Server disk performance analysis optimization method and system based on declarative type management identification

Abstract

The invention relates to the technical field of server disk performance analysis and optimization, in particular to a server disk performance analysis and optimization method and system based on declarative type management identification, comprising the following steps: defining a declarative type management identification model, scanning and collecting server disk information, and intelligently matching the declarative type management identification model. Executing a disk performance analysis test, processing a test result according to a strategy, and performing evaluation to generate an optimal optimization parameter; the method has the beneficial effects that through a server disk performance analysis optimization technology based on declarative type management identification, a declarative type management identification model is defined, server disk information is scanned and collected, the declarative type management identification model is intelligently matched, a disk performance analysis test is executed, and a test result is processed according to a strategy; and the optimal optimization parameters are evaluated and generated, so that the server disk performance analysis optimization method based on declarative type management identification is quickly and automatically realized.

Description

Technical Field

[0001] The present invention relates to the technical field of server disk performance analysis and optimization, and in particular to a server disk performance analysis and optimization method and system based on declarative type management and identification. Background Art

[0002] Different types of server products have different disk types and configuration combinations, and their disk performance is also different. It is impossible to effectively measure and analyze the performance of different types of disks and configuration combinations, and it is also impossible to select the optimal parameter optimization combination in a targeted manner. Summary of the invention

[0003] The purpose of the present invention is to provide a server disk performance analysis and optimization method and system based on declarative type management and identification to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a server disk performance analysis and optimization method based on declarative type management and identification, comprising the following steps:

[0005] Define a declarative type management identification model, which includes a large category of disk combinations, specifically storage node disk combinations and management node disk combinations. Each disk combination involves the number of disks, disk type, disk size, RAID card mode, disk scheduling algorithm, disk access depth parameter, and disk NR request parameter. This definition process is implemented by creating a new model, selecting the node disk combination type, RAID card mode, disk type, setting the disk size, the number of disks of the same disk type, disk scheduling algorithm, disk access depth parameter, disk NR request parameter, disk test combination, average number of disk tests, and saving and naming the current model.

[0006] Preferably, the method further includes the step of scanning and collecting server disk information:

[0007] Install the RAID card management tool in the operating system through an automated script, execute the disk query command, filter the information to obtain the disk type, disk size, number of disks, RAID card mode, disk model, disk manufacturer, and disk firmware version; specifically, download the automated script and execute the installation tool, execute the commands in sequence to query the basic configuration of the RAID card, disk type, disk size, number of disks, disk model, manufacturer, and firmware version, and finally format and save the query results to achieve the scanning and collection process.

[0008] Preferably, the step of intelligently matching the declarative type management recognition model is also included:

[0009] Load multiple groups of defined declarative type management recognition models, and based on the currently collected server disk information, match them with the models through a matching algorithm to obtain the performance test combination that should be carried out under the current server disk configuration; specifically, by loading multiple groups of defined models, formatting the server disk information, and the matching algorithm program, match the RAID card mode, disk type in sequence, and synchronously match the disk quantity and disk size. Determine the test scenario according to a specific intelligent matching algorithm logic, and finally save the matching disk test combination result to implement this matching process; among them, the intelligent matching algorithm logic includes: when the disk quantity of the same disk type is equal to 2 and the disk size is less than 600G, determine this type as the system disk and add the test scenario of forming RAID1; when the disk size is equal to 960G, determine it as the acceleration disk and add the test scenarios of single-disk and 2-disk concurrent tests; when the disk quantity of the same disk type is greater than 4 and less than 7 and the RAID card mode is Raid-mode, determine it as the control node disk and add the test scenario of forming RAID5; when the disk quantity of the same disk type is greater than or equal to 10 and the RAID card mode is Jbod-mode, determine it as the storage node disk and add the test scenarios of single-disk and multi-disk concurrent tests.

[0010] Preferably, it further includes the step of performing disk performance analysis tests:

[0011] According to the matching disk test combination result, perform disk performance analysis tests respectively, and perform disk performance analysis tests under different disk scheduling algorithms, different disk access depth parameters, and different disk NR request parameters, and format and save the disk performance analysis test data; specifically, by loading the matching disk test combination result, cycling the disk scheduling algorithm, disk access depth parameter, and disk NR request parameter configuration, perform disk performance analysis tests, test the disk IOPS, bandwidth, and latency metrics, and the read and write operation types are random read, random write, sequential read, sequential write, and mixed read and write. According to the set disk test average number value, cycle the test process, and finally format and save the disk performance analysis test result data to implement this test process.

