CN112445660A - Test method for performance balance problem of shared storage array - Google Patents

Test method for performance balance problem of shared storage array Download PDF

Info

Publication number
CN112445660A
CN112445660A CN201910801001.9A CN201910801001A CN112445660A CN 112445660 A CN112445660 A CN 112445660A CN 201910801001 A CN201910801001 A CN 201910801001A CN 112445660 A CN112445660 A CN 112445660A
Authority
CN
China
Prior art keywords
test
balance
equilibrium
control module
appointed
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201910801001.9A
Other languages
Chinese (zh)
Other versions
CN112445660B (en
Inventor
何晓斌
高洁
魏巍
余婷
陈起
肖伟
王涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuxi Jiangnan Computing Technology Institute
Original Assignee
Wuxi Jiangnan Computing Technology Institute
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuxi Jiangnan Computing Technology Institute filed Critical Wuxi Jiangnan Computing Technology Institute
Priority to CN201910801001.9A priority Critical patent/CN112445660B/en
Publication of CN112445660A publication Critical patent/CN112445660A/en
Application granted granted Critical
Publication of CN112445660B publication Critical patent/CN112445660B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/22Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing
    • G06F11/2205Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing using arrangements specific to the hardware being tested
    • G06F11/2221Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing using arrangements specific to the hardware being tested to test input/output devices or peripheral units
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/22Detection or location of defective computer hardware by testing during standby operation or during idle time, e.g. start-up testing
    • G06F11/2273Test methods

Abstract

The invention discloses a test method for the performance balance problem of a shared storage array, which is based on the following modules: the control module is used for configuring an equilibrium test end, an equilibrium test partition and an equilibrium test directory, generating and sending an equilibrium test scene command and an equilibrium test command, and collecting and summarizing an equilibrium test result; the test module is used for receiving and executing the command of the management terminal, for example, generating a balance test scene, carrying out balance tests of appointed thread number, appointed file size, appointed block size, appointed sequence write/read and appointed random write/read, recording and returning a balance test result to the control module. The invention can accurately test and judge whether the storage equipment has the performance balance problem when facing a plurality of servers, and can also judge whether the performance balance problem exists among a plurality of different LUNs on a single server.

