CN106610877A - Method for testing hard disk data consistency - Google Patents

Abstract

The invention provides a method for testing hard disk data consistency and belongs to the field of hard disk data testing. The method comprises the following steps of mounting a partition of a hard disk into a file system after hard disk partition and partition formatting, viewing whether data is consistent through copying and comparing data in the file system mounted by the hard disk, and if the data is inconsistent, recording differences; circulating the copy and the comparison between the hard disks for more than once, simulating a long-time operation process of customer business, recording the differences and viewing the differences of the data; and viewing a result to see whether data inconsistency condition is produced in a copy process. According to the method, the operations can simplified and the manpower and the time can be saved.

Description

A kind of method of testing hard disk data consistency

Technical field

The present invention relates to hard disc data measuring technology, more particularly to a kind of method of testing hard disk data consistency.

Background technology

With the continuous development and growth of tide server industry, as the server provider of largest domestic, server Every field has been included in market, especially the main flow field such as finance, internet, the high speed development of internet, the Gao An of information The high reliability of full property and system increasingly becomes company and enterprise, the thing that government pays close attention to the most, and hard disk is used as data storage Main carriers, its stability and reliability be also increasingly taken seriously, thus for hard disk use on our servers when Often there are backup and the dump of hard disc data to occur, during hard disc data is shifted, the data consistency between hard disk It is critically important.Based on requirements above, a kind of method testing hard disk data consistency under system is needed in test process.

The content of the invention

In order to solve problem above, the present invention proposes a kind of method of testing hard disk data consistency.Can apply to The conforming test of hard disc data under various configurations, realizes quick, efficient, accurate, automation the contrast for carrying out difference, and And the result of contrast is simple and clear, the quality to improving server product provides reliable guarantee.

The present invention is easy to the mistake occurred during finding copying and dump.The replicate data on fdisk, By duplication and contrast to data etc. operate, afterwards with initial data contrast come complete pressure test discover whether have number According to uniformity.

The technical scheme is that：

A kind of method of testing hard disk data consistency,

By by fdisk again by partition format after, the subregion of hard disk is mounted in file system, by data（Text Part）Replicate and compare in the file system of hard disk institute carry, check whether that data are consistent and difference is recorded if inconsistent； More than circulation primary replicate and compare between multiple hard disks, the prolonged running of simulation client traffic, record difference is looked into See the otherness of data.Result is then looked at, sees the inconsistent situation of data whether is generated during duplication.

Idiographic flow is：

1）Linux system is installed and obtains root authority,

2）For hard disk or logic magnetic disc subregion and format；

3）The fdisk of formatting is mounted under file system；

4）File is copied to into another file from a file；

5）Relatively the difference of two folder contents, checks whether consistent；

6）If consistent, continue to replicate, then check difference；

If inconsistent, record difference and continue to replicate, then check difference.

Preparation

Need that hard disk to be measured is linked in system, Linux system is installed, Src files are created under/root, non-two are entered In file copy to Src files processed.

Specifically testing procedure is：

1）Subregion is carried out to the hard disk in server system；

2）Fdisk is formatted into into the form of EXT4；

3）Set up equal with hard disk quantity under/opt, and the corresponding file of title；

4）Fdisk is mounted under corresponding file；

5）Recursive copying and the deletion in the subregion of hard disk by the All Files in Src files；

6）The difference for relatively replicating every time；

7）Difference is recorded, test result is returned.

The invention has the beneficial effects as follows

The complexity of manual test when reducing test in the past by the method for automatic running, it is possible to reduce simplify operation, Save manpower and time.

Description of the drawings

Fig. 1 is testing process schematic diagram.

Specific embodiment

Below more detailed elaboration is carried out to present disclosure：

The present invention verification step be：

Hard disk to be measured is linked in system, Linux system is installed, Src files are created under/root, by nonbinary text In part copy to Src files.

When the 1st, running copy and compare, copy number is shown；

2nd, when operation is completed, there are " Test finished1000times！”

3rd, test.log is checked, the number of times being successfully, reproduced can be checked and whether passed through.

Attached script and explanation

Script deal_disk is the script to fdisk and formatting, wherein being 40G to the size of hard disk distribution, is formatted Pattern is the form of ext4.

#!/bin/bash

PATH=/bin:/sbin:/usr/bin:/usr/sbin

export PATH

for i in {b..d};

do

fdisk /dev/sd$i <<ESXU

n

p

1

+40G

w

ESXU

sleep 3s

mkfs.ext4 /dev/sd"$i"1

done

Script mount_disk is establishment file folder and fdisk to be mounted to into the script of corresponding document folder under/root

#/bin/bash

for i in {b..d};

do

umount /root/test/$i

done

rm -rf /root/test

mkdir /root/test

for i in {b..d};

do

mkdir /root/test/$i

done

for i in {b..d};

do

mount /dev/sd"$i"1 /root/test/$i

done

Script copy_and_compare is by the file under the nonbinary file copy in Src to/opt, and at each The script of multiple copy and compare is constantly carried out in file, cp orders used in script are replicated using diff orders Carry out contrasting verification and comparing result being recorded in diffsrc.txt and diff.txt, while can also will appear from the ground of difference Side also records, and is checked after the completion of convenient test, and copy number and result are put into into test.log and Success.log In.

