MongoDB Master-Slave mode deployment

MongoDB deployment adopts master-slave mode, the main node is: Peiwo-turn4, from node to Peiwo-proxy

1, create deployment User and deployment path, log, data storage path

Perform the following actions on the Peiwo-proxy host

Groupadd MongoDB

Useradd-g mongodb-d/usr/local/mongodb MongoDB

Mkdir-p/data0/mongodb/logs #日志文件存放路径

Mkdir-p/data0/mongodb/data/master #数据文件存放路径

Mkdir-p/data0/mongodb/pid #pid文件存放路径

Mkdir-p/usr/local/mongodb/etc #配置文件存放路径

Log, data storage path authorization: Chown-r Mongodb.mongodb/data0/mongodb

Perform the following actions on the PEIWO-TURN4 host

Groupadd MongoDB

Useradd-g mongodb-d/usr/local/mongodb MongoDB

Mkdir-p/data1/mongodb/logs #日志文件存放路径

Mkdir-p/data1/mongodb/data/master #数据文件存放路径

Mkdir-p/data1/mongodb/pid #pid文件存放路径

Mkdir-p/usr/local/mongodb/etc #配置文件存放路径

Log, data storage path authorization: Chown-r Mongodb.mongodb/data1/mongodb

2. Download the software and unzip it

Curl-o http://fastdl.mongodb.org/linux/mongodb-linux-x86_64-1.6.3.tgz && TAR-ZXVF mongodb-linux-x86_ 64-1.6.3.tgz-c/usr/local/mongodb

3, the new configuration file under/usr/local/mongodb/etc

The PEIWO-TURN4 host configuration file is as follows: master.conf

Dbpath=/data1/mongodb/data/master

Logpath=/data1/mongodb/logs/master.log

Pidfilepath=/data1/mongodb/pid/master.pid

Directoryperdb=true

Logappend=true

port=21000

Master = True

source=10.144.160.124:21000

oplogsize=100

Fork=true

Noprealloc=true

The peiwo-proxy host configuration file is as follows: slave.conf

Dbpath=/data0/mongodb/data/slave

Logpath=/data0/mongodb/logs/slave.log

Pidfilepath=/data0/mongodb/pid/slave '. PID

Directoryperdb=true

Logappend=true

port=21000

Slave = True

source=10.144.145.235:21000

oplogsize=100

Fork=true

Noprealloc=true

Autoresync=true

4. Start/stop/Process View

Login to MongoDB server, switch to MongoDB User: Su-mongodb

Start process:./start.sh

Stop process: To end MongoDB, you can killall Mongod, or./mongo in the console, use Admin, and then Db.shutdownserver () to end the MongoDB server. Direct kill-9 will cause the MONGO to exit abnormally and may cause data corruption problems.

View process Run Status:./ps.sh

View log:./taillog.sh

Login MongoDB client:./connect.sh

Master and Slave daily operation and Maintenance command:

1. If you find that the master never synchronizes, manually sync from the top

Db.runcommand ({"Resync": 1})

2. Status Query

Db.runcommand ({"IsMaster": 1})

3, query the main library address on the cluster library

> Use local;

Switched to DB Local

> Db.sources.find ();

{"_id": ObjectId ("5551769ddf22026a698a271d"), "host": "10.144.145.235:21000", "source": "Main", "Syncedto": Timestamp (1431510735, 1)}

4. Check the status of each collection

Db.printcollectionstats ();

5. View master-slave replication status

Db.printreplicationinfo ();

6. Test Master and slave

Create a new database on the primary server

MONGO--port 21000

>show DBS

>use elaindb

>db.blog.save ({title: "new Article"})

Viewing synchronized data from a server

MONGO--port 21000

The operation to insert data from the library will prompt not master

> Use elaindb;

Switched to DB elaindb

> Db.blog.find ();

{"_id": ObjectId ("4e9174b48443c8ef12b30c56"), "title": "New Article"}

To view master-slave synchronization information:

> Db.printreplicationinfo ();

This is a slave, printing slave replication info.

source:10.0.0.11:27001

Syncedto:sun Oct 18:37:12 gmt+0800 (CST)

= secs ago (0.01hrs)

>db.printslavereplicationinfo ();

This is the synchronized data, the test is successful.

7, Slave replacement Master

If the primary server 10.144.145.235 down,

You need to replace the master service with a 10.144.160.124 machine.

The steps are as follows:

Stop 10.144.160.124 process (Mongod)

Kill-2 ' Ps-ef|grep mongod|grep-v Grep|awk ' {print $} '

#删除 local.* in the 10.144.160.124 data Catalog

Rm-rf/data1/mongodb/data/master//local.*

#以--master mode start 10.144.160.124

Appendix II, Switching Master/slave roles

Switch between the main library 10.144.145.235 and the role from the library 10.144.160.124

The steps are as follows: (Command slightly)

Pauses the write operation on the main library with the Fsync command.

To close a service from the library

Emptying the local.* file from the library

Restart from library service with-MASTER option

Performs a write operation from the library, initializes the Oplog, and obtains a synchronous start point

Close from the library service, at which point the new local.* file is already available from the library

Close the main library service and replace the local.* file on the main library with a new local.* file from the library (it is recommended to compress and copy first)

Restart from library service with-MASTER option

Restart the main library service by adding a-fastsync option on the option to start slave

This article is from the "My Ops Time" blog, so be sure to keep this source http://aaronsa.blog.51cto.com/5157083/1741516

MongoDB Master-Slave mode deployment