A method for testing IO performance of storage device

Abstract

The present invention discloses an IO performance test method for a storage device, and relates to the field of storage device performance test. The IO performance test method includes: obtaining historical storage data of a storage device, and establishing a test database based on the historical storage data; obtaining historical usage data of the storage device and dividing it into several test partitions in the storage device; extracting a test data set in the test database and marking target detection data; inputting the test data set into several test partitions in sequence, recording the read and write response time of the target detection data in each test partition, and obtaining partition test data; taking any two partition test data for difference comparison, and calculating the error time according to the difference comparison result; comparing the error time with a preset time difference threshold, and performing a grade assessment according to the comparison result. It evaluates the IO performance of the storage device by comparing the difference between each test partition, reduces the measurement error, and thus improves the accuracy and reliability of the test.

Description

Technical Field

[0001] The present invention relates to the technical field of storage device performance testing, and in particular to an IO performance testing method for a storage device. Background Art

[0002] With the development of the information age, storage technology is constantly updated and developed to meet the storage needs of increasingly abundant data. Storage devices with high performance, high capacity and high transmission efficiency are one of the main technological products.

[0003] IO performance (data reading and writing capability) is one of the important indicators for measuring storage devices. For storage devices that serve in important positions for a long time, it is necessary to test the IO performance of the storage devices regularly to ensure that the storage devices can meet the needs of data transmission. The current IO performance test of storage devices is to test the data reading and writing delay of storage devices through small data blocks. However, since storage devices are affected by interference factors such as network environment fluctuations and the response rate of the operating system when reading and writing data, there is a risk of large errors in the test delay indicator data, resulting in low measurement accuracy and insufficient reliability. Therefore, a measurement method that can reduce the measurement error of storage device delay indicators is needed. Summary of the invention

[0004] The purpose of the present invention is to provide an IO performance testing method for a storage device, which simulates the working environment of a storage device by setting different test partitions, tests the response time of each partition by target data detection, evaluates the IO performance of the storage device by comparing the difference between each test partition, and reduces the measurement error caused by interference factors by caching transmission, thereby improving the accuracy and reliability of the IO performance test of the storage device.

[0005] The present invention provides an IO performance testing method for a storage device, the IO performance testing method comprising:

[0006] S100: Acquire historical storage data of a storage device, and establish a test database based on the historical storage data;

[0007] S110: Acquire historical usage data of a storage device, and divide a plurality of test partitions in the storage device based on the historical usage data;

[0008] S120: extracting a test data set from the test database, randomly extracting a test data set from the test database, randomly selecting data from the test data set, and marking the selected data to obtain target data;

[0009] S130: inputting the test data set into a plurality of the test partitions in sequence, recording the read and write response time of the target detection data in each test partition, and obtaining partition test data;

[0010] The storage device caches the test data set for a preset time and then outputs it;

[0011] Acquire the reading time when the target data is input into the storage device, and acquire the writing time when the target data is output from the storage device;

[0012] Combining the read time, the write time and the preset time, obtaining the read and write response time of the target data;

[0013] The calculation formula for the read and write response time of the target data is:

[0014] T R =(t o -t i )-T s ;

[0015] Where: T R is the read and write response time of the target data, t o is the writing time of the target data, t i is the target data reading time, T s is the preset time;

[0016] S140: Take any two partition test data to perform difference comparison, and calculate the error time according to the difference comparison result;

[0017] S150: Compare the error time with a preset time difference threshold, and grade the IO performance of the storage device according to the comparison result.

[0018] In an optional embodiment, the acquiring historical usage data of the storage device and dividing the storage device into a plurality of test partitions based on the historical usage data comprises:

[0019] Extract the data storage capacity change and read / write frequency of the storage device according to the historical usage data of the storage device;

[0020] A plurality of test partitions are divided in the storage device according to the change of the data storage amount and the read-write frequency.