[0012] Preferably, it further includes the step of processing the test results according to the policy and evaluating and generating the optimal optimization parameters:

[0013] Through the disk data processing strategy, linearly analyze and compare the test results, and save multiple sets of linear analysis combined data; through weighted ratio calculation and horizontal comparison of multiple sets of linear analysis combined data, obtain the optimal linear analysis combined data, and the optimization parameters corresponding to this set of data are the optimal optimization parameter combination generated by the evaluation; specifically, load the disk data processing strategy, disk test index reference values, disk performance analysis test result data, linearly analyze and compare the test result data, and save multiple sets of linear analysis combined data; then load multiple sets of linear analysis combined data, perform weighted ratio calculation, horizontally compare the data, evaluate the horizontal comparison data results, and generate the optimal optimization parameter combination to implement this processing and evaluation process.

[0014] A system for optimizing the analysis of server disk performance based on declarative type management recognition, including a model definition module for defining a declarative type management recognition model. The model includes major disk combinations, specifically storage node disk combinations and management node disk combinations. Each disk combination involves the number of disks, disk type, disk size, RAID card mode, disk scheduling algorithm, disk access depth parameter, and disk NR request parameter; this module realizes model definition by creating a new model, sequentially selecting the node disk combination type, RAID card mode, disk type, setting the disk size, the number of disks of the same disk type, disk scheduling algorithm, disk access depth parameter, disk NR request parameter, disk test combination, and average number of disk tests, and saving and naming the current model.

[0015] Preferably, it further includes an information scanning module for scanning and collecting server disk information. Install the RAID card management tool in the operating system through an automated script, execute the disk query command, and filter the information to obtain the disk type, disk size, number of disks, RAID card mode, disk model, disk manufacturer, and disk firmware version; specifically, realize information scanning by downloading the automated script and executing the installation tool, sequentially executing commands to query the basic configuration of the RAID card, disk type, disk size, number of disks, disk model, manufacturer, and firmware version, and finally formatting and saving the query results.

[0016] Preferably, it further includes a model matching module for intelligently matching a declarative type management recognition model, loading multiple groups of defined declarative type management recognition models, and matching them with the models through a matching algorithm based on the currently collected server disk information to obtain the performance test combination to be carried out under the current server disk configuration; specifically, it is achieved by loading multiple groups of defined models, formatting the server disk information, and a matching algorithm program, successively matching the RAID card mode, disk type and synchronously matching the disk quantity and disk size, determining the test scenario according to a specific intelligent matching algorithm logic, and finally saving the matching disk test combination result to implement model matching; among them, the intelligent matching algorithm logic includes: when the quantity of disks of the same disk type is equal to 2 and the disk size is less than 600G, determining this type as the system disk and adding a test scenario for forming RAID1; when the disk size is equal to 960G, determining it as the acceleration disk and adding a single-disk and two-disk concurrent test scenario; when the quantity of disks of the same disk type is greater than 4 and less than 7 and the RAID card mode is Raid-mode, determining it as the disk of the control node and adding a test scenario for forming RAID5; when the quantity of disks of the same disk type is greater than or equal to 10 and the RAID card mode is Jbod-mode, determining it as the disk of the storage node and adding a single-disk and multi-disk concurrent test scenario.

[0017] Preferably, it further includes a test execution module for performing disk performance analysis tests, and performing disk performance analysis tests under different disk scheduling algorithms, different disk access depth parameters, and different disk NR request parameters according to the matching disk test combination results, and formatting and saving the disk performance analysis test data; specifically, it is achieved by loading the matching disk test combination results, cycling the disk scheduling algorithm, disk access depth parameter, and disk NR request parameter configurations, performing disk performance analysis tests, testing disk IOPS, bandwidth, and latency metrics, and the read and write operation types are random read, random write, sequential read, sequential write, and mixed read and write, and cycling the test process according to the set average number of disk test times value, and finally formatting and saving the disk performance analysis test result data to implement test execution.

