MongoDB复制集的选举原理-创新互联

复制的原理

       复制时基于啊哦做日志 oplog , 相当于 MySQL 中的二进制日志，只记录发生改变的记录。复制是将主节点的oplog 日志同步并应用到其他从节点的过程。

选举的原理

     节点类型分为标准（host）节点 、被动（passive）节点和仲裁（arbiter）节点。

（1）只有标准节点可能被选举为活跃（primary）节点，有选举权。被动节点有完整副本，不可能成为活跃节点，有选举权。仲裁节点不复制数据，不可能成为活跃节点，只有选举权。

（2）标准节点与被动节点的区别：primary 值高者是标准节点，低者则为被动节点。

（3）选举规则是票数高者获胜，primary 是优先权为 0~1000 的值，相当于额为增加 0~1000 的票数。选举结果：票数高者获胜；若票数相同，数据新者获胜。

实验部署

搭建配置4个 MongoDB 实例 mongodb 、mongpdb2 、mongodb3 、mongodb4 。 IP 地址：192.168.213.184

由于MongoDB 数据库安装好后，mongodb 为默认第一个实例，因此只要再搭建2 ,3，4 实例即可

1. 配置复制集

（1）创建数据文件和日志文件存储路径

[root@localhost ~]# mkdir -p /data/mongodb/mongodb{2,3,4}               #创建数据库文件2,3,4
[root@localhost ~]# cd /data/mongodb/
[root@localhost mongodb]# ls
mongodb2 mongodb3 mongodb4
[root@localhost mongodb]# mkdir logs
[root@localhost mongodb]# ls
logs mongodb2 mongodb3 mongodb4
[root@localhost mongodb]# cd logs/
[root@localhost logs]# touch mongodb{2,3,4}.log               #在 /data/mongodb/logs/ 目录下的日志存储路径
[root@localhost logs]# chmod - R 777 /data/mongodb/logs/*.log        #修改目录权限   
[root@localhost logs]# ls -l
总用量 0
-rwxrwxrwx. 1 root root 0 9月 13 18:43 mongodb2.log
-rwxrwxrwx. 1 root root 0 9月 13 18:43 mongodb3.log
-rwxrwxrwx. 1 root root 0 9月 13 18:43 mongodb4.log

（2）配置实例1 的配置文件，启用复制集名称

[root@localhost logs]# vim /etc/mongod.conf
replication:
     replSetName: abc                       #去掉注释，指定复制集名称（名称随便定义）

创建多实例配置文件

[root@localhost logs]# cp -p /etc/mongod.conf /etc/mongod2.conf       复制生成实例2 的配置文件
[root@localhost logs]# vim /etc/mongod2.conf

systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/logs/mongodb2.log            #日志文件位置

storage:
  dbPath: /data/mongodb/mongodb2                       #数据文件位置
  journal:
    enabled: true

port: 27018       #端口号，默认端口号为27017，因此实例2 的端口号要与之区别，实例3和4 依次往后排

配置其他实例，与实例2修改一样，注意端口号

[root@localhost logs]# cp -p /etc/mongod2.conf /etc/mongod3.conf
[root@localhost logs]# cp -p /etc/mongod2.conf /etc/mongod4.conf

（3）启动服务并查看端口是否开启

[root@localhost logs]# mongod -f /etc/mongod.conf –shutdown                     #先关闭实例1 的服务，在开启
killing process with pid: 3997                                 #进程号
[root@localhost logs]# mongod -f /etc/mongod.conf
about to fork child process, waiting until server is ready for connections.
forked process: 4867
child process started successfully, parent exiting

[root@localhost logs]# mongod -f /etc/mongod2.conf                                           #启动实例2 的服务
about to fork child process, waiting until server is ready for connections.
forked process: 4899
child process started successfully, parent exiting

[root@localhost logs]# mongod -f /etc/mongod3.conf                                            #启动实例3 的服务
about to fork child process, waiting until server is ready for connections.
forked process: 4927
child process started successfully, parent exiting

[root@localhost logs]# mongod -f /etc/mongod4.conf                                             #启动实例4 的服务
about to fork child process, waiting until server is ready for connections.
forked process: 4955

