CN101819547A

CN101819547A - Method for testing stability and reliability of storage subsystem

Info

Publication number: CN101819547A
Application number: CN201010132051A
Authority: CN
Inventors: 孙波; 蔡积淼
Original assignee: Langchao Electronic Information Industry Co Ltd
Current assignee: Inspur Electronic Information Industry Co Ltd
Priority date: 2010-03-25
Filing date: 2010-03-25
Publication date: 2010-09-01

Abstract

The invention provides a method for testing the stability and reliability of a storage subsystem. Due to a special application environment and client requirements, a single storage server is generally requested to be provided with more than 12 hard disks, the following test method is made and a test script is formed to perform related test on the reliability and stability of the storage server aiming at the user-related application in order to ensure the reliability and stability for long-term running of the system; because many hard disks are used by the storage server, the product has serious hidden troubles in the stability such as out-of-order magnetic disk, drifting, large-pressure reading and writing environment, disconnection of the magnetic disk, IO error and the like before use of the test method; after use of the test method of the invention, the hidden troubles can be found and solved at the beginning of the development so as to solve similar lot size problem during the product supply and greatly improve the stability and reliability for long-term running of the product.

Description

A kind of method of testing at storage subsystem stability and reliability

Technical field

The present invention relates to a kind of Computer Applied Technology field, specifically a kind of method of testing at storage subsystem stability and reliability.

Background technology

Because the hard disk that storage server uses is more, does not use before this method of testing, has a lot of stable hidden danger in the product, for example: out of order, the drift of disk, problems such as big pressure is read and write under the environment, and disk goes offline, IO reports an error also do not have good solution at present.

Summary of the invention

The purpose of this invention is to provide a kind of method of testing at storage subsystem stability and reliability.

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

1) disk sequence check test: write hard disk positioned in sequence script under the linux system, realize that the order of server configures hard disk is is intermittently read and write, thereby realize the verification of logical order corresponding relation under hard disc physical order and the system;

2) disc hot insert functional test: write hard disk under the linux system and continue pressure read-write script and carry out, when disk is done big pressure read-write operation, pull out certain piece or several hard disks at random, and use script logging related log record and operation note; Again the hard disk that pulls out is turned back to system again according to the physics sequencing afterwards, check and write down the associative operation record, contrast the situation of change of plug front and back device number simultaneously, check hard disc apparatus drift problem whether to exist with this;

IO readwrite performance influence to total system when 3) disk breaks down in the system is tested: write disk under the linux system and continue pressure read-write script and disk I state recording script, when disk is done big pressure read-write operation, pull out a hard disk in the system at random, utilize the system IO information of script logging iostat instrument report simultaneously, thereby whether the IO situation of checking system is normal;

4) disk continues pressure test: write hard disk pressure test script under the linux system, comprise the circular order read-write operation of data blocks such as 4K, 16K, 64K, 128K, 256K, 512K, 1M, with the actual application environment of this analog subscriber.Carry out the pressure test script, 3 working days of circular flow;

The invention has the beneficial effects as follows: the whole even formed product transboundary in fusion design back in part is carried out on the 2-8 road technically can allow the user buy the work that a two-way MP product substitutes low before configuration 4 road MP products with lower cost, necessary if any upgrading, can be at any time be upgraded to 4 tunnel 8 road MP servers even with the mode of expansion module.Avoided purchasing new server once more, reduced space for its deployment, handling cost and purchase cost and risen, other technologies such as combined with virtualization more can realize work such as headachy system applies migration.

Storage server is because of its application circumstances and customer demand, generally require 12 above hard disks of stand-alone configuration, therefore in order to guarantee the long-time running reliability and stability of this system, formulate following method of testing, and the formation test script, at user's related application storage server is done related reliability and stability test:

Embodiment

Storage server has been widely used in the Internet user at present, in order to improve stability and the reliability of this series products in the application process of reaching the standard grade, formulates this method of testing.

Embodiment

One. the environment of test

The tide storage server

The (SuSE) Linux OS environment

Two. testing procedure (implementation method):

Three. test data and defining standard:

1) in the disk sequence check process, logical order can be mapped one by one under the physical sequential of disk and the system, then calculate test and pass through, otherwise test is not passed through;

2) in the disc hot insert functional test, disk is after pulling out, and there are relevant daily record and warning message prompting in system, and after hard disk was inserted again, system can normally discern and use, and then calculate this test and pass through, otherwise test is not passed through;

In the IO readwrite performance influence test to total system when 3) disk breaks down in the system, after a hard disk in the system was removed, therefore the IO performance of other disks was not affected in the system, still can normally read and write, how to calculate this test and pass through, otherwise test is not passed through;

4) disk continues in the pressure test, and system journal does not have and reports an error, and disk does not have relevant informations such as the dish of falling, IO report an error, and disk read-write is normal, how to calculate test to pass through, otherwise test is not passed through;

Four. the effect of test

Because the hard disk that storage server uses is more, does not use before this method of testing, has a lot of stable hidden danger in the product, for example: out of order, the drift of disk, under the big pressure read-write environment, problem such as disk goes offline, IO reports an error; After using this method of testing, this type of potential problem just can be found at the beginning of exploitation and in time be solved, and has avoided occurring similar lot-size problem in the subsequent product supply of material process, has greatly improved the stability and the reliability of product long-time running.

Claims

1. the method for testing at storage subsystem stability and reliability is characterized in that, in order to guarantee the long-time running reliability and stability of this system, formulates method of testing, and forms test script, and testing procedure is as follows:

A. disk sequence check test: under the linux system, write hard disk positioned in sequence script, realize that the order of server configures hard disk is is intermittently read and write, thereby realize the verification of logical order corresponding relation under hard disc physical order and the system;

B. disc hot insert functional test: under the linux system, write hard disk and continue pressure read-write script and carry out, when disk is done big pressure read-write operation, pull out certain piece or several hard disks at random, and use script logging related log record and operation note; Again the hard disk that pulls out is turned back to system again according to the physics sequencing afterwards, check and write down the associative operation record, contrast the situation of change of plug front and back device number simultaneously, check hard disc apparatus drift problem whether to exist with this;

IO readwrite performance influence to total system when c. disk breaks down in the system is tested: write disk under the linux system and continue pressure read-write script and disk I state recording script, when disk is done big pressure read-write operation, pull out a hard disk in the system at random, utilize the system IO information of script logging iostat instrument report simultaneously, thereby whether the IO situation of checking system is normal;

D. disk continues pressure test: write hard disk pressure test script under the linux system, the circular order read-write operation that comprises 4K, 16K, 64K, 128K, 256K, 512K, 1M data block, with the actual application environment of this analog subscriber, carry out the pressure test script, 3 working days of circular flow.