mongodb是最常用的noSql数据库,在数据库排名中已经上升到了前五。这篇文章介绍如何搭建高可用的mongodb(分片+副本)集群。
在搭建集群之前,需要首先了解几个概念:路由,分片、副本集、配置服务器等。
相关概念
mongodb集群架构图:
从图中可以看到有四个组件:mongos、config server、shard、replica set。
mongos,数据库集群请求的入口,所有的请求都通过mongos进行协调,不需要在应用程序添加一个路由选择器,mongos自己就是一个请求分发中心,它负责把对应的数据请求请求转发到对应的shard服务器上。在生产环境通常有多mongos作为请求的入口,防止其中一个挂掉所有的mongodb请求都没有办法操作。
config server,顾名思义为配置服务器,存储所有数据库元信息(路由、分片)的配置。mongos本身没有物理存储分片服务器和数据路由信息,只是缓存在内存里,配置服务器则实际存储这些数据。mongos第一次启动或者关掉重启就会从 config server 加载配置信息,以后如果配置服务器信息变化会通知到所有的 mongos 更新自己的状态,这样 mongos 就能继续准确路由。在生产环境通常有多个 config server 配置服务器,因为它存储了分片路由的元数据,防止数据丢失!
shard,分片(sharding)是指将数据库拆分,将其分散在不同的机器上的过程。将数据分散到不同的机器上,不需要功能强大的服务器就可以存储更多的数据和处理更大的负载。基本思想就是将集合切成小块,这些块分散到若干片里,每个片只负责总数据的一部分,最后通过一个均衡器来对各个分片进行均衡(数据迁移)。
replica set,中文翻译副本集,其实就是shard的备份,防止shard挂掉之后数据丢失。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。
仲裁者(Arbiter),是复制集中的一个MongoDB实例,它并不保存数据。仲裁节点使用最小的资源并且不要求硬件设备,不能将Arbiter部署在同一个数据集节点中,可以部署在其他应用服务器或者监视服务器中,也可部署在单独的虚拟机中。为了确保复制集中有奇数的投票成员(包括primary),需要添加仲裁节点做为投票,否则primary不能运行时不会自动切换primary。
简单了解之后,我们可以这样总结一下,应用请求mongos来操作mongodb的增删改查,配置服务器存储数据库元信息,并且和mongos做同步,数据最终存入在shard(分片)上,为了防止数据丢失同步在副本集中存储了一份,仲裁在数据存储到分片的时候决定存储到哪个节点。
环境准备
系统系统 centos 7.4
三台服务器:192.168.56.101、node2(192.168.56.102)、node3(192.168.56.103)
安装包:yum安装
服务器规划
服务器node1(192.168.56.101) |
服务器node2(192.168.56.102) |
服务器node3(192.168.56.103) |
mongos |
mongos |
mongos |
config server |
config server |
config server |
shard server1 主节点 |
shard server1 副节点 |
shard server1 仲裁 |
shard server2 仲裁 |
shard server2 主节点 |
shard server2 副节点 |
shard server3 副节点 |
shard server3 仲裁 |
shard server3 主节点 |
端口分配:
mongos:20000
config:21000
shard1:27001
shard2:27002
shard3:27003
开放端口
firewall-cmd --add-port=21000/tcp --permanent
firewall-cmd --add-port=21000/tcp
firewall-cmd --add-port=27001/tcp --permanent
firewall-cmd --add-port=27001/tcp
firewall-cmd --add-port=27002/tcp --permanent
firewall-cmd --add-port=27002/tcp
firewall-cmd --add-port=27003/tcp --permanent
firewall-cmd --add-port=27003/tcp
firewall-cmd --add-port=20000/tcp --permanent
firewall-cmd --add-port=20000/tcp
关闭selinux
vi /etc/selinux/config
SELINUX=disabled
重启系统
集群搭建
1、安装mongodb
#yum安装mongodb
cat /etc/yum.repos.d/mongodb.repo
[mongodb-org]
name=MongoDB Repository
baseurl=https://repo.mongodb.org/yum/redhat/$releasever/mongodb-org/3.6/x86_64/
gpgcheck=1
enabled=1
gpgkey=https://www.mongodb.org/static/pgp/server-3.6.asc
yum install mongodb-org
禁用自带服务:
systemctl disable mongod
systemctl stop mongod
路径规划并创建
分别在每台机器建立conf、mongos、config、shard1、shard2、shard3目录,因为mongos不存储数据,只需要建立日志文件目录即可。
配置文件路径:
mkdir -p /etc/mongod/conf.d
pid文件路径
/var/run/mongodb
数据存储路径
#config server数据存储路径
mkdir -p /var/lib/mongo/config/data
#shard server数据存储路径
mkdir -p /var/lib/mongo/shard1/data
mkdir -p /var/lib/mongo/shard2/data
mkdir -p /var/lib/mongo/shard3/data
chown -R mongod:mongod /var/lib/mongo
日志文件路径
/var/log/mongodb
2、config server配置服务器
mongodb3.4以后要求配置服务器也创建副本集,不然集群搭建不成功。
添加配置文件
vi /etc/mongod/conf.d/config.conf
## 配置文件内容
systemLog:
destination: file
logAppend: true
path: /var/log/mongodb/configsvr.log