[0021] In an optional embodiment, dividing the storage device into a plurality of test partitions according to the change in the data storage amount and the read / write frequency includes:

[0022] The data storage amounts of any two test partitions are different, and / or the read and write frequencies of any two test partitions are different.

[0023] In an optional embodiment, randomly selecting data from the test data set and marking the selected data to obtain target data includes:

[0024] Insert the first marker before the selected data and insert the second marker after the selected data.

[0025] In an optional embodiment, obtaining the read time of inputting the target data into the storage device and obtaining the write time of outputting the target data from the storage device comprises:

[0026] The first marker and the second marker of the target test data are read, and the reading time and the writing time of the target test data are marked according to the reading order of the first marker and the second marker.

[0027] In an optional embodiment, taking any two partition test data for difference comparison, and calculating the error time according to the difference comparison result includes:

[0028] The error time data set is obtained by performing pairwise difference calculation on several partition test data through exhaustive method.

[0029] In an optional embodiment, comparing the error time with a preset time difference threshold and rating the IO performance of the storage device according to the comparison result includes:

[0030] Compare the error time in the error time data set with the preset time difference threshold in turn to determine whether the error time is greater than the time difference threshold, if so, it is a failure, if not, it is a pass;

[0031] An error time pass rate of the error time data set is calculated, and an IO performance level of the storage device is evaluated according to the error time pass rate.

[0032] The present invention also provides an IO performance testing device for a storage device, the testing device is used to execute the IO performance testing method, and the testing device comprises:

[0033] Database establishment module: obtains historical storage data of the storage device and establishes a test database based on the historical storage data;

[0034] A test partition division module: obtaining historical usage data of a storage device, and dividing a plurality of test partitions in the storage device based on the historical usage data;

[0035] Test data module: extracting a test data set from the test database, randomly extracting a test data set from the test database, randomly selecting data from the test data set, and marking the selected data to obtain target data;

[0036] Partition test module: input the test data set into a number of test partitions in sequence, record the read and write response time of the target detection data in each test partition, and obtain partition test data;

[0037] The storage device caches the test data set for a preset time and then outputs it;

[0038] Acquire the reading time when the target data is input into the storage device, and acquire the writing time when the target data is output from the storage device;

[0039] Combining the read time, the write time and the preset time, obtaining the read and write response time of the target data;

[0040] The calculation formula for the read and write response time of the target data is:

[0041] T R =(t o -t i )-T s ;

[0042] Where: T R is the read and write response time of the target data, t o is the writing time of the target data, t i is the target data reading time, T s is the preset time;

[0043] Difference comparison module: take any two partition test data for difference comparison, and calculate the error time according to the difference comparison result;

[0044] Rating module: compares the error time with the preset time difference threshold, and rates the IO performance of the storage device according to the comparison result.

[0045] The beneficial effects of the present invention are as follows: different working environments of a storage device are simulated by partitioning, target detection data is tracked and recorded under different test partitions, thereby determining the read and write time difference between different test partitions, and the IO performance of the storage device is evaluated by analyzing the read and write response time difference of different test partitions, and the measurement error caused by interference factors is reduced by caching transmission, thereby improving the reliability and accuracy of the IO performance evaluation of the storage device. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flowchart of an IO performance testing method for a storage device according to an embodiment of the present invention;

[0047] Figure 2 It is a schematic diagram of the test partition construction method according to an embodiment of the present invention;

[0048] Figure 3 It is a schematic diagram of a flow chart of a method for analyzing read and write response time of a test partition according to an embodiment of the present invention;

[0049] Figure 4 It is a schematic diagram of an IO performance testing device of a storage device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Embodiment 1:

[0052] The embodiment of the present invention provides a method for testing the IO performance of a storage device. Figure 1 A schematic flow chart of an IO performance test method for a storage device of the present invention is shown, and the IO performance test method comprises:

[0053] S100: Acquire historical storage data of a storage device, and establish a test database based on the historical storage data.