查看节点端口是否开启

[root@localhost logs]# netstat -ntap
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address          Foreign Address        State      PID/Program name   
tcp       0     0 0.0.0.0:27019         0.0.0.0:*              LISTEN     4927/mongod        
tcp       0     0 0.0.0.0:27020         0.0.0.0:*              LISTEN     4955/mongod        
tcp       0     0 0.0.0.0:27017         0.0.0.0:*              LISTEN     4867/mongod        
tcp       0     0 0.0.0.0:27018         0.0.0.0:*              LISTEN     4899/mongod

（4）查看各实例能否进入 mongodb 数据库 

[root@localhost logs]# mongo
MongoDB shell version v3.6.7
connecting to: mongodb://127.0.0.1:27017
MongoDB server version: 3.6.7

[root@localhost logs]# mongo --port 27018
MongoDB shell version v3.6.7
connecting to: mongodb://127.0.0.1:27018/
MongoDB server version: 3.6.7
Server has startup warnings:

[root@localhost logs]# mongo --port 27019
MongoDB shell version v3.6.7
connecting to: mongodb://127.0.0.1:27019/
MongoDB server version: 3.6.7
Server has startup warnings:

[root@localhost logs]# mongo --port 27020
MongoDB shell version v3.6.7
connecting to: mongodb://127.0.0.1:27020/
MongoDB server version: 3.6.7

2 . 配置复制集的优先级

（1）登录默认实例 mongo，重新配置4个节点 MongoDB 复制集，设置两个标准节点，有个被动节点和一个仲裁节点，命令如下：

根据优先级确定节点: 优先级为 100的为标准节点，端口号为 27017和27018 ，优先级为0 的为被动节点，端口号为27019；仲裁节点为27020

> cfg={"_id":"abc","members":[{"_id":0,"host":"192.168.213.184:27017","priority":100},{"_id":1,"host":"192.168.213.184:27018","priority":100},{"_id":2,"host":"192.168.213.184:27019","priority":0},{"_id":3,"host":"192.168.213.184:27020","arbiterOnly":true}]}
{
    "_id" : "abc",
    "members" : [
        {
            "_id" : 0,
            "host" : "192.168.213.184:27017",                  #标准节点，优先级为100
            "priority" : 100
        },
        {
            "_id" : 1,
            "host" : "192.168.213.184:27018",                 #标准节点
            "priority" : 100
        },
        {
            "_id" : 2,
            "host" : "192.168.213.184:27019",                  #被动节点，优先级为0
            "priority" : 0
        },
        {
            "_id" : 3,
            "host" : "192.168.213.184:27020",                  #仲裁节点
            "arbiterOnly" : true
        }
    ]
}

> rs.initiate(cfg)                   # 初始化复制集
{
    "ok" : 1,
    "operationTime" : Timestamp(1536819122, 1),
    "$clusterTime" : {
        "clusterTime" : Timestamp(1536819122, 1),
        "signature" : {
            "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
            "keyId" : NumberLong(0)
        }
    }
}

（2）使用命令 rs.isMaster() 查看各节点身份

abc:PRIMARY> rs.isMaster()
{
    "hosts" : [                                        #标准节点 
        "192.168.213.184:27017",
        "192.168.213.184:27018"
    ],
    "passives" : [                                   #被动节点
        "192.168.213.184:27019"
    ],
    "arbiters" : [                                    #仲裁节点
        "192.168.213.184:27020"
    ],
    "setName" : "abc",
    "setVersion" : 1,
    "ismaster" : true,
    "secondary" : false,
    "primary" : "192.168.213.184:27017",

（3）此时默认实例已经被选举为活跃（primary）节点，可以在数据库中插入信息了。

abc:PRIMARY> use school                                                #活跃节点（primory）
switched to db school
abc:PRIMARY> db.info.insert({"id":1,"name":"lili"})               #插入两条信息
WriteResult({ "nInserted" : 1 })
abc:PRIMARY> db.info.insert({"id":2,"name":"tom"})
WriteResult({ "nInserted" : 1 })