[0018] Preferably, it further includes a result processing and optimal parameter generation module, which is used to process the test results according to the policy and evaluate and generate the optimal optimization parameters. Through the disk data processing policy, linear analysis and comparison to process the test results, and save multiple linear analysis combination data; through the weighted ratio calculation of multiple groups of linear analysis combination data, horizontal comparison is performed to obtain the optimal linear analysis combination data, and the optimization parameters corresponding to this group of data are the optimal optimization parameter combination evaluated and generated; specifically, by loading the disk data processing policy, disk test index reference values, disk performance analysis test result data, linearly analyzing and comparing the test result data, and saving multiple groups of linear analysis combination data; then loading multiple groups of linear analysis combination data, performing weighted ratio calculation, horizontal comparison of data, evaluating the horizontal comparison data results, and generating the optimal optimization parameter combination to achieve result processing and optimal parameter generation.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] The server disk performance analysis and optimization method and system based on declarative type management recognition proposed by the present invention, through the server disk performance analysis and optimization technology based on declarative type management recognition, defines a declarative type management recognition model, scans and collects server disk information, intelligently matches the declarative type management recognition model, executes disk performance analysis tests, processes the test results according to the policy, evaluates and generates the optimal optimization parameters, and quickly and automatically realizes the server disk performance analysis and optimization method based on declarative type management recognition. Brief Description of the Drawings

[0021] Figure 1 It is a flowchart of the method of the present invention. Detailed Embodiments

[0028] In order to clearly and completely describe the objectives, technical solutions of the present invention, and make the advantages more clearly understood, the following further details the embodiments of the present invention with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are some embodiments of the present invention, rather than all embodiments, and are only used to explain the embodiments of the present invention, not to limit the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0029] Embodiment 1, the present invention provides a technical solution: a server disk performance analysis and optimization method based on declarative type management recognition, including the following steps:

[0030] 1: Define a declarative type management recognition model

[0031] The declarative format type management recognition model includes major categories of disk combinations, storage node disk combinations, and management node disk combinations. Each disk combination involves the number of disks, disk types, disk sizes, RAID card modes, disk scheduling algorithms, disk access depth parameters, disk NR request parameters, etc.

[0032] Define the implementation process of the declarative type management recognition model:

[0033] 1) Create a declarative type management recognition model;

[0034] 2) Select the node disk combination type (storage, management);

[0035] 3) Select the RAID card mode (JBOD mode, RAID mode);

[0036] 4) Select the disk type (SAS, SATA, SSD);

[0037] 5) Set the disk size;

[0038] 6) Set the number of disks of the same disk type;

[0039] 7) Set the disk scheduling algorithm (mq - deadline kyber bfq none);

[0040] 8) Set the disk access depth parameters (32 128 256 512 1024);

[0041] 9) Set the disk NR request parameters (256 1024 2048 5089 8161);

[0042] 10) Set the disk test combination, single disk, multi - disk concurrency, multi - disk group RAID5;

[0043] 11) Set the average number of disk tests;

[0044] 12) Save and name the current model.

[0045] 2: Scan and collect server disk information

[0046] Install the RAID card management tool in the operating system through an automated script, execute the disk query command, and filter the information to obtain information such as disk type, disk size, number of disks, RAID card mode, disk model, disk manufacturer, disk firmware version, etc.

[0047] The implementation process of scanning and collecting server disk information:

[0048] 1) Download the automated script and execute the installation tool;

[0049] 2), Execute the command to query the basic configuration of the RAID card;

[0050] 3), Execute the command to query the disk type;

[0051] 4), Execute the command to query the disk size;

[0052] 5), Execute the command to query the number of disks;

[0053] 6), Execute the command to query the disk model, manufacturer, and firmware version;

[0054] 7), Format and save the query results.

[0055] 3: Intelligent matching declarative type management recognition model

[0056] Load multiple groups of defined declarative type management recognition models. Based on the currently collected disk information of the server, match with the models through a matching algorithm to obtain the performance test combinations that should be carried out under the current server disk configuration.

[0057] Implementation process of intelligent matching declarative type management recognition model:

[0058] 1), Load multiple groups of defined declarative type management recognition models;

[0059] 2), Load the server disk information in a format;

[0060] 3), Load the matching algorithm program;

[0061] 4), Match the RAID card mode;

[0062] 5), Match the disk type, synchronously match the number of disks and the disk size. The intelligent matching algorithm logic is as follows:

[0063] a), If the number of disks of the same disk type is equal to 2 and the disk size is less than 600G, determine this type as the system disk and add the test scenario of forming RAID1; if the disk size is equal to 960G, determine it as the acceleration disk and add the test scenarios of single disk and 2-disk concurrency.

[0064] b), If the number of disks of the same disk type is greater than 4 and less than 7, and the RAID card mode is Raid-mode, determine it as the disk of the control node and add the test scenario of forming RAID5.

[0065] c), If the number of disks of the same disk type is greater than or equal to 10, and the RAID card mode is Jbod-mode, determine it as the disk of the storage node and add the test scenarios of single disk and multi-disk concurrency.

[0066] 6), Save the results of the matched disk test combinations.

[0067] 4: Execute disk performance analysis test

[0068] According to the matching disk test combination results, execute disk performance analysis tests respectively, perform disk performance analysis tests under different disk scheduling algorithms, perform disk performance analysis tests under different disk access depth parameters, perform disk performance analysis tests under different disk NR request parameters, and format and save the disk performance analysis test data.

[0069] Implementation process of executing disk performance analysis test:

[0070] 1). Load the matching disk test combination results;

[0071] 2). Loop through the disk scheduling algorithm parameter configurations, execute disk performance analysis tests, test disk IOPS, bandwidth, and latency metrics, and the read / write operation types are random read, random write, sequential read, sequential write, and mixed read / write;

[0072] 3). Loop through the disk access depth parameter configurations, execute disk performance analysis tests, test disk IOPS, bandwidth, and latency metrics, and the read / write operation types are random read, random write, sequential read, sequential write, and mixed read / write;

[0073] 4). Loop through the disk NR request parameter configurations, execute disk performance analysis tests, test disk IOPS, bandwidth, and latency metrics, and the read / write operation types are random read, random write, sequential read, sequential write, and mixed read / write;

[0074] 5). According to the set disk test average number value, loop through steps 2, 3, and 4;

[0075] 6). Format and save the disk performance analysis test result data.

[0076] 5: Process test results according to policies

[0077] Through the disk data processing policy, linearly analyze and compare the processed test results, and save multiple linear analysis combination data.

[0078] Implementation process of processing test results according to policies:

[0079] 1). Load the disk data processing policy;

[0080] a). For SSD type disks, amplify the access depth parameter 256 by the weighted ratio, and amplify the NR request parameter 5089 by the weighted ratio;

[0081] b). For SAS type disks, amplify the access depth parameter 128 by the weighted ratio, and amplify the NR request parameter 256 by the weighted ratio;

[0082] c), SATA type disks, the access depth parameter 32 is amplified by the weighted ratio, and the NR request parameter 256 is amplified by the weighted ratio;

[0083] 2), Load the reference values of disk test metrics;

[0084] 3), Load the disk performance analysis test result data;

[0085] 4), Linear analysis and comparison of test result data;

[0086] 5), Save multiple sets of linear analysis combined data.

[0087] 6: Evaluate and generate the optimal optimization parameters

[0088] Through the weighted ratio calculation of multiple sets of linear analysis combined data, horizontal comparison is made to obtain the optimal linear analysis combined data, and the optimization parameters corresponding to this set of data are the optimal optimization parameter combination generated by the evaluation.

[0089] The implementation process of evaluating and generating the optimal optimization parameters:

[0090] 1), Load multiple sets of linear analysis combined data;

[0091] 2), Weighted ratio calculation and horizontal comparison of data;

[0092] 3), Evaluate the results of horizontal comparison data;

[0093] 4), Generate the optimal optimization parameter combination.

[0094] In the second embodiment, on the basis of the first embodiment, a system for optimizing the server disk performance analysis based on declarative type management recognition is proposed, including a model definition module for defining a declarative type management recognition model. The model includes large categories of disk combinations, specifically storage node disk combinations and management node disk combinations. Each disk combination involves the number of disks, disk type, disk size, RAID card mode, disk scheduling algorithm, disk access depth parameter, and disk NR request parameter. This module realizes model definition by creating a new model, sequentially selecting the node disk combination type, RAID card mode, disk type, setting the disk size, the number of disks of the same disk type, disk scheduling algorithm, disk access depth parameter, disk NR request parameter, disk test combination, and the average number of disk tests, and saving and naming the current model.

[0095] It also includes an information scanning module, which is used to scan and collect server disk information, install RAID card management tools in the operating system through an automated script, execute disk query commands, and filter information to obtain disk type, disk size, disk quantity, RAID card mode, disk model, disk manufacturer, and disk firmware version; specifically, it realizes information scanning by downloading the automated script and executing the installation tool, sequentially executing commands to query the basic configuration of the RAID card, disk type, disk size, disk quantity, disk model, manufacturer, and firmware version, and finally formatting and saving the query results.

[0096] It also includes a model matching module, which is used to intelligently match the declarative type management recognition model, load multiple groups of defined declarative type management recognition models, and based on the currently collected server disk information, match with the model through a matching algorithm to obtain the performance test combination that should be carried out under the current server disk configuration; specifically, it realizes model matching by loading multiple groups of defined models, formatting server disk information, and the matching algorithm program, sequentially matching the RAID card mode, disk type, and synchronously matching the disk quantity and disk size, determining the test scenario according to the specific intelligent matching algorithm logic, and finally saving the matching disk test combination results; among them, the intelligent matching algorithm logic includes: when the disk quantity of the same disk type is equal to 2 and the disk size is less than 600G, it is determined that this type is the system disk, and the RAID1 test scenario is added; when the disk size is equal to 960G, it is determined as the acceleration disk, and the single-disk and 2-disk concurrent test scenarios are added; when the disk quantity of the same disk type is greater than 4 and less than 7 and the RAID card mode is Raid-mode, it is determined as the control node disk, and the RAID5 test scenario is added; when the disk quantity of the same disk type is greater than or equal to 10 and the RAID card mode is Jbod-mode, it is determined as the storage node disk, and the single-disk and multi-disk concurrent test scenarios are added.

[0097] It also includes a test execution module, which is used to execute disk performance analysis tests, and based on the matching disk test combination results, perform disk performance analysis tests under different disk scheduling algorithms, different disk access depth parameters, and different disk NR request parameters, and format and save the disk performance analysis test data; specifically, it realizes test execution by loading the matching disk test combination results, cycling the disk scheduling algorithm, disk access depth parameter, and disk NR request parameter configurations, executing disk performance analysis tests, testing disk IOPS, bandwidth, and latency metrics, and the read and write operation types are random read, random write, sequential read, sequential write, and mixed read and write, and cycling the test process according to the set average number of disk test times value, and finally formatting and saving the disk performance analysis test result data.

[0098] It further includes a result processing and optimal parameter generation module, which is used to process the test results according to the strategy and evaluate and generate the optimal optimization parameters. Through the disk data processing strategy, linear analysis and comparison are performed to process the test results, and multiple linear analysis combination data are saved; through the weighted ratio calculation and horizontal comparison of multiple groups of linear analysis combination data, the optimal linear analysis combination data is obtained, and the optimization parameters corresponding to this group of data are the optimal optimization parameter combination generated by evaluation; specifically, by loading the disk data processing strategy, disk test index reference values, disk performance analysis test result data, linearly analyzing and comparing the test result data, and saving multiple groups of linear analysis combination data; then loading multiple groups of linear analysis combination data, performing weighted ratio calculation, horizontally comparing the data, evaluating the results of the horizontal comparison data, and generating the optimal optimization parameter combination to achieve result processing and optimal parameter generation.