Description

Test method for performance balance problem of shared storage array
Technical Field
The invention relates to a test method for the performance balance problem of a shared storage array, and belongs to the technical field of computers.
Background
With the rapid development of storage technology, disk arrays have been widely applied to scientific computing, online transaction processing, vod (video on demand), and other applications with high requirements on processing speed, capacity, and reliability. In the age of disk arrays, storage is more and more intelligent and independent, and the electrical independence of storage and a host is a necessary trend. Most of the control systems of a disk array have virtualization functions. At present, RAID virtualization technology is mostly adopted to plan the storage resources of the disk array, and the disks combined by RAID become a huge storage space. This huge storage space cannot be used by only one server, and another logical level, i.e., LUN, is virtualized in this huge storage space, and one host can access multiple LUNs in this huge storage space. Different manufacturers have different LUNs in different dividing methods, and some manufacturers adopt the same partition as a common operating system, and each LUN is a continuous large block in the RAID group space; some manufacturers have adopted a file system to manage LUNs directly to achieve high flexibility. At present, the key technology in the array is LUN, besides RAID, other advanced data processing functions are based on LUN, and the lower layer architecture directly determines the upper layer implementation mode and efficiency.
In order to make the disk array better meet the requirements of various application environments, it is very important to evaluate the performance of the disk array before use. The disk array performance test can mainly reflect indexes such as IOPS (input/output operation number per second), average response time and data transmission rate of the disk array, and has important guiding significance for network storage design and practical application.
At present, some researches have been made on disk array performance tests at home and abroad, and some typical evaluation specifications and evaluation tools, such as SPC, IOMeter, IOZone, fio, Bonnie + +, PostMark and the like, appear. Wherein, spc (storage Performance council) is a storage device Performance industry standard definition and management organization independent of manufacturers. It provides three test benchmarks, SPC-1, SPC-2 and SPC-3. SPC-1 and SPC-2 test the performance of the storage system by using predetermined IO operations, and simulate different types of service applications by different read-write modes and I/O operation sets. SPC-1 mainly simulates the performance of OLTP, Mail Server, etc. applications, while SPC-2 may simulate the performance of large file processing, large-scale database queries, Video On Demand (VOD), etc. applications. The goal of SPC-3 is to provide simulation at the application layer, with the main goal of simulating the performance of applications such as hierarchical storage, storage management, content management, information lifecycle management, etc. The IOZone and Bonnie + + can run on a plurality of operating systems and are good reference tools for file IO operation; the IOMeter can evaluate various common load conditions such as file servers, OLTP and Web services, and the working load specification of the IOMeter becomes a best reference tool for testing the IO performance of the Windows platform; PostMark is used by the famous NAS provider NetApp to test the back-end storage performance of products, is mainly used for testing the performance of a file system in a mail system or an electronic commerce system, and has the characteristic that small files need to be frequently accessed in a large quantity.
At present, in a large-scale storage system, an external storage array is often required to be used to provide a reliable shared storage space among a plurality of servers, and therefore, the storage array needs to be used uniformly among the servers, which is a main problem in the design of the storage array at present. At present, no test tool is used for testing the storage array balance, so that how to test the storage array balance becomes an effort for those skilled in the art.
Disclosure of Invention
The invention aims to provide a method for testing the performance balance problem of a shared storage array, which can accurately test and judge whether the performance balance problem exists when storage equipment faces a plurality of servers or whether the performance balance problem exists among a plurality of different LUNs on a single server.
In order to achieve the purpose, the invention adopts the technical scheme that: a test method for the performance balance problem of a shared storage array is based on the following modules:
the control module runs on an independent control server, is used for configuring an equilibrium test end, an equilibrium test partition and an equilibrium test directory, sending an equilibrium test flow generation command, an equilibrium test starting command and an equilibrium test stopping command and collecting and summarizing an equilibrium test result;
the test module runs on the test server, the test server can see the LUN partition of the shared storage array, and the test module is used for receiving and executing the command of the control module, for example, generating a balance test flow, performing balance test of appointed thread number, appointed file size, appointed block size, appointed sequential write/read or appointed random write/read, recording and returning a balance test result to the control module;
the method comprises the following steps:
s1, constructing an RAID group for the shared storage array, and dividing a plurality of LUN partitions with the same size on the RAID group, wherein the number of the LUN partitions is more than 2 times that of the test terminals;
s2, the test module uniformly uses the default parameters of the file system to format the file system of each LUN partition;
s3, after the file system is formatted, the test module uniformly uses the formatted LUN partitions in the default parameter mounting step S2 of the file system to mount the plurality of formatted LUN partitions to a plurality of test end servers respectively;
s4, the control module configures the test end, the test subarea and the test directory to be tested for the balance according to the result of the step S3, and sets the IP address of the test end, the LUN subarea name and the subarea mounting directory of the shared storage array;
s5, the control module sends the balance test flow generation command to each test terminal to be tested according to the configuration and formulates test parameters, wherein the test parameters comprise: thread number, file size, block size, proportion of sequential write/read type or mixed test, proportion of random write/read type or mixed test;
s6, each testing end receives the balance testing flow generation command from the control module, and locally generates a balance testing flow of the designated LUN partition at the testing end;
s7, the control module sends a balance test starting command to the test module of each test end, and the test module of each test end starts all test flows on the test end after receiving the starting command;
s8, after the test is finished, each test end records the test result of each test partition and transmits the test result back to the control module;
s9, the control module makes a summary analysis to the test result files of all the partitions of the test, and according to the analysis result, judges whether the performance of each partition of the storage device is balanced, that is, whether the storage device stores the performance balance problem when facing a plurality of servers, or whether the storage device stores the performance balance problem among a plurality of different LUNs on a single server.
Due to the application of the technical scheme, compared with the prior art, the invention has the following advantages:
the invention relates to a method for testing the performance balance problem of a shared storage array, which can accurately test and judge whether the performance balance problem exists when a storage device faces a plurality of servers or whether the performance balance problem exists among a plurality of different LUNs on a single server.
Drawings
FIG. 1 is a schematic diagram of the principle of the test method for the performance balance problem of the shared memory array according to the present invention;
FIG. 2 is a flow chart of a method for testing the performance balance of a shared memory array according to the present invention.
Detailed Description
Example (b): a test method for the performance balance problem of a shared storage array is based on a large-scale heterogeneous system and a shared storage array on a single system or a cluster system of a Linux platform and is based on the following modules:
the control module runs on an independent control server, is used for configuring an equilibrium test end, an equilibrium test partition and an equilibrium test directory, sending an equilibrium test flow generation command, an equilibrium test starting command and an equilibrium test stopping command and collecting and summarizing an equilibrium test result;
the test module runs on the test server, the test server can see the LUN partitions of the shared storage array, the number of the test servers of the shared storage array is more than or equal to 2 usually, the number of the LUN partitions which can be supported is more than 4, the test module is used for receiving and executing commands of the control module, for example, a balance test flow is generated, a balance test of specified thread number, specified file size, specified block size, specified sequential write/read or specified random write/read is carried out, and a balance test result is recorded and returned to the control module;
the method comprises the following steps:
s1, constructing an RAID group for the shared storage array, and dividing a plurality of LUN partitions with the same size on the RAID group, wherein the number of the LUN partitions is more than 2 times that of the test terminals;
s2, the test module uniformly uses the default parameters of the file system to format the file system of each LUN partition;
s3, after the file system is formatted, the test module uniformly uses the formatted LUN partitions in the default parameter mounting step S2 of the file system to mount the plurality of formatted LUN partitions to a plurality of test end servers respectively;
s4, the control module configures the test end, the test subarea and the test directory to be tested for the balance according to the result of the step S3, and sets the IP address of the test end, the LUN subarea name and the subarea mounting directory of the shared storage array;
s5, the control module sends the balance test flow generation command to each test terminal to be tested according to the configuration and formulates test parameters, wherein the test parameters comprise: thread number, file size, block size, proportion of sequential write/read type or mixed test, proportion of random write/read type or mixed test;
s6, each testing end receives the balance testing flow generation command from the control module, and locally generates a balance testing flow of the designated LUN partition at the testing end;
s7, the control module sends a balance test starting command to the test module of each test end, and the test module of each test end starts all test flows on the test end after receiving the starting command;
s8, after the test is finished, each test end records the test result of each test partition and transmits the test result back to the control module;
s9, the control module makes a summary analysis to the test result files of all the partitions of the test, and according to the analysis result, judges whether the performance of each partition of the storage device is balanced, that is, whether the storage device stores the performance balance problem when facing a plurality of servers, or whether the storage device stores the performance balance problem among a plurality of different LUNs on a single server.
And comparing the test results of each test between each LUN of each test end, if the variance of the test results is more than 10%, indicating that the performance of the test on different LUNs on the same test end is unbalanced, or summing the test results on each LUN of each test item on each test end, and comparing the test results between each test end, if the variance of the test results is more than 10%, indicating that the performance of the test on different test ends is unbalanced.
The examples are further explained below:
the testing method is designed for testing the balance of the shared storage array on a single system or a cluster system based on a Linux platform, can accurately judge whether the performance balance problem exists between a plurality of servers of the storage equipment or not, and can also judge whether the performance balance problem exists between a plurality of different LUNs on a single server or not.
The test method design comprises a control end and a test end.
The main functions of the control end comprise: configuring a balance test end, configuring a balance test partition, configuring a balance test directory, sending a balance test scene generation command, sending a balance test command, and collecting and summarizing balance test results.
The main functions of the test end include: receiving and executing the command of the management terminal, for example, generating a balance test scene, performing balance tests of a specified thread number, a specified file size, a specified block size, a specified sequence of writing/reading and a specified random writing/reading, and recording and returning a balance test result.
The preparation work includes:
firstly, a manufacturer constructs an RAID group for a shared storage array provided by the manufacturer, and divides a plurality of LUN partitions with the same size on the RAID group according to a certain distribution mode;
then, ext3 file system formatting is carried out on each LUN partition;
and mounting after the formatting is finished (when judging whether the storage equipment has the storage performance balance problem or not when facing to a plurality of servers, the plurality of formatted LUN partitions are respectively mounted on the designated mounting catalogues of the plurality of servers, and when judging whether the storage equipment has the storage performance balance problem or not when facing to 1 server, the plurality of formatted LUN partitions are mounted on the designated mounting catalogues of the single server).
The above preparation work needs only one operation.
After the preparation work is completed, the balance test can be performed, and the specific test flow is as follows:
1. the control end configures a test end, a test partition and a test directory to be subjected to the balance test, and sets an IP address of the distributed test end, a partition name of the shared storage array and a partition mounting directory;
2. the control end concurrently sends an equilibrium test scene generation command to each test end to be subjected to the equilibrium test according to the configuration;
3. after receiving the scene generation command, each distributed test end generates an equilibrium test scene in an equilibrium test partition appointed by the command and simulates an actual use scene;
4. after the balance test scene is generated, the control end sends balance test set commands of specified thread number, specified file size, specified block size, specified sequence writing/reading and specified random writing/reading to the distributed test end;
5. after receiving the test command, each distributed test end executes balance set test of appointed thread number, appointed file size, appointed block size, appointed sequence write/read and appointed random write/read to respective test partition appointed by the command;
6. after the test is finished, each test end records the test result of each test partition and transmits the test result back to the control end;
7. the control end carries out summary analysis on all the subarea test result files of the test;
8. the user can accurately judge whether the performance of each partition of the storage equipment is balanced according to the analysis result, so that whether the performance of the storage equipment is balanced facing a plurality of servers or whether the performance of a plurality of different LUNs on a single server is balanced can be accurately judged.
To facilitate a better understanding of the invention, the terms used herein will be briefly explained as follows:
OLTP: on-line Transaction Processing, also known as Transaction-oriented Processing, is an online Transaction process.
LUN: logical Unit Number, i.e., the Number of Logical units.
When the test method for the performance balance problem of the shared storage array is adopted, the test method can accurately test and judge whether the performance balance problem exists when the storage equipment faces a plurality of servers or not, and can also judge whether the performance balance problem exists among a plurality of different LUNs on a single server or not.
The above embodiments are merely illustrative of the technical ideas and features of the present invention, and the purpose thereof is to enable those skilled in the art to understand the contents of the present invention and implement the present invention, and not to limit the protection scope of the present invention. All equivalent changes and modifications made according to the spirit of the present invention should be covered within the protection scope of the present invention.