storage:
dbPath: /var/lib/mongo/config/data
journal:
enabled: true
processManagement:
fork: true
pidFilePath: /var/run/mongodb/configsvr.pid
timeZoneInfo: /usr/share/zoneinfo
net:
port: 21000
bindIp: 0.0.0.0
maxIncomingConnections: 20000
replication:
replSetName: csReplSet
sharding:
clusterRole: configsvr
创建systemctl unit文件mongod-configsvr.service
[Unit]
Description=Mongodb Config Server
After=network.target
Documentation=https://docs.mongodb.org/manual
[Service]
User=mongod
Group=mongod
Environment="OPTIONS=-f /etc/mongod/conf.d/config.conf"
ExecStart=/usr/bin/mongod $OPTIONS
ExecStartPre=/usr/bin/mkdir -p /var/run/mongodb
ExecStartPre=/usr/bin/chown mongod:mongod /var/run/mongodb
ExecStartPre=/usr/bin/chmod 0755 /var/run/mongodb
PermissionsStartOnly=true
PIDFile=/var/run/mongodb/configsvr.pid
Type=forking
# file size
LimitFSIZE=infinity
# cpu time
LimitCPU=infinity
# virtual memory size
LimitAS=infinity
# open files
LimitNOFILE=64000
# processes/threads
LimitNPROC=64000
# locked memory
LimitMEMLOCK=infinity
# total threads (user+kernel)
TasksMax=infinity
TasksAccounting=false
# Recommended limits for for mongod as specified in
# http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
[Install]
WantedBy=multi-user.target
启动三台服务器的config server
systemctl enable mongod-configsvr
systemctl start mongod-configsvr
登录任意一台配置服务器,初始化配置副本集
#连接
mongo 127.0.0.1:21000
#config变量
config = {
_id : "csReplSet",
members : [
{_id : 1, host : "192.168.56.101:21000" },
{_id : 2, host : "192.168.56.102:21000" },
{_id : 3, host : "192.168.56.103:21000" }
]
}
#初始化副本集
rs.initiate(config)
其中,"_id" : " csReplSet "应与配置文件中配置的 replicaction.replSetName 一致,"members" 中的 "host" 为三个节点的 ip 和 port
3、配置分片副本集(三台机器)
1) 设置第一个分片副本集
添加配置文件
vi /etc/mongod/conf.d/shard1.conf
## 配置文件内容
systemLog:
destination: file
logAppend: true
path: /var/log/mongodb/shard1.log
storage:
dbPath: /var/lib/mongo/shard1/data
journal:
enabled: true
processManagement:
fork: true
pidFilePath: /var/run/mongodb/shard1.pid
timeZoneInfo: /usr/share/zoneinfo
net:
port: 27001
bindIp: 0.0.0.0
maxIncomingConnections: 20000
replication:
replSetName: shard1
sharding:
clusterRole: shardsvr
创建systemctl unit文件mongod-shard1.service
[Unit]
Description=Mongodb Shard1 Server
After=network.target
Documentation=https://docs.mongodb.org/manual
[Service]
User=mongod
Group=mongod
Environment="OPTIONS=-f /etc/mongod/conf.d/shard1.conf"
ExecStart=/usr/bin/mongod $OPTIONS
ExecStartPre=/usr/bin/mkdir -p /var/run/mongodb
ExecStartPre=/usr/bin/chown mongod:mongod /var/run/mongodb
ExecStartPre=/usr/bin/chmod 0755 /var/run/mongodb
PermissionsStartOnly=true
PIDFile=/var/run/mongodb/shard1.pid
Type=forking
# file size
LimitFSIZE=infinity
# cpu time
LimitCPU=infinity
# virtual memory size
LimitAS=infinity
# open files
LimitNOFILE=64000
# processes/threads
LimitNPROC=64000
# locked memory
LimitMEMLOCK=infinity
# total threads (user+kernel)
TasksMax=infinity
TasksAccounting=false
# Recommended limits for for mongod as specified in
# http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
[Install]
