CN102929666A - Method for deploying system quickly - Google Patents

Abstract

The invention provides a method for deploying a system quickly. The method comprises the following steps of: 1) making a backup for a needed operating system on a mother disc; 2) inserting a hard disc of a machine to be deployed together with the mother disc into a same machine and performing system startup information and operating system data reproduction by using Linux instructions and scripts; and 3) inserting the hard disc with reproduced data into the machine to be deployed to complete deployment. The method can effectively solve the problems of long time consumed for system deployment when the size of a data center becomes larger, requirement for addition of the drive of a hardware driver, and difficulty in manual deployment. Compared with a conventional system deployment method, the method can complete deployment only by the installation of a simple Linux system without purchase and installation of professional software. Compared with the conventional system deployment method, the method facilitates quick deployment of the operating system which is in complete consistency with the system in the mother disc. A user does not need worrying about the problem of inconsistent files caused by a damaged mirror image or installation medium.

Description

A kind of method of rapid deployment system

Technical field

The present invention relates to the Computer Applied Technology field, specifically a kind of method of rapid deployment system.

Background technology

Along with the development of computer technology and the universal expansion of computer utility, the continuous expansion of data center's scale, a large amount of servers are applied in data center, because the renewal speed of bottom hardware is more and more faster, deployment for operating system, be engraved in during requirement and change, and along with the increase of number, the deployment of operating system becomes the operation personnel's of data center a more unmanageable problem.There is more problem in the Ghost software of present business software such as Symantec aspect linux, thus the deployment of operating system especially the deployment of (SuSE) Linux OS become a hard problem.

Being scaled to of hard-disk capacity is a difficult problem in operation personnel system of the data center deployment.Need to carry out traditional subregion to hard disk in system's deployment, format.But occupied the more time for the jumbo formatting hard disk time, the time of especially present large capacity 2T 3T formatting hard disk approximately needs about 30 minutes more.A large amount of time is all for the format for fdisk.

So traditional system's mounting means rapid deployment require too wide in the gap.

Dispose the problem of time-consuming expense manpower for tradition, this method has proposed the method that a kind of new system disposes, thereby having reduced the time that system disposes, has saved human cost, thereby guaranteed the server product use of can reaching the standard grade fast, given full play to the value of fund input in the dilation process.

Summary of the invention

The method that the purpose of this invention is to provide a kind of rapid deployment system.

The objective of the invention is to realize in the following manner, concrete steps are as follows:

A. the backup of system's stamper: for the operating system of needs, use first as requested traditional LINUX PXE mode to do a backup at system's stamper of a capacity 2T, at stamper a centos 5.6x64 position system is installed; Stamper divides 7 subregions, wherein:

1) root partition, size, path :/dev/sda1

2) SWAP subregion: exchange partition, big or small 20G, path :/dev/sda2

3) Boot subregion: boot partition, big or small 300M, path :/dev/sda3

4) Home subregion: system main partition, big or small 20G, path/dev/sda4

5) Linshi subregion: interim subregion is deposited interim number, big or small 20G, path :/dev/sda5

6) Log subregion: be used for existing interim log information, big or small 10G, path/dev/sda6

7) Data subregion: storage master file data are used all residue subregions, big or small 1.8T/dev/sda7;

The hard disk of machine that B. will be to be disposed and stamper are inserted in the enterprising enforcement of uniform machinery and carry out system's log-on message and operating system data Replica with the linux script, and step is as follows:

Hard disk and the stamper of machine to be disposed are inserted on the machine of one 24 dish position, 24 dishes are arranged under the system, wherein the drive of stamper is/dev/sda, remaining 23 hard disk is destination disk, drive is from/dev/sdb to/dev/sdx, adopt copy command with copy script with os starting information and other data Replica in target hard disk:

1) calls the mbr file dd if=/dev/sda of=/home/mbr.img bs=512 count=1 of stamper

2) use script that 23 remaining dishes are carried out copying of stamper, write as two shell scripts herein, script information is as follows:

#!?/bin/bash

For I in thus b..x}/according to the hard disk number carry out mark herein 23 the dish from b to x/

do

Dd if=/home/mbr.img of=/dev/sd $ the i} bs=512 count=1/mbr information of stamper is write on each hard disk, execute this operation after, the mbr information of every hard disk just complete/

/ bin/bash $ i}/call $ the i} shell script to each subregion back up/

done

$ the i} script is write according to each name of calling, and this sentences b is example, and follow-up c writes according to b to x, then corresponding alphabetical b is revised as c and gets final product to the letter of x, and the order of b script is as follows:

Dd if=/dev/sda1 of=/dev/sdb1 bs=4K/system partitioning all carries out dd/

Dd if=/dev/sda2 of=/dev/sdb2 bs=4K/swap subregion all carries out dd/

Dd if=/dev/sda3 of=/dev/sdb3 bs=4K/boot subregion all carries out dd/

The new countless certificates of system of installing of dd if=/dev/sda4 of=/dev/sdb4 bs=4K count=10000/home subregion only need to be write out basic nodal information and get final product, so locate only need write a small amount of count/

dd?if=/dev/sda5?of=/dev/sdb5?bs=4K

dd?if=/dev/sda6?of=/dev/sdb6?bs=4K

Move behind the above-mentioned script about 2 minutes time, the system of 23 dishes namely backs up and finishes;

C. hard disk is taken off on the machine that directly is put into the needs startup, i.e. the completion system rapid deployment.

Excellent effect of the present invention is: the time that the saving system disposes: compare with traditional linux PXE installation, the softwares such as ghost of windows, this method can be finished a large amount of systems and dispose, thereby save the valuable time within the shorter time, has also reduced maintainer's workload.

Save cost: be different from traditional system and dispose, the replication form that this method proposes can be faster system is disposed

The difficulty of deployment system length consuming time when this method can effectively be improved data center's increase in size needs to add the problems such as the driving of hardware driver, artificial deployment difficulty.Compare traditional Account Dept's arranging method, this method only need to be installed simple linux system and can be finished affectedly, need not carry out the purchase of professional software and install.

Compare the legacy system dispositions method, this method can be disposed fast and be finished, and and being consistent fully of stamper, need not to worry because mirror image or the inconsistent problem of file that media damage causes is installed.

Embodiment

1. for the operating system of needs, do as requested the stamper backup of a system;

We use traditional LINUX PXE mode at the hard disk of a capacity 2T centos 5.6x64 position system to be installed herein.About time of 1 hour of this process need;

Wherein we divide 7 subregions

/ 30G root partition/dev/sda1 SWAP 20G exchange partition/dev/sda2/boot boot partition 300M/dev/sda3

Main partition 20G/the dev/sda4 of/home system/interim subregion of linshi is deposited ephemeral data 20G/dev/sda5

/ log is used for existing interim log information etc. to be used for 10G/dev/sda6 herein

/ data is used for the master file subregion and uses the nearly 1.8T/dev/sda7 of remaining all subregions

Here select another one that the machine of linux system is installed, and this machine hypothesis can be placed 24 hard disks, and the system that we can carry out 24 machines simultaneously disposes;

2. the hard disk of machine that will be to be disposed and stamper are inserted on the machine of one 24 dish position can see 24 dishes under our system, wherein the drive of stamper be/drive of remaining 23 dish of dev/sda arrives/dev/sdx for/dev/sdb

3. take out the mbr file dd if=/dev/sda of=/home/mbr.img bs=512 count=1 of stamper

4. use script that 23 remaining dishes are carried out copying of stamper, write as two shell scripts herein.

First is that the script information of main script is as follows

#!?/bin/bash

For I in thus b..x}/according to the hard disk number carry out mark herein 23 the dish from b to x/

do

Dd if=/home/mbr.img of=/dev/sd $ the i} bs=512 count=1/mbr information of stamper is write on each hard disk, execute this operation after, the mbr information of every hard disk just complete/

{ i}/invoke script program backs up each subregion/bin/bash $./

done

About { i} writes according to each name of calling, and this sentences b and is row, and follow-up c writes according to b to x and then corresponding alphabetical b is revised as c and gets final product to the letter of x;

The order of b script is as follows:

Dd if=/dev/sda1 of=/dev/sdb1 bs=4K/system partitioning all carries out dd/

Dd if=/dev/sda1 of=/dev/sdb1 bs=4K/swap subregion all carries out dd/

Dd if=/dev/sda1 of=/dev/sdb1 bs=4K/boot subregion all carries out dd/

The new countless certificates of system of installing of dd if=/dev/sda1 of=/dev/sdb1 bs=4K count=10000/home subregion only need to be write out basic nodal information and get final product, so locate only need write a small amount of count/

The new countless certificates of system of installing in dd if=/dev/sda1 of=/dev/sdb1 bs=4K/district only need to be write out basic nodal information and get final product, so locate only need write a small amount of count/

The new countless certificates of system of installing in dd if=/dev/sda1 of=/dev/sdb1 bs=4K/district only need to be write out basic nodal information and get final product, so locate only need write a small amount of count/

Move behind the above-mentioned script about 2 minutes time, the system of 23 dishes namely backs up and finishes.

Hard disk taken off directly to be put on the machine that need to start get final product.

Dispose separately with every hard disk of each machine and to compare, time efficiency has improved 690 times and has shortened greatly system deployment time.

Except the described technical characterictic of instructions, be the known technology of those skilled in the art.

Claims (1)

1. the method for a rapid deployment system is characterized in that concrete steps are as follows:

A. the backup of system's stamper: for the operating system of needs, use first as requested traditional LINUX PXE mode to do a backup at system's stamper of a capacity 2T, at stamper a centos 5.6x64 position system is installed; Stamper divides 7 subregions, wherein:

Root partition, size, path :/dev/sda1

SWAP subregion: exchange partition, big or small 20G, path :/dev/sda2

Boot subregion: boot partition, big or small 300M, path :/dev/sda3

Home subregion: system main partition, big or small 20G, path/dev/sda4

The Linshi subregion: interim subregion is deposited interim number, big or small 20G, path :/dev/sda5

Log subregion: be used for existing interim log information, big or small 10G, path/dev/sda6

The Data subregion: storage master file data are used all residue subregions, big or small 1.8T/dev/sda7;

The hard disk of machine that B. will be to be disposed and stamper are inserted in the enterprising enforcement of uniform machinery and carry out system's log-on message and operating system data Replica with the linux script, and step is as follows:

Hard disk and the stamper of machine to be disposed are inserted on the machine of one 24 dish position, 24 dishes are arranged under the system, wherein the drive of stamper is/dev/sda, remaining 23 hard disk is destination disk, drive is from/dev/sdb to/dev/sdx, adopt copy command with copy script with os starting information and other data Replica in target hard disk:

Call the mbr file dd if=/dev/sda of=/home/mbr.img bs=512 count=1 of stamper

Use script that 23 remaining dishes are carried out copying of stamper, write as two shell scripts herein, script information is as follows:

#!?/bin/bash

For I in thus b..x}/according to the hard disk number carry out mark herein 23 the dish from b to x/

do

Dd if=/home/mbr.img of=/dev/sd $ the i} bs=512 count=1/mbr information of stamper is write on each hard disk, execute this operation after, the mbr information of every hard disk just complete/

/ bin/bash $ i}/call $ the i} shell script to each subregion back up/

done

$ the i} script is write according to each name of calling, and this sentences b is example, and follow-up c writes according to b to x, then corresponding alphabetical b is revised as c and gets final product to the letter of x, and the order of b script is as follows:

Dd if=/dev/sda1 of=/dev/sdb1 bs=4K/system partitioning all carries out dd/

Dd if=/dev/sda2 of=/dev/sdb2 bs=4K/swap subregion all carries out dd/

Dd if=/dev/sda3 of=/dev/sdb3 bs=4K/boot subregion all carries out dd/

The new countless certificates of system of installing of dd if=/dev/sda4 of=/dev/sdb4 bs=4K count=10000/home subregion only need to be write out basic nodal information and get final product, so locate only need write a small amount of count/

dd?if=/dev/sda5?of=/dev/sdb5?bs=4K

dd?if=/dev/sda6?of=/dev/sdb6?bs=4K

Move behind the above-mentioned script about 2 minutes time, the system of 23 dishes namely backs up and finishes;

C. hard disk is taken off on the machine that directly is put into the needs startup, i.e. the completion system rapid deployment.