您的位置：首页 > 数据库 > Mongodb

mongoDB集群之分片集群

2020-08-18 17:12 1646 查看

ps auxf|grep mongo |grep -v grep|xargs kill -9 杀死所有redis的进程

分片是把大型数据集进行分区成更小的可管理的片,这些数据片分散到不同的mongoDB节点, 这些节点组成了分片集群。总体分为3块分片服务器配置服务器路由器(mongos)组成：

如果有旧的数据先清除:

rm -rf /soft/mongosplit/node27021/db/*
rm -rf /soft/mongosplit/node27031/db/*
rm -rf /soft/mongosplit/node27041/db/*

rm -rf /soft/mongosplit/node27022/db/*
rm -rf /soft/mongosplit/node27032/db/*
rm -rf /soft/mongosplit/node27042/db/*

rm -rf /soft/mongosplit/node27023/db/*
rm -rf /soft/mongosplit/node27033/db/*
rm -rf /soft/mongosplit/node27043/db/*

rm -rf /soft/mongosplit/node27025/db/*
rm -rf /soft/mongosplit/node27035/db/*
rm -rf /soft/mongosplit/node27045/db/*

1:分片服务器配置全部3个分片（由于空间不足，分片放一台,配置和路由放另一台） 还是要在mongosplit目录下复制13个节点：先配置好一个节点再复制。在第一个节点上创建db,logs两个目录: mkdir node27021/{db,logs} 1》分片1：每个分片3台作为一个可复制集

首先: vim mgdb.conf 内容如下：

replication:
replSetName: configRS
oplogSizeMB: 50

storage:
dbPath: "/usr/local/soft/mongosplit/note27021/db"
systemLog:
destination: file
path: "/usr/local/soft/mongosplit/note27021/logs/mongodb.log"
net:
port: 27021
bindIp: 127.0.0.1,192.168.42.100
processManagement:
fork: true
setParameter:
enableLocalhostAuthBypass: false
sharding:
clusterRole: shardsvr

同一片内只需要修改IP 和端口号和文件路径

不同片内还要改这个：replSetName: configRS2 或者configRS3 ,如下：

replication:
replSetName: configRS
oplogSizeMB: 50

storage:
dbPath: "/usr/local/soft/mongosplit/note27022/db"
systemLog:
destination: file
path: "/usr/local/soft/mongosplit/note27022/logs/mongodb.log"
net:
port: 27022
bindIp: 127.0.0.1,192.168.42.100
processManagement:
fork: true
setParameter:
enableLocalhostAuthBypass: false
sharding:
clusterRole: shardsvr

启动：

/usr/local/soft/mongosplit/note27021/bin/mongod --config /usr/local/soft/mongosplit/note27021/mgdb.conf &
/usr/local/soft/mongosplit/note27031/bin/mongod --config /usr/local/soft/mongosplit/note27031/mgdb.conf &
/usr/local/soft/mongosplit/note27041/bin/mongod --config /usr/local/soft/mongosplit/note27041/mgdb.conf &

配置复制集

./mongo --port 27021
/usr/local/soft/mongosplit/note27021/bin/mongo --port 27021

use admin

rs.initiate({
    _id: "configRS",
    version: 1,
    members: [{
        _id: 0,
        host: "192.168.42.100:27021"
    }]
});

rs.add("192.168.42.100:27031"); //有几个节点就执行几次方法
rs.add("192.168.42.100:27041");

之后第2片如下：

replication:
replSetName: configRS2
oplogSizeMB: 50

storage:
dbPath: "/usr/local/soft/mongosplit/note27042/db"
systemLog:
destination: file
path: "/usr/local/soft/mongosplit/note27042/logs/mongodb.log"
net:
port: 27042
bindIp: 127.0.0.1,192.168.42.100
processManagement:
fork: true
setParameter:
enableLocalhostAuthBypass: false
sharding:
clusterRole: shardsvr

启动：

/usr/local/soft/mongosplit/note27022/bin/mongod --config /usr/local/soft/mongosplit/note27022/mgdb.conf &
/usr/local/soft/mongosplit/note27032/bin/mongod --config /usr/local/soft/mongosplit/note27032/mgdb.conf &
/usr/local/soft/mongosplit/note27042/bin/mongod --config /usr/local/soft/mongosplit/note27042/mgdb.conf &

配置复制集

./mongo --port 27022
/usr/local/soft/mongosplit/note27032/bin/mongo --port 27032

use admin

rs.initiate({
    _id: "configRS2",
    version: 1,
    members: [{
        _id: 0,
        host: "192.168.42.100:27022"
    }]
});

rs.add("192.168.42.100:27032"); //有几个节点就执行几次方法
rs.add("192.168.42.100:27042");

之后第3片如下：

replication:
replSetName: configRS3
oplogSizeMB: 50

storage:
dbPath: "/usr/local/soft/mongosplit/note27043/db"
systemLog:
destination: file
path: "/usr/local/soft/mongosplit/note27043/logs/mongodb.log"
net:
port: 27043
bindIp: 127.0.0.1,192.168.42.100
processManagement:
fork: true
setParameter:
enableLocalhostAuthBypass: false
sharding:
clusterRole: shardsvr

启动：

/usr/local/soft/mongosplit/note27023/bin/mongod --config /usr/local/soft/mongosplit/note27023/mgdb.conf &
/usr/local/soft/mongosplit/note27033/bin/mongod --config /usr/local/soft/mongosplit/note27033/mgdb.conf &
/usr/local/soft/mongosplit/note27043/bin/mongod --config /usr/local/soft/mongosplit/note27043/mgdb.conf &

配置复制集

./mongo --port 27023
./usr/local/soft/mongosplit/note27023/bin/mongo --port 27023

use admin

rs.initiate({
    _id: "configRS3",
    version: 1,
    members: [{
        _id: 0,
        host: "192.168.42.100:27023"
    }]
});

rs.add("192.168.42.100:27033"); //有几个节点就执行几次方法
rs.add("192.168.42.100:27043");

以上完成分片的：

配置 Config 服务器（3台） 节点下 vim mgdb.conf配置如下：

replication:
replSetName: editRS
oplogSizeMB: 50

storage:
dbPath: "/usr/local/soft/mongosplit/note27045/db"
systemLog:
destination: file
path: "/usr/local/soft/mongosplit/note27045/logs/mongodb.log"
net:
port: 27045
bindIp: 127.0.0.1,192.168.42.101
processManagement:
fork: true
setParameter:
enableLocalhostAuthBypass: false
sharding:
clusterRole: configsvr

注意： replSetName: editRS 名称不一样了 也是一样要启动： /usr/local/soft/mongosplit/note27025/bin/mongod --config /usr/local/soft/mongosplit/note27025/mgdb.conf &
/usr/local/soft/mongosplit/note27035/bin/mongod --config /usr/local/soft/mongosplit/note27035/mgdb.conf &
/usr/local/soft/mongosplit/note27045/bin/mongod --config /usr/local/soft/mongosplit/note27045/mgdb.conf &
也是一样要配置可复制集：

./mongo --port 27025
/usr/local/soft/mongosplit/note27025/bin/mongo --port 27025

use admin

rs.initiate({
    _id: "editRS",
    version: 1,
    members: [{
        _id: 0,
        host: "192.168.42.101:27025"
    }]
});

rs.add("192.168.42.101:27035"); //有几个节点就执行几次方法
rs.add("192.168.42.101:27045");

mongos 路由器一台（27066) 节点下 vim mgdb.conf配置如下： systemLog:
destination: file
path: "/usr/local/soft/mongosplit/note27066/logs/mongodb.log"
net:
port: 27055
bindIp: 127.0.0.1,192.168.42.101
processManagement:
fork: true
setParameter:
enableLocalhostAuthBypass: false
sharding:
configDB:editRS/192.168.42.101:27025,192.168.42.101:27035,192.168.42.101:27045 指定：configDB:editRS 然后启动之后添加分区： /usr/local/soft/mongosplit/note27066/bin/mongos --config /usr/local/soft/mongosplit/note27066/mgdb.conf &

use admin;

sh.addShard("configRS/192.168.42.100:27021,192.168.42.100:27031,192.168.42.100:27041");

sh.addShard("configRS2/192.168.42.100:27022,192.168.42.100:27032,192.168.42.100:27042");

//configRS 这个是复制集的名称
sh.addShard("configRS3/192.168.42.100:27023,192.168.42.100:27033,192.168.42.100:27043");

对lison数据库启动分片： sh.enableSharding("lison");

对 orders 集合进行分片，分片键为 sh.shardCollection("lison.orders",{"useCode":"hashed"});

删除分片 db.runCommand( { removeShard: "xxx" } ) 查看状态： sh.status();