WantedBy=multi-user.target
启动三台服务器的shard1 server
systemctl enable mongod-shard1
systemctl start mongod-shard1
登陆任意一台非仲裁节点服务器,初始化副本集
mongo 127.0.0.1:27001
#使用admin数据库
use admin
#定义副本集配置,第三个节点的 "arbiterOnly":true 代表其为仲裁节点。
config = {
_id : "shard1",
members : [
{_id : 1, host : "192.168.56.101:27001" , priority:2},
{_id : 2, host : "192.168.56.102:27001" , priority:1},
{_id : 3, host : "192.168.56.103:27001" , arbiterOnly: true }
]
}
#初始化副本集配置
rs.initiate(config);
2) 设置第二个分片副本集
添加配置文件
vi /etc/mongod/conf.d/shard2.conf
## 配置文件内容
systemLog:
destination: file
logAppend: true
path: /var/log/mongodb/shard2.log
storage:
dbPath: /var/lib/mongo/shard2/data
journal:
enabled: true
processManagement:
fork: true
pidFilePath: /var/run/mongodb/shard2.pid
timeZoneInfo: /usr/share/zoneinfo
net:
port: 27002
bindIp: 0.0.0.0
maxIncomingConnections: 20000
replication:
replSetName: shard2
sharding:
clusterRole: shardsvr
创建systemctl unit文件mongod-shard2.service
[Unit]
Description=Mongodb Shard2 Server
After=network.target
Documentation=https://docs.mongodb.org/manual
[Service]
User=mongod
Group=mongod
Environment="OPTIONS=-f /etc/mongod/conf.d/shard2.conf"
ExecStart=/usr/bin/mongod $OPTIONS
ExecStartPre=/usr/bin/mkdir -p /var/run/mongodb
ExecStartPre=/usr/bin/chown mongod:mongod /var/run/mongodb
ExecStartPre=/usr/bin/chmod 0755 /var/run/mongodb
PermissionsStartOnly=true
PIDFile=/var/run/mongodb/shard2.pid
Type=forking
# file size
LimitFSIZE=infinity
# cpu time
LimitCPU=infinity
# virtual memory size
LimitAS=infinity
# open files
LimitNOFILE=64000
# processes/threads
LimitNPROC=64000
# locked memory
LimitMEMLOCK=infinity
# total threads (user+kernel)
TasksMax=infinity
TasksAccounting=false
# Recommended limits for for mongod as specified in
# http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
[Install]
WantedBy=multi-user.target
启动三台服务器的shard2 server
systemctl enable mongod-shard2
systemctl start mongod-shard2
登陆任意一台非仲裁节点服务器,初始化副本集
mongo 127.0.0.1:27002
#使用admin数据库
use admin
#定义副本集配置,第一个节点的 "arbiterOnly":true 代表其为仲裁节点。
config = {
_id : "shard2",
members : [
{_id : 1, host : "192.168.56.101:27002" , arbiterOnly: true },
{_id : 2, host : "192.168.56.102:27002" , priority:2},
{_id : 3, host : "192.168.56.103:27002" , priority:1}
]
}
#初始化副本集配置
rs.initiate(config);
3) 设置第三个分片副本集
添加配置文件
vi /etc/mongod/conf.d/shard3.conf
## 配置文件内容
systemLog:
destination: file
logAppend: true
path: /var/log/mongodb/shard3.log
storage:
dbPath: /var/lib/mongo/shard3/data
journal:
enabled: true
processManagement:
fork: true
pidFilePath: /var/run/mongodb/shard3.pid
timeZoneInfo: /usr/share/zoneinfo
net:
port: 27003
bindIp: 0.0.0.0
maxIncomingConnections: 20000
replication:
replSetName: shard3
sharding:
clusterRole: shardsvr
创建systemctl unit文件mongod-shard3.service
[Unit]
Description=Mongodb Shard3 Server
After=network.target
Documentation=https://docs.mongodb.org/manual
[Service]
User=mongod
Group=mongod
Environment="OPTIONS=-f /etc/mongod/conf.d/shard3.conf"
ExecStart=/usr/bin/mongod $OPTIONS
ExecStartPre=/usr/bin/mkdir -p /var/run/mongodb
ExecStartPre=/usr/bin/chown mongod:mongod /var/run/mongodb
ExecStartPre=/usr/bin/chmod 0755 /var/run/mongodb
PermissionsStartOnly=true
PIDFile=/var/run/mongodb/shard3.pid
Type=forking
# file size
LimitFSIZE=infinity
# cpu time
LimitCPU=infinity
# virtual memory size
LimitAS=infinity
# open files
LimitNOFILE=64000
# processes/threads
LimitNPROC=64000
# locked memory
LimitMEMLOCK=infinity
# total threads (user+kernel)
TasksMax=infinity
TasksAccounting=false
# Recommended limits for for mongod as specified in
# http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
[Install]
WantedBy=multi-user.target
启动三台服务器的shard3 server
systemctl enable mongod-shard3
systemctl start mongod-shard3
登陆任意一台非仲裁节点服务器,初始化副本集
mongo 127.0.0.1:27003
#使用admin数据库
use admin
#定义副本集配置,第二个节点的 "arbiterOnly":true 代表其为仲裁节点。
config = {
_id : "shard3",
members : [
{_id : 1, host : "192.168.56.101:27003" , priority:1},
{_id : 2, host : "192.168.56.102:27003" , arbiterOnly: true},
{_id : 3, host : "192.168.56.103:27003" , priority:2}
]
}
#初始化副本集配置
rs.initiate(config);
4、配置路由服务器 mongos
先启动配置服务器和分片服务器,后启动路由实例启动路由实例:(三台机器)
vi /etc/mongod/conf.d/mongos.conf
## 配置文件内容
systemLog:
destination: file
logAppend: true
path: /var/log/mongodb/mongos.log
processManagement:
fork: true
pidFilePath: /var/run/mongodb/mongos.pid
timeZoneInfo: /usr/share/zoneinfo
net:
port: 20000
bindIp: 0.0.0.0
maxIncomingConnections: 20000
sharding:
configDB: csReplSet/192.168.56.101:21000, 192.168.56.102:21000, 192.168.56.103:21000
#注意监听的配置服务器,只能有1个或者3个 csReplSet为配置服务器的副本集名字
configDB: csReplSet/192.168.56.101:21000, 192.168.56.102:21000, 192.168.56.103:21000
创建systemctl unit文件mongod-mongos.service
[Unit]
Description=Mongodb Mongos Server
After=network.target mongod-shard1.service mongod-shard2.service mongod-shard3.service
Documentation=https://docs.mongodb.org/manual
[Service]
User=mongod
Group=mongod
Environment="OPTIONS=-f /etc/mongod/conf.d/mongos.conf"
ExecStart=/usr/bin/mongos $OPTIONS
ExecStartPre=/usr/bin/mkdir -p /var/run/mongodb
ExecStartPre=/usr/bin/chown mongod:mongod /var/run/mongodb
ExecStartPre=/usr/bin/chmod 0755 /var/run/mongodb
PermissionsStartOnly=true
PIDFile=/var/run/mongodb/mongos.pid
Type=forking
# file size
LimitFSIZE=infinity
# cpu time
LimitCPU=infinity
# virtual memory size
LimitAS=infinity
# open files
LimitNOFILE=64000
# processes/threads
LimitNPROC=64000
# locked memory
LimitMEMLOCK=infinity
# total threads (user+kernel)
TasksMax=infinity
TasksAccounting=false
# Recommended limits for for mongod as specified in
# http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
[Install]
WantedBy=multi-user.target
启动三台服务器的mongos server
systemctl enable mongod-mongos
systemctl start mongod-mongos
5、启用分片
目前搭建了mongodb配置服务器、路由服务器,各个分片服务器,不过应用程序连接到mongos路由服务器并不能使用分片机制,还需要在程序里设置分片配置,让分片生效。
登陆任意一台mongos
mongo 127.0.0.1:20000
#使用admin数据库
user admin
#串联路由服务器与分配副本集
sh.addShard("shard1/192.168.56.101:27001,192.168.56.102:27001,192.168.56.103:27001")
sh.addShard("shard2/192.168.56.101:27002,192.168.56.102:27002,192.168.56.103:27002")