（4）查看日志记录所有操作，在默认数据库 local 中的oplog.rs  查看

abc:PRIMARY> show dbs
admin  0.000GB
config 0.000GB
local  0.000GB
school 0.000GB
abc:PRIMARY> use local                 #进入 local 数据库
switched to db local
abc:PRIMARY> show collections   #显示
me
oplog.rs
replset.election
replset.minvalid
startup_log
system.replset
system.rollback.id
abc:PRIMARY> db.oplog.rs.find()    #查看日志记录所有操作

通过日志记录，可以找到刚才加入的两条信息

"school.info", "ui" : UUID("2b9c93b6-a58a-4021-a2b4-33f9b19925d8"), "wall" : ISODate("2018-09-13T06:24:44.537Z"), "o" : { "_id" : ObjectId("5b9a02aca106c3eab9c639e5"), "id" : 1, "name" : "lili" } }
{ "ts" : Timestamp(1536819899, 1), "t" : NumberLong(1), "h" : NumberLong("-1447313909384631008"), "v" : 2, "op" : "i", "ns" : "school.info", "ui" : UUID("2b9c93b6-a58a-4021-a2b4-33f9b19925d8"), "wall" : ISODate("2018-09-13T06:24:59.186Z"), "o" : { "_id" : ObjectId("5b9a02bba106c3eab9c639e6"), "id" : 2, "name" : "tom" } }

3.模拟节点故障，如果主节点出现故障，另一个标准节点会选举成为新的主节点。

（1）先查看当前状态，使用命令 rs.status()

abc:PRIMARY> rs.status()

"members" : [
        {
            "_id" : 0,                  
            "name" : "192.168.213.184:27017",              #标准节点，端口27017，作为主节点
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 2481,
            "optime" : {
                "ts" : Timestamp(1536820334, 1),
                "t" : NumberLong(1)
            },

{
            "_id" : 1,
            "name" : "192.168.213.184:27018",                #标准节点 ，端口27018 为从节点
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 1213,
            "optime" : {
                "ts" : Timestamp(1536820324, 1),
                "t" : NumberLong(1)
            },

{
            "_id" : 2,
            "name" : "192.168.213.184:27019",              #被动节点， 端口27019位从节点
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 1213,
            "optime" : {
                "ts" : Timestamp(1536820324, 1),
                "t" : NumberLong(1)
            },

{
            "_id" : 3,
            "name" : "192.168.213.184:27020",                   #仲裁节点
            "health" : 1,
            "state" : 7,
            "stateStr" : "ARBITER",
            "uptime" : 1213,
            "lastHeartbeat" : ISODate("2018-09-13T06:32:14.152Z"),

（2）模拟主节点故障，关闭主节点服务，登录另一个标准节点端口 27018 ,查看麟能够一个标准节点是否被选举为主节点

abc:PRIMARY> exit
bye
[root@localhost logs]# mongod -f /etc/mongod.conf –shutdown                      #关闭主节点服务
killing process with pid: 4821
[root@localhost logs]# mongo --port 27018                                                         #指定27018 端口进入数据库

abc:PRIMARY> rs.status()                                                             #查看状态

"members" : [
        {
            "_id" : 0,
            "name" : "192.168.213.184:27017",
            "health" : 0,                                                #健康之为 0 ，说明端口27017 已经宕机了
            "state" : 8,
            "stateStr" : "(not reachable/healthy)",
            "uptime" : 0,
            "optime" : {
                "ts" : Timestamp(0, 0),
                "t" : NumberLong(-1)
            },

{
            "_id" : 1,
            "name" : "192.168.213.184:27018",              #此时另一台标准节点被选举为主节点，端口为 27018
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",

            "uptime" : 2812,
            "optime" : {
                "ts" : Timestamp(1536820668, 1),
                "t" : NumberLong(2)
            },

{
            "_id" : 2,
            "name" : "192.168.213.184:27019",
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 1552,
            "optime" : {
                "ts" : Timestamp(1536820668, 1),
                "t" : NumberLong(2)
            },

{
        "_id" : 3,
        "name" : "192.168.213.184:27020",
        "health" : 1,
        "state" : 7,
        "stateStr" : "ARBITER",
        "uptime" : 1552,

