lastHeartbeat" : ISODate("2018-11-12T08:05:14Z"),

"pingMs" : 0,

"errmsg" : "syncing to: 192.168.70.214:27017"

{

"_id" : 6,

"name" : "192.168.70.214:27017",

"health" : 1,

"state" : 1,

"stateStr" : "PRIMARY",

"uptime" : 64,

"optime" : Timestamp(1542009850000, 1),

"optimeDate" : ISODate("2018-11-12T08:04:10Z"),

"lastHeartbeat" : ISODate("2018-11-12T08:05:14Z"),

"pingMs" : 1

{

"_id" : 7,

"name" : "192.168.70.215:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 64,

"optime" : Timestamp(1542009850000, 1),

"optimeDate" : ISODate("2018-11-12T08:04:10Z"),

"lastHeartbeat" : ISODate("2018-11-12T08:05:14Z"),

"pingMs" : 0,

"errmsg" : "syncing to: 192.168.70.214:27017"

{

"_id" : 8,

"name" : "192.168.70.216:27017",

"health" : 1,

"state" : 2,

"stateStr" : "SECONDARY",

"uptime" : 64,

"optime" : Timestamp(1542009850000, 1),

"optimeDate" : ISODate("2018-11-12T08:04:10Z"),

"lastHeartbeat" : ISODate("2018-11-12T08:05:14Z"),

"pingMs" : 1,

"errmsg" : "syncing to: 192.168.70.214:27017"

}

"ok" : 1

}

repset:SECONDARY>

以上所述就是小编给大家介绍的《MongDB集群容灾方案步骤》，希望对大家有所帮助，如果大家有任何疑问请给我留言，小编会及时回复大家的。在此也非常感谢大家对码农网的支持！

查看所有标签

猜你喜欢:

本站部分资源来源于网络，本站转载出于传递更多信息之目的，版权归原作者或者来源机构所有，如转载稿涉及版权问题，请联系我们。

码农书籍

算法精解

Kyle Loudon / 肖翔、陈舸 / 机械工业出版社 / 2012-8 / 79.00元

本书是数据结构和算法领域的经典之作，十余年来，畅销不衰！全书共分为三部分：第一部分首先介绍了数据结构和算法的概念，以及使用它们的原因和意义，然后讲解了数据结构和算法中最常用的技术——指针和递归，最后还介绍了算法的分析方法，旨在为读者学习这本书打下坚实的基础；第二部分对链表、栈、队列、集合、哈希表、堆、图等常用数据结构进行了深入阐述；第三部分对排序、搜索数值计算、数据压缩、数据加密、图算法、几何算法......一起来看看《算法精解》这本书的介绍吧!

码农工具

MongDB集群容灾方案步骤

算法精解

html转js在线工具

UNIX 时间戳转换