sh.addShard("shard3/192.168.56.101:27003,192.168.56.102:27003,192.168.56.103:27003")
#查看集群状态
sh.status()
mongos> sh.status()
--- Sharding Status ---
sharding version: {
"_id" : 1,
"minCompatibleVersion" : 5,
"currentVersion" : 6,
"clusterId" : ObjectId("5a86b6255d128f35cb22de20")
}
shards:
{ "_id" : "shard1", "host" : "shard1/192.168.56.101:27001,192.168.56.102:27001", "state" : 1 }
{ "_id" : "shard2", "host" : "shard2/192.168.56.102:27002,192.168.56.103:27002", "state" : 1 }
{ "_id" : "shard3", "host" : "shard3/192.168.56.101:27003,192.168.56.103:27003", "state" : 1 }
active mongoses:
"3.6.2" : 3
autosplit:
Currently enabled: yes
balancer:
Currently enabled: yes
Currently running: no
Failed balancer rounds in last 5 attempts: 0
Migration Results for the last 24 hours:
No recent migrations
databases:
{ "_id" : "config", "primary" : "config", "partitioned" : true }
6、测试
目前配置服务、路由服务、分片服务、副本集服务都已经串联起来了,但我们的目的是希望插入数据,数据能够自动分片。连接在mongos上,准备让指定的数据库、指定的集合分片生效。
mongo 127.0.0.1:20000
#设置分片chunk大小
use config
db.settings.save({ "_id" : "chunksize", "value" : 1 })
设置1M是为了测试,否则要插入大量数据才能分片。
#指定test分片生效
sh.enableSharding("test")
#指定数据库里需要分片的集合和片键
use test
db.users.createIndex({user_id : 1})
use admin
sh.shardCollection("test.users", {user_id: 1})
我们设置testdb的 table1 表需要分片,根据 user_id 自动分片到 shard1 ,shard2,shard3 上面去。要这样设置是因为不是所有mongodb 的数据库和表 都需要分片!
测试分片配置结果
mongo 127.0.0.1:20000
use test;
for (var i = 1; i <=1000000; i++){
db.users.save({user_id: i, username: "user"+i});
}
#查看分片情况如下,部分无关信息省掉了
sh.status()
--- Sharding Status ---
sharding version: {
"_id" : 1,
"minCompatibleVersion" : 5,
"currentVersion" : 6,
"clusterId" : ObjectId("5a8919101e5f8edd5feddf36")
}
shards:
{ "_id" : "shard1", "host" : "shard1/192.168.56.101:27001,192.168.56.102:27001", "state" : 1 }
{ "_id" : "shard2", "host" : "shard2/192.168.56.102:27002,192.168.56.103:27002", "state" : 1 }
{ "_id" : "shard3", "host" : "shard3/192.168.56.101:27003,192.168.56.103:27003", "state" : 1 }
active mongoses:
"3.6.2" : 3
autosplit:
Currently enabled: yes
balancer:
Currently enabled: yes
Currently running: no
Collections with active migrations:
test.users started at Sun Feb 18 2018 21:14:30 GMT+0800 (CST)
Failed balancer rounds in last 5 attempts: 0
Migration Results for the last 24 hours:
43 : Success
1 : Failed with error ‘aborted‘, from shard1 to shard3
139 : Failed with error ‘aborted‘, from shard1 to shard2
databases:
…
{ "_id" : "test", "primary" : "shard3", "partitioned" : true }
test.users
shard key: { "user_id" : 1 }
unique: false
balancing: true
chunks:
shard1 34
shard2 31
shard3 31
too many chunks to print, use verbose if you want to force print
可以看到数据分到3个分片,各自分片chunk 数为: shard1 : 34,shard2 : 31,shard3 : 31。已经成功了!
后期运维
启动
mongodb的启动顺序是,先启动配置服务器,在启动分片,最后启动mongos.
systemctl start mongod-configsvr
systemctl start mongod-shard1
systemctl start mongod-shard2
systemctl start mongod-shard3
systemctl start mongod-mongos
关闭:
systemctl stop mongod-mongos
systemctl stop mongod-shard3
systemctl stop mongod-shard2
systemctl stop mongod-shard1
systemctl stop mongod-configsvr