Claims (1)

1. A test method for the performance balance problem of a shared storage array is based on the shared storage array on a single system or a cluster system of a Linux platform, and is characterized in that: based on the following modules:
the control module runs on an independent control server, is used for configuring an equilibrium test end, an equilibrium test partition and an equilibrium test directory, sending an equilibrium test flow generation command, an equilibrium test starting command and an equilibrium test stopping command and collecting and summarizing an equilibrium test result;
the test module runs on the test server, the test server can see the LUN partition of the shared storage array, and the test module is used for receiving and executing the command of the control module, for example, generating a balance test flow, performing balance test of appointed thread number, appointed file size, appointed block size, appointed sequential write/read or appointed random write/read, recording and returning a balance test result to the control module;
the method comprises the following steps:
s1, constructing an RAID group for the shared storage array, and dividing a plurality of LUN partitions with the same size on the RAID group, wherein the number of the LUN partitions is more than 2 times that of the test terminals;
s2, the test module uniformly uses the default parameters of the file system to format the file system of each LUN partition;
s3, after the file system is formatted, the test module uniformly uses the formatted LUN partitions in the default parameter mounting step S2 of the file system to mount the plurality of formatted LUN partitions to a plurality of test end servers respectively;
s4, the control module configures the test end, the test subarea and the test directory to be tested for the balance according to the result of the step S3, and sets the IP address of the test end, the LUN subarea name and the subarea mounting directory of the shared storage array;
s5, the control module sends the balance test flow generation command to each test terminal to be tested according to the configuration and formulates test parameters, wherein the test parameters comprise: thread number, file size, block size, proportion of sequential write/read type or mixed test, proportion of random write/read type or mixed test;
s6, each testing end receives the balance testing flow generation command from the control module, and locally generates a balance testing flow of the designated LUN partition at the testing end;
s7, the control module sends a balance test starting command to the test module of each test end, and the test module of each test end starts all test flows on the test end after receiving the starting command;
s8, after the test is finished, each test end records the test result of each test partition and transmits the test result back to the control module;
s9, the control module makes a summary analysis to the test result files of all the partitions of the test, and according to the analysis result, judges whether the performance of each partition of the storage device is balanced, that is, whether the storage device stores the performance balance problem when facing a plurality of servers, or whether the storage device stores the performance balance problem among a plurality of different LUNs on a single server.
CN201910801001.9A 2019-08-28 2019-08-28 Test method for performance balance problem of shared storage array Active CN112445660B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910801001.9A CN112445660B (en) 2019-08-28 2019-08-28 Test method for performance balance problem of shared storage array

