MongoDB高阶特性：分片
二、分片
一）简介
首先我们先来了解一下Mongo集群的概念，Mongo集群有3个主要组件
- ConfigServer：在集群中扮演存储整个集群的配置信息，负责配置存储，如果需要高可用的ConfigServer那么需要3个节点。
- Shard：分片，存储真实的数据，每一个Shard分片都负责存储集群中的数据，例如一个集群有3个分片，然后我们定义分片规则为哈希，那么整个集群的数据就会（分割）到3个分片中的某一个分片，那么分片是特别重要的，如果集群中的一个分片全部崩溃了那么集群将不可用，所以我们要保证集群的高可用，那么我们需要一个分片配置3个节点，2个副本集一个仲裁节点，仲裁节点类似于Redis的哨兵模式，如果发现主节点挂了那么让另一个副本集进行数据存储。
- Mongos：Mongos我们可以理解为整个集群的入口，类似于Kafka的Broker代理，也就是客户端，我们通过客户端连接集群进行查询。
下面是MongoDB的官方集群架构图，我们看到Mongos是一个路由，他们的信息都存储在ConfigServer中，我们通过Mongos进行添加，然后根据条件将数据进行分片到分片的副本集中
二）搭建
1、docker-compose-mongodb-shard.yml
version: '3'
services:
shard1:
image: mongo:5.0.8
container_name: mongo_shard1
command: mongod --replSet shard1 --port 27011 --keyFile /data/mongodb/keyFile --directoryperdb --shardsvr
volumes:
- /etc/localtime:/etc/localtime
- ./mongodb-cluster/shard1:/data/db
- ./mongodb-cluster/keyFile:/data/mongodb/keyFile
networks:
- weiji-net
shard2:
image: mongo:5.0.8
container_name: mongo_shard2
command: mongod --replSet shard2 --port 27012 --keyFile /data/mongodb/keyFile --directoryperdb --shardsvr
volumes:
- /etc/localtime:/etc/localtime
- ./mongodb-cluster/shard2:/data/db
- ./mongodb-cluster/keyFile:/data/mongodb/keyFile
networks:
- weiji-net
shard3:
image: mongo:5.0.8
container_name: mongo_shard3
command: mongod --replSet shard3 --port 27013 --keyFile /data/mongodb/keyFile --directoryperdb --shardsvr
volumes:
- /etc/localtime:/etc/localtime
- ./mongodb-cluster/shard3:/data/db
- ./mongodb-cluster/keyFile:/data/mongodb/keyFile
networks:
- weiji-net
config1:
image: mongo:5.0.8
container_name: mongo_config1
command: mongod --replSet fates-mongo-config --port 27021 --keyFile /data/mongodb/keyFile --directoryperdb --configsvr
volumes:
- /etc/localtime:/etc/localtime
- ./mongodb-cluster/config1:/data/db
- ./mongodb-cluster/keyFile:/data/mongodb/keyFile
networks:
- weiji-net
config2:
image: mongo:5.0.8
container_name: mongo_config2
command: mongod --replSet fates-mongo-config --port 27022 --keyFile /data/mongodb/keyFile --directoryperdb --configsvr
volumes:
- /etc/localtime:/etc/localtime
- ./mongodb-cluster/config2:/data/db
- ./mongodb-cluster/keyFile:/data/mongodb/keyFile
networks:
- weiji-net
config3:
image: mongo:5.0.8
container_name: mongo_config3
command: mongod --replSet fates-mongo-config --port 27023 --keyFile /data/mongodb/keyFile --directoryperdb --configsvr
volumes:
- /etc/localtime:/etc/localtime
- ./mongodb-cluster/config3:/data/db
- ./mongodb-cluster/keyFile:/data/mongodb/keyFile
networks:
- weiji-net
mongos:
image: mongo:5.0.8
container_name: mongo_mongos
command: mongos --configdb fates-mongo-config/config1:27021,config2:27022,config3:27023 --keyFile /data/mongodb/keyFile --bind_ip 0.0.0.0 --port 27017
ports:
- 27017:27017
volumes:
- /etc/localtime:/etc/localtime
- ./mongodb-cluster/mongos:/data/db
- ./mongodb-cluster/keyFile:/data/mongodb/keyFile
