Table of Contents

Introduction

A replica set provides MongoDB with redundancy and increased data availability by replicating the database to multiple servers. The primary member of the replica set can be used for read and write operations whereas the secondary members are available for read-only operations.

A replica set must consist of an odd number of MongoDB servers. If a single server fails, the replica set members must be able elect a new primary server. An election majority can also be achieved with two database servers by including a third MongoDB arbiter. The arbiter does not have to maintain a copy of the database, but it does need to be a separate server instance.

Requirements

  • Minimum of three server instances
  • MongoDB 2.6 installed and running
  • Each server instance must allow TCP traffic from each replica set member over port 27017
  • Each server instance hostname must be resolvable from each replica set member

As mentioned in the above requirements, each replica set member must be able to communicate with the other replica set members by their hostname. The hostname will default to the fully qualified domain name of each server. The server hostnames can be added to DNS, but the recommended method is to add all replica set member hostnames to their /etc/hosts file. For example:

10.10.200.1    mongodb01.yourdomain.com mongodb01
10.10.200.2 mongodb02.yourdomain.com mongodb02
10.10.200.3 mongodb03.yourdomain.com mongodb03

Create Replica Set

MongoDB must successfully be running on the primary server instance. The standalone instance of MongoDB can become a replica set member by assigning a replica set name in the MongoDB configuration file. To create a replica set called "rs0", add the following line to /etc/mongod.conf.

replSet=rs0

MongoDB will need to be restarted for the change to take affect.

sudo /etc/init.d/mongod restart

Use the mongo shell to connect once the service is running again.

mongo

Initiate the replica set from within the mongo shell.

rs.initiate()

The initial replica set configuration can be verified using rs.conf().

rs.conf()
{
"_id" : "rs0",
"version" : 1,
"members" : [
{
"_id" : 0,
"host" : "mongodb01.yourdomain.com:27017"
}
]
}

At this stage, the mongo shell prompt should also indicate the replica name and replica member role. For example:

rs0:PRIMARY>

The replica set is now created but more members must be added to achieve redundancy and increased data availability.

Add Secondary Members

The secondary MongoDB server instances must be successfully running MongoDB and accessible by the other members. The configuration file on the secondary members must reflect the same replica set name as the primary member. Add the following line to the /etc/mongod.conf on all secondary members.

replSet=rs0

Restart the MongoDB service.

sudo /etc/init.d/mongod restart

The secondary members can now be added from the mongo shell on the primary server using rs.add().

rs0:PRIMARY> rs.add("mongodb02.yourdomain.com")
rs0:PRIMARY> rs.add("mongodb03.yourdomain.com")

The rs.status() command can be used to verify the status of the replica set:

rs.status()

Add an Arbiter

An arbiter is an optional member of a replica set which stores no data. Instead, it is available to break election ties in a situation where there is an even number of MongoDB members.

An arbiter member does not require the same performance capabilities of a full MongoDB server, but it must reside on a separate server instance. Never configure an arbiter on an existing replica set member.

To add an arbiter, first install MongoDB on a server instance accessible by the other replica set members and add the replica set name to /etc/mongod.conf.

replSet=rs0

Unnecessary data can be reduced on the arbiter server by adding the following files to /etc/mongod.conf.

journal.enabled=false
smallFiles=true
preallocDataFiles=false

Now log into the primary replica set member and launch the mongo shell.

mongo

The arbiter can be added with the following command:

rs.addArb("arbiter01.yourdomain.com:27017")

原文地址:https://devops.profitbricks.com/tutorials/configure-mongodb-replica-set/#create-replica-set

Configure MongoDB Replica Set的更多相关文章

  1. 关于MongoDb Replica Set的故障转移集群——实战篇

    如果你还不了解Replica Set的相关理论,请猛戳传送门阅读笔者的上一篇博文. 因为Replica Set已经属于MongoDb的进阶应用,下文中关于MongoDb的基础知识笔者就不再赘述了,请参 ...

  2. MongoDb Replica Set中使用的地址

    Unable to connect to a member of the replica set matching the read preference Primary 今天尝试使用MongoDB ...