#!/bin/bash

TestCount=1000

rm diff.txt

rm diffsrc.txt

rm SuccessCount.log

rm test.log

for i in {b..d};

do

rm -rvf /root/test/$i/*

done

cp -a /root/Src/* /root/test/b

diff /root/Src /root/test/b |tee -a diffsrc.txt

diffsrc=$(cat diffsrc.txt | wc -l)

if [ "$diffsrc" != "0" ]; then

echo "Compare with the source files fail !!!"

exit 1

fi

count=0

###copy from disk to disk###

while (true);do

######copy from b to c

cp -a /root/test/b/* /root/test/c

count=$(($count + 1))

diff /root/test/b /root/test/c |tee -a diff.txt

diffnum=$(cat diff.txt |wc -l)

if [ "$diffnum" != "0" ];then

echo "failed at "$count "times test" | tee -a diff.txt

echo "failed at when copy from sdb to sdc"|tee -a diff.txt

cat SuccessCount.log >>test.log

break

fi

echo "Successfully copy files "$count"times !" |tee -a SuccessCount.log

rm -rvf /root/test/b/*

if [ -n "${TestCount}" ];then

if [ "$count" == "$TestCount" ];then

echo "Test finished"$count"times!"

cat SuccessCount.log>> test.log

echo "PASS !!!">> test.log

exit

fi

fi

######copy from c to d

cp -a /root/test/c/* /root/test/d

count=$(($count + 1))

diff /root/test/c /root/test/d |tee -a diff.txt

diffnum=$(cat diff.txt |wc -l)

if [ "$diffnum" != "0" ];then

echo "failed at "$count "times test" | tee -a diff.txt

echo "failed at when copy from sdc to sdd"|tee -a diff.txt

cat SuccessCount.log >>test.log

break

fi

echo "Successfully copy files "$count"times !" |tee -a SuccessCount.log

rm -rvf /root/test/c/*

if [ -n "${TestCount}" ];then

if [ "$count" == "$TestCount" ];then

echo "Test finished"$count"times!"

cat SuccessCount.log>> test.log

echo "PASS !!!">> test.log

exit

fi

fi

######copy from d to b

cp -a /root/test/d/* /root/test/b

count=$(($count + 1))

diff /root/test/d /root/test/b |tee -a diff.txt

diffnum=$(cat diff.txt |wc -l)

if [ "$diffnum" != "0" ];then

echo "failed at "$count "times test" | tee -a diff.txt

echo "failed at when copy from sdd to sdb"|tee -a diff.txt

cat SuccessCount.log >>test.log

break

fi

echo "Successfully copy files "$count"times !" |tee -a SuccessCount.log

rm -rvf /root/test/d/*

if [ -n "${TestCount}" ];then

if [ "$count" == "$TestCount" ];then

echo "Test finished"$count"times!"

cat SuccessCount.log>> test.log

echo "PASS !!!">> test.log

exit

fi

fi

done。

This method can apply to the conforming test of hard disc data under various configurations, realize it is quick, efficiently, accurately, The contrast for carrying out difference of automation, and the result for contrasting is simple and clear, and the quality to improving server product is provided can The guarantee leaned on.

Claims (4)

1. a kind of method of testing hard disk data consistency, it is characterised in that

By by fdisk again by partition format after, the subregion of hard disk is mounted in file system, by existing to data Replicate and compare in the file system of hard disk institute carry, check whether that data are consistent and difference is recorded if inconsistent；

More than circulation primary replicate and compare between multiple hard disks, the prolonged running of simulation client traffic, it is poor to record The different otherness for checking data；

Result is then looked at, sees the inconsistent situation of data whether is generated during duplication.

2. method according to claim 1, it is characterised in that

Idiographic flow is：

1）Linux system is installed and obtains root authority,

2）For hard disk or logic magnetic disc subregion and format；

3）The fdisk of formatting is mounted under file system；

4）File is copied to into another file from a file；

5）Relatively the difference of two folder contents, checks whether consistent；

6）If consistent, continue to replicate, then check difference；

If inconsistent, record difference and continue to replicate, then check difference.

3. method according to claim 2, it is characterised in that

Before test, need that hard disk to be measured is linked in system, Linux system be installed, and Src files are created under/root, By in nonbinary file copy to Src files.

4. method according to claim 3, it is characterised in that

Testing procedure is：

1）Subregion is carried out to the hard disk in server system；

2）Fdisk is formatted into into the form of EXT4；

3）Set up equal with hard disk quantity under/opt, and the corresponding file of title；

4）Fdisk is mounted under corresponding file；

5）Recursive copying and the deletion in the subregion of hard disk by the All Files in Src files；

6）The difference for relatively replicating every time；

7）Difference is recorded, test result is returned.