depends_on:
- config1
- config2
- config3
networks:
- weiji-net
networks:
weiji-net:
external: true
name: app_net
2、准备keyFile
openssl rand -base64 128 > ./mongodb-cluster/keyFile
sudo chmod 600 keyFile
sudo chown 999:999 keyFile
3、配置config集群
docker-compose -f docker-compose-mongodb-shard.yml exec config1 bash -c "echo 'rs.initiate({_id: \"fates-mongo-config\",configsvr: true, members: [{ _id : 0, host : \"config1:27021\" },{ _id : 1, host : \"config2:27022\" }, { _id : 2, host : \"config3:27023\" }]})' | mongo --port 27021"
4、初始化3个分片服务器
docker-compose -f docker-compose-mongodb-shard.yml exec shard1 bash -c "echo 'rs.initiate({_id: \"shard1\",members: [{ _id : 0, host : \"shard1:27011\" }]})' | mongo --port 27011"
docker-compose -f docker-compose-mongodb-shard.yml exec shard2 bash -c "echo 'rs.initiate({_id: \"shard2\",members: [{ _id : 0, host : \"shard2:27012\" }]})' | mongo --port 27012"
docker-compose -f docker-compose-mongodb-shard.yml exec shard3 bash -c "echo 'rs.initiate({_id: \"shard3\",members: [{ _id : 0, host : \"shard3:27013\" }]})' | mongo --port 27013"
5、将mongo分片加入 mongos
docker-compose -f docker-compose-mongodb-shard.yml exec mongos bash -c "echo 'sh.addShard(\"shard1/shard1:27011\")' | mongo"
docker-compose -f docker-compose-mongodb-shard.yml exec mongos bash -c "echo 'sh.addShard(\"shard2/shard2:27012\")' | mongo"
docker-compose -f docker-compose-mongodb-shard.yml exec mongos bash -c "echo 'sh.addShard(\"shard3/shard3:27013\")' | mongo"
6、登陆 mongos ,创建用户
创建 admin用户，通过 admin 用户创建 其他用户及数据库
# 进入容器
docker exec -it 18b325c28de7 /bin/bash
# 进入mongo命令模式
mongo
use admin
show collections
# 创建 admin用户
admin = db.getSiblingDB("admin")
admin.createUser({user: "admin",pwd: "111111", roles: [{role: "userAdminAnyDatabase", db: "admin"} ,{role: "clusterAdmin", db: "admin"}]});
# 切换为 admin
db.auth("admin","111111")
show collections
# 创建不同数据库的用户
db.createUser({user:"testUser",pwd:"111111", roles:[{role:"readWrite",db:"account"}]});
db.createUser({user:"tom",pwd:"111111", roles:[{role:"readWrite",db:"video"}]});
# testUser 在account 库占增加 表 t_user
db.auth("testUser","111111")
use account
show collections
db.createCollection("t_user")
show collections
# 切换 成 tom 用户，但 tom 用户没有权限 查看 account库
use admin
db.auth("tom","111111")
use account
show collections
>> Warning: unable to run listCollections, attempting to approximate collection names by parsing connectionStatus
7、显示分片
mongo 分片 键目前有两种:
- 范围
- hash分片
操作
- 指定 video 为分片库
- 指定分片 collection 为 person，如果 collection 已经存在，则需要创建索引！！
# 切换为admin角色
use admin
db.auth("admin","111111")
# 指定 video 为分片库
sh.enableSharding("'video")
# 指定分片 collection 为 person， collection 不存在，不需要额外创建 索引,这里指定 hash 分片
sh.shardCollection("video.person",{id:"hashed"})
# 查看 MongoDB 分片集群的状态和配置信息
sh.status()
--- Sharding Status ---
shards:
{ "_id" : "shard1", "host" : "shard1/shard1:27018", "state" : 1 }
{ "_id" : "shard2", "host" : "shard2/shard2:27018", "state" : 1 }
{ "_id" : "shard3", "host" : "shard3/shard3:27018", "state" : 1 }
active mongoses:
"4.2.21" : 1
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:
682 : Success
databases:
{ "_id" : "'video", "primary" : "shard3", "partitioned" : true, "version" : { "uuid" : UUID("6d262f75-62e9-4b83-a2ae-d688bd59fba5"), "lastMod" : 1 } }
{ "_id" : "account", "primary" : "shard2", "partitioned" : true, "version" : { "uuid" : UUID("37b1fcf4-2923-461f-ac54-e2bfb312cdcd"), "lastMod" : 1 } }
{ "_id" : "config", "primary" : "config", "partitioned" : true }
{ "_id" : "video", "primary" : "shard3", "partitioned" : true, "version" : { "uuid" : UUID("71bc47c0-001b-466c-92ba-f9eee14af382"), "lastMod" : 1 } }
video.person
shard key: { "id" : "hashed" }
unique: false
balancing: true
chunks:
shard1 2
shard2 2
shard3 2
{ "id" : { "$minKey" : 1 } } -->> { "id" : NumberLong("-6148914691236517204") } on : shard1 Timestamp(1, 0)
{ "id" : NumberLong("-6148914691236517204") } -->> { "id" : NumberLong("-3074457345618258602") } on : shard1 Timestamp(1, 1)
{ "id" : NumberLong("-3074457345618258602") } -->> { "id" : NumberLong(0) } on : shard2 Timestamp(1, 2)
{ "id" : NumberLong(0) } -->> { "id" : NumberLong("3074457345618258602") } on : shard2 Timestamp(1, 3)
{ "id" : NumberLong("3074457345618258602") } -->> { "id" : NumberLong("6148914691236517204") } on : shard3 Timestamp(1, 4)
{ "id" : NumberLong("6148914691236517204") } -->> { "id" : { "$maxKey" : 1 } } on : shard3 Timestamp(1, 5)
# 可以看见每个分片所管理的hash 范围
8、新增数据，观察分片
# 切换为tom
use admin
db.auth("tom","111111")
user video
# 初始分片数据
db.person.stats()
# 插入数据
var arr=[];
for(var i=0;i<300;i++){
arr.push({"id":i,"name":"user"+i});
}
db.person.insertMany(arr);
# 查看当前 collection 分片情况
db.person.stats()
{
"sharded" : true,
"capped" : false,
"nindexes" : 2,
"nchunks" : 6,
"shards" : {
"shard1" : {
"ns" : "video.person",
"size" : 5473836,
"count" : 100205,
"avgObjSize" : 54,
"storageSize" : 1699840,
"capped" : false
"nindexes" : 2,
"indexBuilds" : [ ],
"totalIndexSize" : 4583424,
},
"shard2" : {
"ns" : "video.person",
"size" : 5451608,
"count" : 99791,
"avgObjSize" : 54,
"storageSize" : 1740800,
"capped" : false
"operationTime" : Timestamp(1664434149, 1)
},
"shard3" : {
"ns" : "video.person",
"size" : 5463446,
"count" : 100004,
"avgObjSize" : 54,
"storageSize" : 1703936,
"capped" : false
}
}
}
可以发现，hashed 分片键 把文档都 很均匀的分布在每个分片
9、一些其它命令
# 先让当前库支持分片
sh.enableSharding("'video")
# 当前集合加入分片，递增片键(timestamp已经作为索引了)
sh.shardCollection('video.person',{id:1})
# 刷新路由
db.adminCommand("flushRouterConfig")
# 让当前分片支持平衡
sh.enableBalancing("video.person")
# 开启平衡
sh.startBalancer()
#查看详细分片信息
sh.status({"verbose":1})
# 查看该表分片数据信息
db.getSiblingDB("video").t_video.getShardDistribution()