[0054] Specifically, historical storage data of the storage device is obtained, and the historical storage data of the storage device is aggregated to form a database. The main working purpose of the storage device and the type of stored data can be determined through the historical storage data. The operator can supplement and expand the database according to the main working data environment of the storage device.

[0055] Furthermore, by using the historical storage data of the storage device to construct a test database for the storage device, the IO performance test of the storage device can be tailored to the actual use requirements of the storage device, thereby improving the accuracy of the IO performance test of the storage device.

[0056] S110: Acquire historical usage data of a storage device, and divide the storage device into a plurality of test partitions based on the historical usage data.

[0057] Specifically, the historical usage data includes the high-frequency usage time of the storage device. Based on the high-frequency usage time of the storage device, relevant parameters such as the network environment and processing system status during the time period are obtained, and a test environment is constructed based on the relevant parameters, so that the storage device can test its IO capabilities under the actual working environment.

[0058] Specifically, Figure 2A schematic flow chart of a test partition construction method according to an embodiment of the present invention is shown, wherein obtaining historical usage data of a storage device and dividing a plurality of test partitions in the storage device based on the historical usage data comprises:

[0059] S111: Extracting data storage capacity changes and read / write frequencies of the storage device according to historical usage data of the storage device.

[0060] The data storage capacity change of the storage device is extracted from the historical usage data of the storage device, that is, the change of the overall data storage capacity of the storage device in different working time periods. Based on the change of the storage capacity, the storage capacity of the storage device when it is working can be determined, and the data of each storage capacity of the storage device can be extracted from the storage capacity change.

[0061] The read and write rates of the storage device are obtained from the historical usage data of the storage device. During actual use, the system will adjust the read and write rates of the storage device according to the data transmission requirements, so that the read and write rates of the storage device remain within an appropriate range, meet the data transmission requirements, and reduce the energy efficiency loss of the storage device.

[0062] In this embodiment, the maximum read rate of the storage device is 3600MB / s, and the maximum write speed of the storage device is 3200MB / s. The range of the read and write rates of the storage device in daily operation is determined based on the historical usage of the storage device.

[0063] S112: Divide the storage device into a plurality of test partitions according to the change in the data storage amount and the read / write frequency.

[0064] Specifically, the data storage capacity of any two test partitions is different, and / or the read and write frequencies of any two test partitions are different. According to the historical usage data of the storage device, the test partitions of the storage device are sorted out as follows:

[0065] Table 1: Test partition arrangement table

[0066]

[0067] Specifically, the total capacity of each test partition is the same, and the remaining capacity of each test partition is adjusted according to the historical usage data of the storage device. By setting multiple groups of test partitions with different capacities or different read and write rates, the test effect of the storage device is more accurate.

[0068] Furthermore, by dividing the storage device into a plurality of test partitions, the plurality of test partitions can reflect the actual working state of the storage device, thereby improving the accuracy of the IO capability test of the storage device.

[0069] Specifically, data is sorted on the storage device to ensure that the storage device has sufficient disk space for IO performance testing, test partitions are divided according to the disk space of the storage device, part of the test data is extracted from the test database, and input into different test partitions according to the requirements of the test partitions, thereby adjusting the capacity of each test partition.

[0070] S120: extracting a test data set from the test database, and marking target detection data in the test data set.

[0071] Specifically, a test data set is extracted from the test database, and a data set of a preset capacity is obtained from the test database by random crawling. The preset capacity can be adjusted according to the size of the test partition of the storage device.

[0072] Furthermore, data is randomly selected from the test data set, and the selected data is marked to obtain target data. By randomly selecting single data from the test data set and marking it as target data, after the test data set is input, the accuracy of the read and write performance test of the test partition data is improved by tracking the read and write status of the target data.

[0073] Specifically, a first marker is inserted before the selected data, and a second marker is inserted after the selected data. When the storage device reads or writes the first marker or the second marker, the storage device feeds back an information packet to the test system, and the test system can obtain the read status or write status of the target data according to the information packet fed back by the storage device.

[0074] Furthermore, the target data is marked by adding a marker to avoid confusion between the target data and the gas data of the test data set, thereby improving the accuracy of the test.

[0075] Specifically, the target data can be marked by a special symbol marking method, that is, the first marker and the second marker are special symbols, and the character database used by the special symbols is different from the character database used by the test data set, thereby avoiding data duplication between the first marker and the second marker and the test data set, thereby avoiding detection errors.

[0076] S130: Input the test data set into a plurality of the test partitions in sequence, record the read and write response time of the target detection data in each test partition, and obtain partition test data.

[0077] Specifically, Figure 3 A schematic flow chart of a method for analyzing the read and write response time of a test partition according to an embodiment of the present invention is shown. The method comprises:

[0078] S131: The storage device caches the test data set for a preset time and then outputs it.

[0079] Specifically, by setting the preset time for data caching, the storage device has sufficient time for data caching between reading data and writing data, thereby reducing the impact of network fluctuations on the read and write performance of the storage device and improving the accuracy of the storage device IO performance test.

[0080] S132: Obtain the reading time of the target data input into the storage device, and obtain the writing time of the target data output from the storage device;

[0081] Specifically, when the storage device reads the test data set, it will first read the first marker and feedback the information package to the test system, and the test system obtains the information that the first marker is read based on the content of the information package; similarly, when the storage device reads the second marker, the storage device feedback the information package to the test system, and the test system obtains the information that the second marker is read.

[0082] Furthermore, the test system is provided with a reading order and a reading time interval of the first marker and the second marker. After the first marker is read and the reading information of the second marker is obtained within the reading time interval, the test system can determine that the target data is in a read state, thereby determining the reading time of the target data.

[0083] Furthermore, when writing data, the storage device writes in the reverse direction of the reading order, that is, the storage device preferentially writes the last read data and completes the writing operation of all data in sequence. After the storage device writes the second marker, the storage device feeds back the information packet to the test system. After the storage device writes the first marker, the storage device feeds back the information packet to the test system. The test system determines the writing time of the target data according to the triggering order and triggering interval time of the first marker and the second marker.

[0084] S133: Combining the read time, the write time and the preset time, obtain the read and write response time of the target data.

[0085] Specifically, the writing time of the target data is subtracted from the reading time of the target data to obtain the total reading and writing time of the target data, and the reading and writing response time of the target data is subtracted from the total reading and writing time.

[0086] Furthermore, the read and write response time can reflect the data reading and writing capability of the test partition for the test data set.

[0087] Specifically, the calculation formula for the read and write response time of the target data is:

[0088] T R =(t o -t i )-T s ;

[0089] Where: T R is the read and write response time of the target data, t o is the writing time of the target data, t i is the target data reading time, T s is the preset time.

[0090] S140: Take any two partition test data to perform difference comparison, and calculate the error time according to the difference comparison result.

[0091] Specifically, the error time data set is obtained by performing pairwise difference calculations on several partitioned test data through an exhaustive method. That is, one of the test partitions is taken, and the test data of the test partition is calculated difference with the test data of other test partitions in turn. After completing the difference calculation of one test partition, the difference comparison is repeated for the remaining test partitions until all test partitions complete the pairwise difference calculation, and the difference calculation results are summarized to obtain the error time data set.

[0092] Furthermore, the error time data of the error time data set reflects the difference in data reading and writing performance of the storage device under different working environments.

[0093] S150: Compare the error time with a preset time difference threshold, and grade the IO performance of the storage device according to the comparison result.

[0094] The error time in the error time data set is compared with the preset time difference threshold in turn to determine whether the error time is greater than the time difference threshold. If so, it fails, otherwise, it passes.

[0095] Specifically, the error time data of the error time data set are compared with a preset time difference threshold in turn. The time difference threshold is set according to the factory debugging data of the storage device. By comparing and analyzing the error time with the time difference threshold, it is determined whether the error time of the storage device in different working environments meets the error requirements set by the factory for the storage device.

[0096] An error time pass rate of the error time data set is calculated, and an IO performance level of the storage device is evaluated according to the error time pass rate.

[0097] Specifically, by calculating the error time pass rate of the error time data set, if the pass rate is above 80%, the IO performance level of the storage device is excellent; if the pass rate is between 60% and 80%, the IO performance level of the storage device is good; if the pass rate is below 60%, the IO performance level of the storage device is unsatisfactory, that is, the storage device cannot meet the needs of daily storage work, and the storage device needs to be maintained or replaced.

[0098] Specifically, the IO performance test of the storage device also includes a throughput test on the test partition, that is, recording the data read and write volume and the read and write rate of each test partition in real time, calculating the data throughput of each test partition, summarizing the data throughput of each test partition and calculating the average throughput of the test partition, and judging whether the storage device meets the requirements of the working environment according to the average throughput of the storage device and the working environment of the storage device.

[0099] The embodiment of the present invention provides an IO performance testing method for a storage device, which simulates different working environments of a storage device by partitioning, tracks and records target detection data in different test partitions, thereby determining the read and write time difference between different test partitions, and evaluates the IO performance of the storage device by analyzing the read and write response time difference of different test partitions. The measurement error caused by interference factors is reduced by caching transmission, thereby improving the reliability and accuracy of the IO performance evaluation of the storage device.

[0100] Embodiment 2:

[0101] Specifically, Figure 4 A schematic diagram of an IO performance testing device for a storage device in an embodiment of the present invention is shown. The testing device is used to perform the IO performance testing method. The testing device includes:

[0102] Database establishment module 10: acquiring historical storage data of a storage device, and establishing a test database based on the historical storage data;

[0103] The test partition division module 20 is configured to obtain historical usage data of a storage device and divide the storage device into a plurality of test partitions based on the historical usage data.

[0104] Test data module 30: extract a test data set from the test database, randomly extract a test data set from the test database, randomly select data from the test data set, and mark the selected data to obtain target data;

[0105] Partition test module 40: input the test data set into a plurality of test partitions in sequence, record the read and write response time of the target detection data in each test partition, and obtain partition test data;

[0106] The storage device caches the test data set for a preset time and then outputs it;

[0107] Acquire the reading time when the target data is input into the storage device, and acquire the writing time when the target data is output from the storage device;

[0108] Combining the read time, the write time and the preset time, obtaining the read and write response time of the target data;

[0109] The calculation formula for the read and write response time of the target data is:

[0110] T R =(t o -t i )-T s ;

[0111] Where: T R is the read and write response time of the target data, t o is the writing time of the target data, t i is the target data reading time, T s is the preset time;

[0112] Difference comparison module 50: taking any two partition test data for difference comparison, and calculating the error time according to the difference comparison result;

[0113] The rating module 60 compares the error time with a preset time difference threshold, and rates the IO performance of the storage device according to the comparison result.

[0114] An embodiment of the present invention provides an IO performance testing device for a storage device. The testing device simulates different working environments of a storage device by partitioning, tracks and records target detection data in different test partitions, thereby determining the read and write time differences between different test partitions, and evaluates the IO performance of the storage device by analyzing the read and write response time differences of different test partitions. The measurement error caused by interference factors is reduced by caching transmission, thereby improving the reliability and accuracy of the IO performance evaluation of the storage device.

[0115] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the software module can be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the technical field.