其中， health 为 1 代表健康，0 代表宕机。 state 为 1 代表主节点，为 2 代表从节点

（3）将标准节点服务全部关闭，查看被动节点是否会被选举为主节点

abc:PRIMARY> exit
bye
[root@localhost logs]# mongod -f /etc/mongod2.conf –shutdown                    #关闭主节点服务
killing process with pid: 4853
[root@localhost logs]# mongo --port 27019                                                               #开启备用节点服务

abc:SECONDARY> rs.status()

"members" : [
        {
            "_id" : 0,
            "name" : "192.168.213.184:27017",                                  #此时两个标准节点都处于宕机状态，健康值为 0
            "health" : 0,
            "state" : 8,
            "stateStr" : "(not reachable/healthy)",
            "uptime" : 0,
            "optime" : {
                "ts" : Timestamp(0, 0),
                "t" : NumberLong(-1)
            },

{
            "_id" : 1,
            "name" : "192.168.213.184:27018",
            "health" : 0,
            "state" : 8,
            "stateStr" : "(not reachable/healthy)",
            "uptime" : 0,
            "optime" : {
                "ts" : Timestamp(0, 0),
                "t" : NumberLong(-1)
            },

{
            "_id" : 2,
            "name" : "192.168.213.184:27019",                           #被动节点并没有被选举为主节点，说明被动节点不可能成为活跃节点
            "health" : 1,
            "state" : 2,
            "stateStr" : "SECONDARY",
            "uptime" : 3102,
            "optime" : {
                "ts" : Timestamp(1536820928, 1),
                "t" : NumberLong(2)
            },

{
            "_id" : 3,
            "name" : "192.168.213.184:27020",
            "health" : 1,
            "state" : 7,
            "stateStr" : "ARBITER",
            "uptime" : 1849,

（4）再次启动标准节点服务，查看主节点能否恢复。

abc:SECONDARY> exit
bye
[root@localhost logs]# mongod -f /etc/mongod.conf                                    #启动实例1 的服务
about to fork child process, waiting until server is ready for connections.
forked process: 39839
child process started successfully, parent exiting
[root@localhost logs]# mongod -f /etc/mongod2.conf                                  #启动实例2的服务
about to fork child process, waiting until server is ready for connections.
forked process: 39929
child process started successfully, parent exiting
[root@localhost logs]# mongo

{
            "_id" : 0,
            "name" : "192.168.213.184:27017",                          #端口27017 恢复为主节点，（与标准节点启动顺序有关，先启动的为主节点）
            "health" : 1,
            "state" : 1,
            "stateStr" : "PRIMARY",
            "uptime" : 25,
            "optime" : {
                "ts" : Timestamp(1536821324, 1),
                "t" : NumberLong(3)
            },

{
        "_id" : 1,
        "name" : "192.168.213.184:27018",
        "health" : 1,
        "state" : 2,
        "stateStr" : "SECONDARY",
        "uptime" : 14,
        "optime" : {
            "ts" : Timestamp(1536821324, 1),
            "t" : NumberLong(3)
        },

{
        "_id" : 2,
        "name" : "192.168.213.184:27019",
        "health" : 1,
        "state" : 2,
        "stateStr" : "SECONDARY",
        "uptime" : 23,
        "optime" : {

{
        "_id" : 3,
        "name" : "192.168.213.184:27020",
        "health" : 1,
        "state" : 7,
        "stateStr" : "ARBITER",
        "uptime" : 23,

由此可以看出：只有标准节点可能被选举为活跃节点（主 primary），被动节点不可能成为活跃节点，有选举权。仲裁节点不可能成为活跃节点。

另外有需要云服务器可以了解下创新互联cdcxhl.cn，海内外云服务器15元起步，三天无理由+7*72小时售后在线，公司持有idc许可证，提供“云服务器、裸金属服务器、高防服务器、香港服务器、美国服务器、虚拟主机、免备案服务器”等云主机租用服务以及企业上云的综合解决方案，具有“安全稳定、简单易用、服务可用性高、性价比高”等特点与优势，专为企业上云打造定制，能够满足用户丰富、多元化的应用场景需求。