  3. MongoDB 学习笔记(三) MongoDB (replica set) 集群配置

    MongoDB Replica Sets的结构类似于以集群,完全可以把他当成一个集群,因为他确实与集群实现的作用是一样的:如果其中一个节点出现故障,其他的节点会马上将业务接管过来.而无需停机操作 Mo ...

  4. mongodb replica set 和 nodejs中使用mongoose连接replica

    一.mongodb replication 介绍 官网上的第一句话就是Replication is the process of synchronizing data across multiple ...

  5. MongoDB Replica Set搭建集群

    MongoDB做集群,版本3.2官网推荐的集群方式Replica Set 准备服务器3台 两个standard节点(这两个节点直接可以互切primary secondary). 一个arbiter节点 ...

  6. mongodb replica set 配置高性能多服务器详解

    mongodb的多服务器配置,以前写过一篇文章,是master-slave模式的,请参考:详解mongodb 主从配置.master-slave模式,不能自动实现故障转移和恢复.所以推荐大家使用mon ...

  7. (1)解锁 MongoDB replica set核心姿势

    副本集Replica Set是一个术语,定义具有多节点的数据库集群,这些节点具有主从复制(master-slave replication) 且节点之间实现了自动故障转移. 这样的结构通常需要具有奇数 ...

  8. MongoDB replica set IDs do not match

    在搭建MongoDB(版本 3.2.9)的Replica Set时,使用 rs.status() 查看Replica Set的状态,发现一个成员异常:replica set IDs do not ma ...

  9. MongoDB Replica Set 选举过程

    什么是选举? 选举是副本集选择某个成员成为primary的过程.primary是一个副本集中唯一能够接收写操作的成员. 下面的事件能够引发一次选举: 第一次初始化一个副本集 Primary失效.rep ...

随机推荐

  1. Oracle11gr2_ADG管理之跳归档恢复dg实战

    模拟故障 关闭备库 SQL> shutdown immediate; Database closed. Database dismounted. ORACLE instance shut dow ...

  2. LINUX的SSH下FTP到远程服务器Entering Passive Mode失败解决

    LINUX 系统FTP连接远程服务器经常出现在传输文件或者发出 ls命令时候出现 "Entering Passive Mode "然后就再也无法运作了.该工作主要是因为LINUX的 ...

  3. Stall Reservations(贪心+优先队列)

    Description Oh those picky N (1 <= N <= 50,000) cows! They are so picky that each one will onl ...

  4. 使用avalon实现SKU组合查询功能

    SKU(stock keeping unit库存量单位)组合查询是网上商场一个非常常用的功能.具体来说,一件商品会有许多型号,许多颜色,许多产地,许多码寸,而满足用户选中的这些条件的具体商品可能有库存 ...

  5. Python之面向过程和面向对象的区别

    一.面向过程 1.面向过程:核心是过程二字,过程指的是解决问题的步骤,好比如设计一条流水线,是一种机械式的思维方式. 就是程序从上到下一步步执行,一步步从上到下,从头到尾的解决问题 .基本设计思路就是 ...

  6. 安装 Windows Service

    1.打开 VS 命令行窗口 2. installutil /u  service文件路径  (卸载原有服务) 3, installutil /i  service 文件路径 (安装服务)

  7. mfs教程(三)

    mfs文件系统(三) 使用  MooseFS 一.挂载文件系统 启动管理服务器(master server)和数据服务器(chunkservers) (chunkservers一个是必需的,但至少两个 ...

  8. opennebula kvm 创建虚拟机错误

    Thu Jul :: : Error executing image transfer script: Error copying localhost.localdomain:/app/openneb ...

  9. python字符串大小写转换

    str = "www.w3cSChool.cn"print(str.upper()) # 把所有字符中的小写字母转换成大写字母print(str.lower()) # 把所有字符中 ...

  10. .NET读取服务器或本地文件

    //把本地文件信息读入数据流中                    FileStream stream = new FileStream(path, FileMode.Open, FileAcces ...