[0099] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing server disk performance analysis based on declarative type management recognition, characterized in that: Including the following steps: Define a declarative type management recognition model. The model includes major categories of disk combinations, specifically storage node disk combinations and management node disk combinations. Each disk combination involves the number of disks, disk type, disk size, RAID card mode, disk scheduling algorithm, disk access depth parameter, and disk NR request parameter. This definition process is achieved by creating a new model, sequentially selecting the node disk combination type, RAID card mode, disk type, setting the disk size, the number of disks of the same disk type, disk scheduling algorithm, disk access depth parameter, disk NR request parameter, disk test combination, and the average number of disk tests, and then saving and naming the current model.

2. The server disk performance analysis and optimization method based on declarative type management recognition according to claim 1, characterized in that: It also includes the step of scanning and collecting server disk information: Install a RAID card management tool in the operating system through an automated script, execute a disk query command, and filter the information to obtain the disk type, disk size, number of disks, RAID card mode, disk model, disk manufacturer, and disk firmware version. Specifically, this scanning and collection process is achieved by downloading the automated script and executing the installation tool, sequentially executing commands to query the basic configuration of the RAID card, disk type, disk size, number of disks, disk model, manufacturer, and firmware version, and finally formatting and saving the query results.

3. The server disk performance analysis and optimization method based on declarative type management recognition according to claim 2, wherein: It also includes the step of intelligently matching the declarative type management recognition model: Load multiple groups of defined declarative type management recognition models. Based on the currently collected server disk information, match with the models through a matching algorithm to obtain the performance test combination that should be carried out under the current server disk configuration. Specifically, this matching process is achieved by loading multiple groups of defined models, formatting the server disk information, and the matching algorithm program, sequentially matching the RAID card mode and disk type and synchronously matching the number of disks and disk size, determining the test scenario according to a specific intelligent matching algorithm logic, and finally saving the result of the matched disk test combination. Among them, the intelligent matching algorithm logic includes: when the number of disks of the same disk type is equal to 2 and the disk size is less than 600G, determine this type as the system disk and add a test scenario for forming RAID1; when the disk size is equal to 960G, determine it as an acceleration disk and add test scenarios for single-disk and 2-disk concurrent testing; when the number of disks of the same disk type is greater than 4 and less than 7 and the RAID card mode is Raid-mode, determine it as a management and control node disk and add a test scenario for forming RAID5; when the number of disks of the same disk type is greater than or equal to 10 and the RAID card mode is Jbod-mode, determine it as a storage node disk and add test scenarios for single-disk and multi-disk concurrent testing.

4. The server disk performance analysis and optimization method based on declarative type management recognition according to claim 3, wherein: It also includes the step of performing disk performance analysis and testing: According to the results of the matched disk test combinations, perform disk performance analysis tests respectively. Conduct disk performance analysis tests under different disk scheduling algorithms, different disk access depth parameters, and different disk NR request parameters, and format and save the disk performance analysis test data. Specifically, by loading the results of the matched disk test combinations, cycling through the configurations of disk scheduling algorithms, disk access depth parameters, and disk NR request parameters, perform disk performance analysis tests, test disk IOPS, bandwidth, and latency metrics. The read and write operation types are random read, random write, sequential read, sequential write, and mixed read and write. According to the set value of the average number of disk tests, cycle through the test process, and finally format and save the disk performance analysis test result data to implement this test process.

5. The server disk performance analysis and optimization method based on declarative type management recognition according to claim 4, characterized in that: It also includes the step of processing the test results according to the policy and evaluating to generate optimal optimization parameters: Through the disk data processing policy, linearly analyze and compare the test results, and save multiple linear analysis combination data; through weighted ratio calculation and horizontal comparison of multiple groups of linear analysis combination data, obtain the optimal linear analysis combination data, and the optimization parameters corresponding to this group of data are the optimal optimization parameter combinations generated by evaluation. Specifically, by loading the disk data processing policy, disk test index reference values, and disk performance analysis test result data, linearly analyze and compare the test result data, and save multiple groups of linear analysis combination data; then load multiple groups of linear analysis combination data, perform weighted ratio calculation, horizontally compare the data, evaluate the results of the horizontal comparison data, and generate the optimal optimization parameter combination to implement this processing and evaluation process.