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910801001.9A CN112445660B (en) 2019-08-28 2019-08-28 Test method for performance balance problem of shared storage array

Publications (2)

Publication Number Publication Date
CN112445660A true CN112445660A (en) 2021-03-05
CN112445660B CN112445660B (en) 2022-09-13

Family

ID=74742119

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910801001.9A Active CN112445660B (en) 2019-08-28 2019-08-28 Test method for performance balance problem of shared storage array

Country Status (1)

Country Link
CN (1) CN112445660B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115801633A (en) * 2023-02-08 2023-03-14 湖南国科亿存信息科技有限公司 SAN storage multi-client sharing concurrency performance test method and system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105224637A (en) * 2015-09-24 2016-01-06 珠海许继芝电网自动化有限公司 A kind of based on PostgreSQL database active and standby/the comprehensive method of cluster application
CN107678901A (en) * 2017-10-30 2018-02-09 郑州云海信息技术有限公司 Disk performance test system and method based on Linux distributed systems

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105224637A (en) * 2015-09-24 2016-01-06 珠海许继芝电网自动化有限公司 A kind of based on PostgreSQL database active and standby/the comprehensive method of cluster application
CN107678901A (en) * 2017-10-30 2018-02-09 郑州云海信息技术有限公司 Disk performance test system and method based on Linux distributed systems

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115801633A (en) * 2023-02-08 2023-03-14 湖南国科亿存信息科技有限公司 SAN storage multi-client sharing concurrency performance test method and system
CN115801633B (en) * 2023-02-08 2023-04-28 湖南国科亿存信息科技有限公司 Multi-client sharing concurrency performance testing method and system for SAN storage

Also Published As

Publication number Publication date
CN112445660B (en) 2022-09-13

Similar Documents

Publication Publication Date Title
US9336340B1 (en) Evaluating management operations
EP2843537B1 (en) Method and systems for simulating a workload of a storage system
US7213113B2 (en) System and method for preparation of workload data for replaying in a data storage environment
US6260110B1 (en) Virtual tape system with variable size
CN102568522B (en) The method of testing of hard disk performance and device
CN101616174B (en) Method for optimizing system performance by dynamically tracking IO processing path of storage system
CN107092541A (en) A kind of method that utilization virtual machine carries out storage performance test
CN102789410A (en) Method for testing performance of storage system
US6751702B1 (en) Method for automated provisioning of central data storage devices using a data model
CN102915270B (en) Method for recording storage I/O (input/output) requirement and verifying pressure simulation
CN105808443A (en) Data migration method, apparatus and system
CN106534373A (en) IO performance test method and apparatus of SAN storage device
US9720818B2 (en) Scalable, distributed, fault-tolerant test framework
CN112445660B (en) Test method for performance balance problem of shared storage array
CN101883019B (en) Test method for verifying video application of storage server
CN110046069A (en) A kind of method, apparatus, equipment and the readable storage medium storing program for executing of storage performance test
US20220229768A1 (en) Method and Apparatus for Generating Simulated Test IO Operations
CN105607975A (en) Method, device and system for testing disk arrays
US20160036632A1 (en) Computer system
CN115248782B (en) Automatic testing method and device and computer equipment
US20070094393A1 (en) Intelligent logical unit provisioning
CN110825607A (en) Big data-based computer network test system
US20070088810A1 (en) Apparatus, system, and method for mapping a storage environment
CN112558858B (en) Method for acquiring access characteristics of super-large-scale application external memory based on dynamic IO perception
CN108572914B (en) Method and device for generating log for pressure test

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant