Deployment and operation of mongodb partitioned cluster

preparation ： Three servers are used for partition cluster planning , Fill in the relevant information in the table below , Test whether the communication is interconnected , If there is no interworking , Configure communication interworking . see Linux Firewall status , Check if it is closed , If not closed , Turn off the firewall . see Linux in SELinux Whether it is invalid , If not , Then it is modified as invalid .

virtual machine name	Host name	IP Address	Shard1	Shard2	Shard3	mongos	Config Server
Hadoop01	Hadoop01	192.168.121.134	27018 The main node	27020 Arbitration node	27019 Secondary node	27021	27022 The main node
Hadoop02	Hadoop02	192.168.121.135	27019 Secondary node	27018 The main node	27020 Arbitration node	27021	27022 Secondary node
Hadoop03	Hadoop03	192.168.121.136	27020 Arbitration node	27019 Secondary node	27018 The main node		27022 Secondary node

1、 Create related folders and files , install mongodb. Create related folders and files on the first server , install mongodb, And distribute the relevant directories to the other two , Make the directory and installation of the three servers the same .（20 branch ）

# Create a folder to use later

[[email protected]~]#touch/opt/servers/mongodb_demo/shardcluster/configServer/logs/config_server.log

[[email protected] ~]# touch /opt/servers/mongodb_demo/shardcluster/shard/logs/shard1.log

[[email protected] ~]# touch /opt/servers/mongodb_demo/shardcluster/shard/logs/shard2.log

[[email protected] ~]# touch /opt/servers/mongodb_demo/shardcluster/shard/logs/shard3.log

[[email protected] ~]# touch /opt/servers/mongodb_demo/shardcluster/mongos/logs/mongos.log

[[email protected] ~]# tar -zxvf /opt/software/mongodb-linux-x86_64-rhel62-3.6.23.tgz -C /opt/servers/mongodb_demo/shardcluster/

# Copy the folder to the other two virtual machines

scp -r /opt/servers/mongodb_demo/shardcluster/ [email protected]:/opt/servers/mongodb_demo/

scp -r /opt/servers/mongodb_demo/shardcluster/ [email protected]:/opt/servers/mongodb_demo/

2、 Deploy configuration server .（20 branch ）

3、 Deploy the shard server .（20 branch ）

Configure the partitioned master node 、 Secondary node 、 Arbitration node

./mongo --host hadoop01 --port 27018

rs.initiate()

rs.add("hadoop02:27019")

rs.addArb("hadoop03:27020")

rs.status()

./mongo --host hadoop02 --port 27018

rs.initiate()

rs.add("hadoop03:27019")

rs.addArb("hadoop01:27020")

rs.status()

./mongo --host hadoop03 --port 27018

rs.initiate()

rs.add("hadoop01:27019")

rs.addArb("hadoop03:27020")

rs.status()

4、 Deploy routing server .（20 branch ）

./mongos -f /opt/servers/mongodb_demo/shardcluster/mongos/configFile/mongodb_mongos.conf

5、 Start the fragmentation function , On the server nosql01 In the login mongos Of MongoDB client , Switch to database gateway, Add... To a sharded cluster 3 A shard .（20 branch ）

./mongo --host hadoop01 --port 27021

use gateway

sh.addShard("shard1/hadoop01:27018,hadoop02:27019,hadoop03:27020")

sh.addShard("shard2/hadoop01:27020,hadoop02:27018,hadoop03:27019")

sh.addShard("shard3/hadoop01:27019,hadoop02:27020,hadoop03:27018")

MongoDB Shard operation

1. Sign in 01 The server MongoDB client

./mongo --host hadoop01 --port 27021

2、 Switch to database config, Set partition block （chunk） The size is 1M. It is mainly to facilitate the display of slicing operation , And we'll slice chunk（ block ） Set to 1M, So that inserting a small amount of data can reflect the effect of fragmentation .

3、 Create a test database , Simulate inserting... Into a collection of databases 50000 Documents （ Documentation except for the default _id Field , It also includes id and name Field , Default _id Fields do not need to insert values , Automatic system generation , You can use for Cycle generation id and name value ）.

for(i=1;i<=50000;i++){db.user.insert({"id":i,"name":"jack"+i})}

4、 Partition the database ： Switch to database gateway, Slice the test database

Use gateway

sh.enableSharding("school")

5、 Slice the set ： Switch to database school, The collection of test databases is based on sharding keys id Create index , Switch to database gateway, With "id" As a set of partitioned key pairs user Slice .

6、 Switch to database gateway, Check the slice information .