6. A system for optimizing the analysis of server disk performance based on declarative type management recognition according to claim 5, characterized in that: It includes a model definition module for defining a declarative type management recognition model. The model includes major disk combinations, specifically storage node disk combinations and management node disk combinations. Each disk combination involves the number of disks, disk type, disk size, RAID card mode, disk scheduling algorithm, disk access depth parameter, and disk NR request parameter. This module realizes model definition by creating a new model, sequentially selecting the node disk combination type, RAID card mode, disk type, setting the disk size, the number of disks of the same disk type, disk scheduling algorithm, disk access depth parameter, disk NR request parameter, disk test combination, and the average number of disk tests, and saving and naming the current model.

7. The server disk performance analysis and optimization system based on declarative type management recognition according to claim 6, wherein: It also includes an information scanning module for scanning and collecting server disk information. Install a RAID card management tool in the operating system through an automated script, execute a disk query command, and filter the information to obtain the disk type, disk size, number of disks, RAID card mode, disk model, disk manufacturer, and disk firmware version. Specifically, by downloading the automated script and executing the installation tool, sequentially execute commands to query the basic configuration of the RAID card, disk type, disk size, number of disks, disk model, manufacturer, and firmware version, and finally format and save the query results to implement information scanning.

8. An optimization system for server disk performance analysis based on declarative type management recognition according to claim 7, characterized in that: It further includes a model matching module, which is used to intelligently match the declarative type management recognition model, load multiple groups of defined declarative type management recognition models, and match them with the models through a matching algorithm based on the currently collected server disk information, so as to obtain the performance test combination that should be carried out under the current server disk configuration; specifically, it is realized by loading multiple groups of defined models, formatting the server disk information, and the matching algorithm program, sequentially matching the RAID card mode, disk type and synchronously matching the disk quantity and disk size, determining the test scenario according to the specific intelligent matching algorithm logic, and finally saving the matching disk test combination result to achieve model matching; among them, the intelligent matching algorithm logic includes: when the quantity of disks of the same disk type is equal to 2 and the disk size is less than 600G, it is determined that this type is the system disk, and the test scenario of forming RAID1 is added; when the disk size is equal to 960G, it is determined as the acceleration disk, and the single-disk and 2-disk concurrent test scenarios are added; when the quantity of disks of the same disk type is greater than 4 and less than 7, and the RAID card mode is Raid-mode, it is determined as the management node disk, and the test scenario of forming RAID5 is added; when the quantity of disks of the same disk type is greater than or equal to 10, and the RAID card mode is Jbod-mode, it is determined as the storage node disk, and the single-disk and multi-disk concurrent test scenarios are added.

9. The server disk performance analysis and optimization system based on declarative type management recognition according to claim 8, characterized in that: It further includes a test execution module, which is used to execute the disk performance analysis test. According to the matching disk test combination result, the disk performance analysis test is carried out under different disk scheduling algorithms, different disk access depth parameters, and different disk NR request parameters, and the disk performance analysis test data is formatted and saved; specifically, it is realized by loading the matching disk test combination result, looping the disk scheduling algorithm, disk access depth parameter, and disk NR request parameter configuration, executing the disk performance analysis test, testing the disk IOPS, bandwidth, and latency indicators, and the read and write operation types are random read, random write, sequential read, sequential write, and mixed read and write. According to the set disk test average number value, the test process is looped, and finally the disk performance analysis test result data is formatted and saved to achieve test execution.

10. The server disk performance analysis and optimization system based on declarative type management recognition according to claim 9, characterized in that: It further includes a result processing and optimal parameter generation module, which is used to process the test results according to the policy and evaluate and generate the optimal optimization parameters. Through the disk data processing policy, the test results are linearly analyzed and compared, and multiple linear analysis combination data are saved; through the weighted ratio calculation of multiple groups of linear analysis combination data and horizontal comparison, the optimal linear analysis combination data is obtained, and the optimization parameters corresponding to this group of data are the optimal optimization parameter combination evaluated and generated; specifically, it is realized by loading the disk data processing policy, disk test index reference value, and disk performance analysis test result data, linearly analyzing and comparing the test result data, and saving multiple groups of linear analysis combination data; then loading multiple groups of linear analysis combination data, performing weighted ratio calculation, horizontal comparison of data, evaluating the horizontal comparison data result, and generating the optimal optimization parameter combination to achieve result processing and optimal parameter generation.

Images