[0116] The above is a detailed introduction to the IO performance testing method of a storage device provided by an embodiment of the present invention. In this article, specific examples are used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this field, without departing from the principle of the present invention, several improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A method for testing the IO performance of a storage device, characterized in that: The IO performance testing method comprises: S100: Acquire historical storage data of a storage device, and establish a test database based on the historical storage data; S110: Acquire historical usage data of a storage device, and divide a plurality of test partitions in the storage device based on the historical usage data; S120: extracting a test data set from the test database, randomly extracting a test data set from the test database, randomly selecting data from the test data set, and marking the selected data to obtain target data; S130: inputting the test data set into a plurality of the test partitions in sequence, recording the read and write response time of the target data in each test partition, and obtaining partition test data; The storage device caches the test data set for a preset time and then outputs it; Acquire the reading time when the target data is input into the storage device, and acquire the writing time when the target data is output from the storage device; Combining the read time, the write time and the preset time, obtaining the read and write response time of the target data; The calculation formula for the read and write response time of the target data is: ； in: is the read and write response time of the target data, is the writing time of the target data, is the time when the target data is read, is the preset time; S140: Take any two partition test data to perform difference comparison, and calculate the error time according to the difference comparison result; S150: Compare the error time with a preset time difference threshold, and grade the IO performance of the storage device according to the comparison result.

2. The IO performance testing method of a storage device according to claim 1, characterized in that: The acquiring of historical usage data of the storage device and dividing the storage device into a plurality of test partitions based on the historical usage data comprises: Extract the data storage capacity change and read / write frequency of the storage device according to the historical usage data of the storage device; A plurality of test partitions are divided in the storage device according to the change of the data storage amount and the read-write frequency.

3. The IO performance testing method of a storage device according to claim 2, characterized in that: The dividing of the storage device into a plurality of test partitions according to the change of the data storage amount and the read / write frequency comprises: The data storage amounts of any two test partitions are different, and / or the read and write frequencies of any two test partitions are different.

4. The IO performance testing method of a storage device according to claim 1, characterized in that: The randomly selecting data from the test data set and marking the selected data to obtain the target data includes: Insert the first marker before the selected data and insert the second marker after the selected data.

5. The IO performance testing method of a storage device according to claim 4, characterized in that: The obtaining of the read time of the target data input into the storage device and the obtaining of the write time of the target data output by the storage device include: The first marker and the second marker of the target data are read, and the reading time and the writing time of the target data are marked according to the reading order of the first marker and the second marker.

6. The IO performance testing method of a storage device according to claim 1, characterized in that: The method of taking any two partition test data for difference comparison and calculating the error time according to the difference comparison result includes: The error time data set is obtained by performing pairwise difference calculation on several partition test data through exhaustive method.

7. The IO performance testing method of a storage device according to claim 6, characterized in that: The step of comparing the error time with a preset time difference threshold and evaluating the IO performance of the storage device according to the comparison result includes: Compare the error time in the error time data set with the preset time difference threshold in turn to determine whether the error time is greater than the time difference threshold, if so, it is a failure, if not, it is a pass; An error time pass rate of the error time data set is calculated, and an IO performance level of the storage device is evaluated according to the error time pass rate.

8. An IO performance testing device for a storage device, characterized in that: The testing device is used to perform the IO performance testing method according to any one of claims 1 to 7, and the testing device comprises: Database establishment module: obtains historical storage data of the storage device and establishes a test database based on the historical storage data; A test partition division module: obtaining historical usage data of a storage device, and dividing a plurality of test partitions in the storage device based on the historical usage data; Test data module: extracting a test data set from the test database, randomly extracting a test data set from the test database, randomly selecting data from the test data set, and marking the selected data to obtain target data; Partition test module: input the test data set into a number of test partitions in sequence, record the read and write response time of the target data in each test partition, and obtain partition test data; The storage device caches the test data set for a preset time and then outputs it; Acquire the reading time when the target data is input into the storage device, and acquire the writing time when the target data is output from the storage device; Combining the read time, the write time and the preset time, obtaining the read and write response time of the target data; The calculation formula for the read and write response time of the target data is: ； in: is the read and write response time of the target data, is the writing time of the target data, is the time when the target data is read, is the preset time; Difference comparison module: take any two partition test data for difference comparison, and calculate the error time according to the difference comparison result; Rating module: compares the error time with the preset time difference threshold, and rates the IO performance of the storage device according to the comparison result.

Images