Method and system for performance optimization of openGauss database for NUMA CPU architecture

Abstract

The embodiment of the application relates to the technical field of databases, in particular to an openGauss database performance optimization method and system for a NUMA CPU architecture, and the embodiment of the application is based on the AI frequent item of the openGauss database running text record to realize the performance optimization decision process of the openGauss database running text record, so that the performance optimization decision of the openGauss database running text record with different performance defects can be realized, and the applicability of the performance optimization decision of the openGauss database running text record is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of database, in particular to an openGauss database performance optimization method and system for NUMA CPU architecture. BACKGROUND

[0002] openGauss is an enterprise-level open source relational database, which is issued with the Mulan permissive license v2, provides extreme performance for multi-core architecture, full-link business, data security, AI-based tuning and efficient operation and maintenance capabilities. openGauss deeply integrates the research and development experience in the database field for many years, combines with the needs of enterprise-level scenarios, and continuously builds competitive features. At the same time, openGauss is also an open source and free database platform, which encourages community contribution and cooperation. With the continuous popularization of openGauss application, database performance optimization for openGauss has become inevitable. SUMMARY

[0003] In order to at least overcome the above-mentioned deficiencies in the prior art, one of the purposes of the present application is to provide an openGauss database performance optimization method and system for NUMA CPU architecture.

[0004] The present application provides an openGauss database performance optimization method for NUMA CPU architecture, which is applied to a database performance optimization system, and the method comprises the following steps:

[0005] Obtain an openGauss database running text record to be analyzed, mine a first original database running state frequent item and a second original database running state frequent item of the openGauss database running text record to be analyzed, perform AI frequent item aggregation on the first original database running state frequent item and the second original database running state frequent item, and obtain a database running state aggregated frequent item; wherein the feature capacity of the first original database running state frequent item is greater than the feature capacity of the second original database running state frequent item;

[0006] Perform dynamic filtering processing on the database running state aggregated frequent item to obtain a target database running state frequent item corresponding to the openGauss database running text record to be analyzed, and determine a target defect AI frequent item corresponding to the openGauss database running text record to be analyzed according to the database running state aggregated frequent item and the target database running state frequent item;

[0007] The target defect AI frequent item is used for AI frequent item adjustment on the target database running state frequent item, and a database running state upgrade frequent item is obtained. The database running state upgrade frequent item is decoded into an openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed.

[0008] In some possible embodiments, the first original database running state frequent item and the second original database running state frequent item of the openGauss database running text record to be analyzed are obtained by:

[0009] The openGauss database running text record to be analyzed is input into an AI frequent item mining model, and x AI frequent item mining dynamic filtering components in the AI frequent item mining model are used to mine a dynamic filtering AI frequent item corresponding to each AI frequent item mining dynamic filtering component in the openGauss database running text record to be analyzed. x is a positive integer.

[0010] The x dynamic filtering AI frequent items are subjected to dynamic filtering processing, and the first original database running state frequent item and the second original database running state frequent item of the openGauss database running text record to be analyzed are obtained.

[0011] In some possible embodiments, the first original database running state frequent item and the second original database running state frequent item of the openGauss database running text record to be analyzed are obtained by:

[0012] The x AI frequent item mining dynamic filtering components are divided into a first dynamic filtering component and a second dynamic filtering component. The level of the first dynamic filtering component in the AI frequent item mining model is less than the level of the second dynamic filtering component in the AI frequent item mining model.

[0013] The dynamic filtering AI frequent item corresponding to the first dynamic filtering component is subjected to dynamic filtering processing, and the first original database running state frequent item of the openGauss database running text record to be analyzed is obtained.

[0014] The dynamic filtering AI frequent item corresponding to the second dynamic filtering component is subjected to dynamic filtering processing, and the second original database running state frequent item of the openGauss database running text record to be analyzed is obtained.

[0015] In some possible embodiments, the AI frequent item aggregation on the first original database running status frequent item and the second original database running status frequent item to obtain a database running status aggregated frequent item comprises:

[0016] The first original database running status frequent item is subjected to frequent item derivation processing to obtain a first derived AI frequent item corresponding to the first original database running status frequent item;

[0017] The second original database running status frequent item is subjected to frequent item derivation processing to obtain a second derived AI frequent item corresponding to the second original database running status frequent item; the first derived AI frequent item and the second derived AI frequent item have the same AI frequent item feature capacity;

[0018] The first derived AI frequent item and the second derived AI frequent item are subjected to AI frequent item combination to obtain a combined AI frequent item;

[0019] The combined AI frequent item is subjected to AI frequent item information aggregation by an AI frequent item aggregation dynamic filtering component to obtain a database running status aggregated frequent item corresponding to the combined AI frequent item.

[0020] In some possible embodiments, the dynamic filtering processing on the database running status aggregated frequent item to obtain a target database running status frequent item corresponding to the openGauss database running text record to be analyzed comprises:

[0021] The database running status aggregated frequent item is subjected to dynamic filtering processing by an AI frequent item disassembly filtering component to obtain a database running status disassembly frequent item, and the database running status disassembly frequent item is subjected to trigger processing to obtain an AI frequent item relationship network;

[0022] The database running status aggregated frequent item is subjected to feature enhancement based on the AI frequent item relationship network to obtain a target database running status frequent item corresponding to the openGauss database running text record to be analyzed.

[0023] In some possible embodiments, the target defect AI frequent item corresponding to the openGauss database running text record to be analyzed is determined according to the database running status aggregated frequent item and the target database running status frequent item, comprising:

[0024] An AI frequent item difference between the database running status aggregated frequent item and the target database running status frequent item is determined as an original defect AI frequent item;

[0025] Obtain R defect AI frequent item attention surfaces included in the original defect AI frequent item, and among y local defect AI frequent items corresponding to the R defect AI frequent item attention surfaces, respectively, integrate the local defect AI frequent items under the same AI frequent item distribution label to obtain the target defect AI frequent item corresponding to the openGauss database running text record to be analyzed; R is a positive integer, and y is a positive integer.

[0026] In some possible embodiments, the target defect AI frequent item is obtained by the target defect AI frequent item, and the target database running state frequent item is obtained by the target defect AI frequent item. The database running state upgrade frequent item is obtained by the database running state upgrade frequent item, and the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed is obtained by the database running state upgrade frequent item.

[0027] Obtain the defect heat map corresponding to the target defect AI frequent item, and determine the AI frequent item to be adjusted in the target database running state frequent item according to the defect heat map;

[0028] The AI frequent item is adjusted to obtain the database running state upgrade frequent item, and the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed is obtained by the database running state upgrade frequent item.

[0029] In some possible embodiments, the target defect AI frequent item is obtained by the target defect AI frequent item, and the target database running state frequent item is obtained by the target defect AI frequent item. The database running state upgrade frequent item is obtained by the database running state upgrade frequent item, and the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed is obtained by the database running state upgrade frequent item.

[0030] The target defect AI frequent item and the target database running state frequent item are input into a state frequent item adjustment network, and a defect heat map of the target defect AI frequent item is obtained by the state frequent item adjustment network;

[0031] In the mark model in the state frequent item adjustment network, a database running state mark frequent item of the target database running state frequent item is obtained based on the defect heat map, and an openGauss database running text record adjustment AI frequent item of the target database running state frequent item is generated according to the database running state mark frequent item;

[0032] The openGauss database is run based on the text record adjustment AI frequent item, and the target database running state frequent item is adjusted by the AI frequent item to obtain a database running state upgrade frequent item.

[0033] The database running state upgrade frequent item is decoded into an openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed.

[0034] In some possible embodiments, the target defect AI frequent item is input into a state frequent item adjustment network together with the target database running state frequent item, the state frequent item adjustment network includes p local adjustment models, the p local adjustment models include a first local adjustment model and a second local adjustment model, and p is a positive integer.

[0035] In the first local adjustment model, the target database running state frequent item is adjusted by the AI frequent item based on the target defect AI frequent item to obtain a first adjustment AI frequent item, and the target defect AI frequent item is changed based on an AI frequent item adjustment result to obtain a first defect AI frequent item.

[0036] In the second local adjustment model, the first defect AI frequent item is used to adjust the first adjustment AI frequent item by the AI frequent item to obtain a database running state upgrade frequent item.

[0037] In the second local adjustment model, the first defect AI frequent item is used to adjust the first adjustment AI frequent item by the AI frequent item to obtain a database running state upgrade frequent item.

[0038] The database running state upgrade frequent item is decoded into an openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed.

[0039] The embodiment of the application further provides a database performance optimization system, including a processor, a memory and a bus connected with the processor, wherein the processor and the memory complete mutual communication through the bus; the processor is used to call program instructions in the memory to execute the openGauss database performance optimization method for the NUMA CPU architecture.

[0040] The embodiment of the present application also provides a computer readable storage medium, which stores a program, and the program is executed by a processor to realize the openGauss database performance optimization method for NUMA CPU architecture. Advantages

[0041] In the embodiment of the present application, the openGauss database running text record to be analyzed is obtained, the first original database running state frequent item and the second original database running state frequent item of the openGauss database running text record to be analyzed are mined, the AI frequent item aggregation is performed on the first original database running state frequent item and the second original database running state frequent item, and the database running state aggregated frequent item is obtained; the feature capacity of the first original database running state frequent item is greater than that of the second original database running state frequent item; the dynamic filtering processing is performed on the database running state aggregated frequent item, the target database running state frequent item corresponding to the openGauss database running text record to be analyzed is obtained, the target defect AI frequent item corresponding to the openGauss database running text record to be analyzed is determined according to the database running state aggregated frequent item and the target database running state frequent item; the AI frequent item adjustment is performed on the target database running state frequent item through the target defect AI frequent item, the database running state upgraded frequent item is obtained, and the database running state upgraded frequent item is decoded into the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed.

[0042] With the embodiment of the present application, the database performance optimization system can mine the intermediate AI frequent items (the first original database running state frequent item and the second original database running state frequent item) of the openGauss database running text record to be analyzed, instead of directly obtaining the final AI frequent item of the openGauss database running text record to be analyzed, so that the database performance optimization system can obtain the AI frequent items of the openGauss database running text record to be analyzed from different description angles, and can obtain multiple types of data from part to whole of the openGauss database running text record to be analyzed, thereby guaranteeing the richness of the obtained data, and based on the mined first original database running state frequent item and the second original database running state frequent item, the openGauss database running text record to be analyzed can be subjected to first state adjustment, and the target database running state frequent item obtained by the first state adjustment can be subjected to frequent item upgrading based on the obtained target defect AI frequent item, so as to realize the performance optimization decision of the openGauss database running text record to be analyzed, and thereby improve the timeliness of the performance optimization decision of the openGauss database running text record. Since the above-mentioned idea is based on the AI frequent item of the openGauss database running text record itself to realize the performance optimization decision process of the openGauss database running text record, the performance optimization decision of the openGauss database running text record with different performance defects can be realized, and the applicability of the performance optimization decision of the openGauss database running text record is improved. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0044] Figure 1 A flow chart of an openGauss database performance optimization method for NUMA CPU architecture provided by the embodiment of the present application.

[0045] Figure 2 A block diagram of a database performance optimization system provided by the embodiment of the present application.

[0046] FIG.

[0047] 100-database performance optimization system;

[0048] 101 - processor; 102 - memory; 103 - bus. Embodiments

[0049] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present application can be more thoroughly understood and the scope of the present application can be accurately conveyed to those skilled in the art.

[0050] In order to better understand the above technical solutions, the technical solutions of the present application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solutions of the present application, and not limitations of the technical solutions of the present application. In the case of no conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0051] Figure 1 A flowchart of the method for optimizing the performance of an openGauss database for a NUMA CPU architecture according to an embodiment of the present application is applied to a database performance optimization system and includes S10-S30.

[0052] S10, obtaining an openGauss database running text record to be analyzed, mining a first original database running state frequent item and a second original database running state frequent item of the openGauss database running text record to be analyzed, performing AI frequent item aggregation on the first original database running state frequent item and the second original database running state frequent item, and obtaining a database running state aggregated frequent item.

[0053] In the embodiments of the present application, the feature capacity of the first original database running state frequent item is greater than the feature capacity of the second original database running state frequent item.

[0054] The database performance optimization system is a server integrated with a CPU based on a NUMA architecture, which is used for feature mining and analysis processing of an openGauss database running text record, so as to make a decision on the performance optimization of the openGauss database. The database running state frequent item can be understood as a running state feature of the database. The embodiments of the present application can perform AI frequent item mining (feature extraction) through an AI neural network, so as to obtain a corresponding running state frequent item, and can perform a series of feature processing on the running state frequent item, such as aggregation processing, splicing processing, weighting processing, derivation processing, and decoding processing. The feature capacity is used to reflect the detail carrying capacity of the database running state frequent item.

[0055] S20, performing dynamic filtering processing on the database running state aggregated frequent items to obtain target database running state frequent items corresponding to the openGauss database running text record to be analyzed, and determining target defect AI frequent items corresponding to the openGauss database running text record to be analyzed according to the database running state aggregated frequent items and the target database running state frequent items.

[0056] In the embodiment of the application, the dynamic filtering processing can be understood as a convolution operation, which can be implemented by a convolutional neural network (CNN) for example. Further, the target defect AI frequent items are used to reflect the feature vectors corresponding to the defect state of the openGauss database during operation.

[0057] S30, performing AI frequent item adjustment on the target database running state frequent items through the target defect AI frequent items to obtain database running state upgrade frequent items, and translating the database running state upgrade frequent items into openGauss database running performance optimization records corresponding to the openGauss database running text record to be analyzed.

[0058] In the embodiment of the application, the AI frequent item adjustment can be understood as feature optimization or feature upgrade, such as performing upgrade optimization processing on the corresponding running state features, so as to obtain the database running state upgrade frequent items. The database running state upgrade frequent items can indicate the state feature description of the openGauss database after performance upgrade. Further, by performing feature translation on the database running state upgrade frequent items, the openGauss database running performance optimization records in the form of text can be obtained. It can be understood that, in combination with the database running state upgrade frequent items and the openGauss database running performance optimization records, the performance optimization decision of the openGauss database can be realized. For example, the business-level openGauss database performance optimization improvement can be performed through the database running state upgrade frequent items and the openGauss database running performance optimization records, so as to improve the accuracy and efficiency of the openGauss database performance optimization improvement.

[0059] The S10 includes S101 and S102.

[0060] S101, input the openGauss database running text record to be analyzed into an AI frequent item mining model, pass through x AI frequent item mining dynamic filtering components in the AI frequent item mining model, and mine the corresponding dynamic filtering AI frequent item of the openGauss database running text record to be analyzed in each AI frequent item mining dynamic filtering component.

[0061] Wherein, x is a positive integer.

[0062] S102, dynamically filter the x dynamic filtering AI frequent items to obtain the first original database running state frequent item and the second original database running state frequent item of the openGauss database running text record to be analyzed.

[0063] Further, the dynamic filtering processing of the x dynamic filtering AI frequent items in S102 obtains the first original database running state frequent item and the second original database running state frequent item of the openGauss database running text record to be analyzed, including S1021-S1023.

[0064] S1021, divide the x AI frequent item mining dynamic filtering components into a first dynamic filtering component and a second dynamic filtering component.

[0065] Wherein, the level of the first dynamic filtering component in the AI frequent item mining model is less than the level of the second dynamic filtering component in the AI frequent item mining model. Further, the level can be understood as the depth of the network component.

[0066] S1022, dynamically filter the dynamic filtering AI frequent item corresponding to the first dynamic filtering component to obtain the first original database running state frequent item of the openGauss database running text record to be analyzed.

[0067] S1023, dynamically filter the dynamic filtering AI frequent item corresponding to the second dynamic filtering component to obtain the second original database running state frequent item of the openGauss database running text record to be analyzed.

[0068] In this way, the first original database running state frequent item and the second original database running state frequent item can be accurately and completely obtained.

[0069] In some examples, the AI frequent item aggregation of the first original database running state frequent item and the second original database running state frequent item in S10 to obtain the database running state aggregated frequent item includes: performing frequent item derivation processing on the first original database running state frequent item to obtain a first derived AI frequent item corresponding to the first original database running state frequent item; performing frequent item derivation processing on the second original database running state frequent item to obtain a second derived AI frequent item corresponding to the second original database running state frequent item; the first derived AI frequent item and the second derived AI frequent item have the same AI frequent item feature capacity; performing AI frequent item combination on the first derived AI frequent item and the second derived AI frequent item to obtain a combined AI frequent item; and performing AI frequent item information aggregation on the combined AI frequent item by an AI frequent item aggregation dynamic filtering component to obtain a database running state aggregated frequent item corresponding to the combined AI frequent item.

[0070] In this way, through the feature aggregation processing, the comprehensive running feature performance capability of the database running state aggregated frequent item can be improved.

[0071] In some example embodiments, the dynamic filtering processing of the database running state aggregated frequent item in S20 to obtain the target database running state frequent item corresponding to the openGauss database running text record to be analyzed includes S201 and S202.

[0072] S201, performing dynamic filtering processing on the database running state aggregated frequent item by an AI frequent item disassembly filtering component to obtain a database running state disassembly frequent item, and performing trigger processing on the database running state disassembly frequent item to obtain an AI frequent item relationship network;

[0073] S202, performing feature strengthening on the database running state aggregated frequent item based on the AI frequent item relationship network to obtain the target database running state frequent item corresponding to the openGauss database running text record to be analyzed.

[0074] In this way, the feature recognition degree and feature performance capability of the target database running state frequent item can be improved.

[0075] In some examples, the determining the target defect AI frequent item corresponding to the openGauss database running text record to be analyzed in the step S20 comprises: distinguishing AI frequent items between the database running state aggregated frequent item and the target database running state frequent item, and determining the AI frequent items as original defect AI frequent items; obtaining R defect AI frequent item attention surfaces included in the original defect AI frequent items, and performing AI frequent item integration on local defect AI frequent items under the same AI frequent item distribution label in y local defect AI frequent items corresponding to the R defect AI frequent item attention surfaces respectively, to obtain the target defect AI frequent item corresponding to the openGauss database running text record to be analyzed; R is a positive integer, and y is a positive integer. In this way, the target defect AI frequent item can be completely determined, and omission of the target defect AI frequent item can be avoided.

[0076] In some optional embodiments, the obtaining the database running state upgrade frequent item by adjusting the target database running state frequent item according to the target defect AI frequent item in the step S30, and translating the database running state upgrade frequent item into the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed comprises: obtaining a defect heat map corresponding to the target defect AI frequent item, and determining a to-be-adjusted AI frequent item in the target database running state frequent item according to the defect heat map; adjusting the to-be-adjusted AI frequent item to obtain a database running state upgrade frequent item, and performing feature translation on the database running state upgrade frequent item to obtain the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed. In this way, targeted and accurate performance optimization decisions can be made in combination with the defect heat map.

[0077] In some optional embodiments, the obtaining, by the target defect AI frequent item, the database running state frequent item, and performing AI frequent item adjustment on the database running state frequent item to obtain a database running state upgrade frequent item, and decoding the database running state upgrade frequent item into an openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed in the S30, comprises: inputting the target defect AI frequent item and the target database running state frequent item into a state frequent item adjustment network, and obtaining a defect heat map of the target defect AI frequent item through the state frequent item adjustment network; in a label model in the state frequent item adjustment network, obtaining a database running state label frequent item of the target database running state frequent item based on the defect heat map, and generating an openGauss database running text record adjustment AI frequent item of the target database running state frequent item according to the database running state label frequent item; performing AI frequent item adjustment on the target database running state frequent item based on the openGauss database running text record adjustment AI frequent item to obtain a database running state upgrade frequent item; and decoding the database running state upgrade frequent item into an openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed. In this way, targeted and accurate performance optimization decisions can be made in combination with the defect heat map.

[0078] In some optional embodiments, the obtaining, by the target defect AI frequent item, the database running state frequent item, and performing AI frequent item adjustment on the database running state frequent item to obtain a database running state upgrade frequent item, and decoding the database running state upgrade frequent item into an openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed, in S30, includes: inputting the target defect AI frequent item and the target database running state frequent item into a state frequent item adjustment network; the state frequent item adjustment network includes p local adjustment models, the p local adjustment models include a first local adjustment model and a second local adjustment model, and p is a positive integer; in the first local adjustment model, the target defect AI frequent item is adjusted based on the target database running state frequent item to obtain a first adjusted AI frequent item, and the target defect AI frequent item is changed based on the AI frequent item adjustment result to obtain a first defect AI frequent item; in the second local adjustment model, the first defect AI frequent item is adjusted based on the first adjusted AI frequent item to obtain the database running state upgrade frequent item; and the database running state upgrade frequent item is decoded into the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed. In this way, the database running state upgrade frequent item and the openGauss database running performance optimization record can be accurately and completely determined, so that the openGauss database performance optimization improvement at the business level is performed through the database running state upgrade frequent item and the openGauss database running performance optimization record, to improve the accuracy and efficiency of the openGauss database performance optimization improvement.

[0079] The embodiment of the present application provides a computer readable storage medium, which stores a program, and the program is executed by a processor to implement the openGauss database performance optimization method for NUMA CPU architecture.

[0080] The embodiment of the present application provides a processor for running a program, wherein the program is executed to perform the openGauss database performance optimization method for NUMA CPU architecture.

[0081] In the embodiment of the present application, as Figure 2As shown, the database performance optimization system 100 includes at least one processor 101, and at least one memory 102 connected with the processor 101, a bus 103; wherein the processor 101, the memory 102 complete mutual communication through the bus 103; the processor 101 is used to call the program instruction in the memory 102, to execute the above-mentioned openGauss database performance optimization method for NUMA CPU architecture.

[0082] The application is described with reference to flowcharts and / or block diagrams of the method, database performance optimization system (system) and computer program product according to the embodiments of the application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of the flows and / or blocks in the flowchart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a machine that implements the function specified in the flowchart and / or block diagram. Figure 1 The function specified in one flow or multiple flows and / or blocks. Figure 1 The device that realizes the function specified in one block or multiple blocks.

[0083] In a typical configuration, the database performance optimization system includes one or more processors (CPU), memory and bus. The database performance optimization system can also include input / output interface, network interface, etc.

[0084] The memory can include non-permanent memory in the computer readable medium, random access memory (RAM) and / or non-volatile memory, etc. such as read-only memory (ROM) or flash memory (flash RAM), the memory includes at least one memory chip. The memory is an example of the computer readable medium.

[0085] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage computer-readable storage media, or any other non-transmission medium that can be used to store information that can be accessed by a database performance optimization system. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0086] It should also be noted that the terms "comprising", "comprising" or any other variant thereof are intended to cover non-exclusive inclusions, so that a process, method, article or computer-readable storage medium including a series of elements includes not only those elements, but also other elements not explicitly listed or inherent to such a process, method, article or computer-readable storage medium. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or computer-readable storage medium including the element.

[0087] Those skilled in the art will appreciate that embodiments of the present application can be provided as a method, system or computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0088] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the scope of claims of the present application.

Claims

1. A performance optimization method for the openGauss database on a NUMA CPU architecture, characterized in that, The method, applied to a database performance optimization system, includes: The openGauss database runtime text records to be analyzed are obtained. The first and second original frequent database runtime status items of the openGauss database runtime text records to be analyzed are mined. The first and second original frequent database runtime status items are aggregated using AI frequent item aggregation to obtain the aggregated frequent database runtime status items. The feature capacity of the first original frequent database runtime status items is greater than the feature capacity of the second original frequent database runtime status items. Dynamic filtering is performed on the frequent items of the database running status aggregation to obtain the frequent items of the target database running status corresponding to the running text record of the openGauss database to be analyzed. Based on the frequent items of the database running status aggregation and the frequent items of the target database running status, the frequent items of the target defect AI corresponding to the running text record of the openGauss database to be analyzed are determined. By adjusting the AI ​​frequent items of the target database running status through the AI ​​frequent items of the target defect AI, the database running status upgrade frequent items are obtained, and the database running status upgrade frequent items are decoded into the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed. The process involves adjusting the frequent AI items of the target database operating status using the frequent AI items of the target defect to obtain frequent database operating status upgrade items. These frequent database operating status upgrade items are then decoded into OpenGauss database performance optimization records corresponding to the OpenGauss database operating text records to be analyzed. This includes: Obtain the defect heatmap corresponding to the frequent items of the target defect AI, and determine the frequent AI items to be adjusted in the frequent items of the target database operation status based on the defect heatmap; The AI ​​frequent item to be adjusted is adjusted to obtain the database running status upgrade frequent item. The database running status upgrade frequent item is then translated to obtain the openGauss database running performance optimization record corresponding to the openGauss database running text record to be analyzed.

2. The method according to claim 1, characterized in that, The mining of frequent items in the first and second original database running states of the openGauss database running text records to be analyzed includes: The openGauss database running text records to be analyzed are input into the AI ​​frequent item mining model. After passing through x AI frequent item mining dynamic filtering components in the AI ​​frequent item mining model, the dynamic filtered AI frequent items corresponding to the running text records of the openGauss database to be analyzed in each AI frequent item mining dynamic filtering component are mined; x is a positive integer. Dynamic filtering is applied to x frequent AI terms to obtain the first and second frequent terms of the original database running state of the openGauss database running text records to be analyzed. The process of performing dynamic filtering on x frequent AI terms to obtain the first and second frequent terms of the original database running status of the openGauss database running text records to be analyzed includes: The x AI frequent item mining dynamic filtering components are divided into a first dynamic filtering component and a second dynamic filtering component; the level of the first dynamic filtering component in the AI ​​frequent item mining model is lower than the level of the second dynamic filtering component in the AI ​​frequent item mining model. Dynamic filtering is performed on the frequent AI terms corresponding to the first dynamic filtering component to obtain the first original database running status frequent terms of the openGauss database running text records to be analyzed. Dynamic filtering is performed on the frequent terms of the dynamic filtering AI corresponding to the second dynamic filtering component to obtain the second original database running status frequent terms of the openGauss database running text records to be analyzed.

3. The method according to claim 1, characterized in that, The step of aggregating frequent items of the first and second original database operating states using AI to obtain aggregated frequent items of database operating states includes: Frequent item derivation processing is performed on the frequent items of the first original database running status to obtain the first derived AI frequent items corresponding to the frequent items of the first original database running status. The frequent items of the second original database running status are subjected to frequent item derivation processing to obtain the second derived AI frequent items corresponding to the frequent items of the second original database running status; the first derived AI frequent items and the second derived AI frequent items have the same AI frequent item feature capacity. The first derived AI frequent term and the second derived AI frequent term are combined to obtain the combined AI frequent term; The AI ​​frequent item aggregation dynamic filtering component aggregates the AI ​​frequent item information of the combined AI frequent items to obtain the database running status aggregation frequent items corresponding to the combined AI frequent items.

4. The method according to claim 1, characterized in that, The dynamic filtering process performed on the frequent items of the aggregated database running status to obtain the frequent items of the target database running status corresponding to the openGauss database running text records to be analyzed includes: The frequent items of the database operation status aggregation are dynamically filtered using an AI frequent item decomposition and filtering component to obtain frequent items of the database operation status decomposition. The frequent items of the database operation status decomposition are then triggered to obtain the AI ​​frequent item relationship network. Based on the AI ​​frequent item relationship network, the frequent items of the database operation status are aggregated and feature-enhanced to obtain the frequent items of the target database operation status corresponding to the openGauss database operation text records to be analyzed.

5. The method according to claim 1, characterized in that, The step of determining the target defect AI frequent items corresponding to the openGauss database running text records to be analyzed, based on the frequent items aggregated from the database running status and the frequent items from the target database running status, includes: The AI ​​frequent items difference between the aggregated frequent items of the database running status and the frequent items of the target database running status are determined as the original defect AI frequent items; Obtain the R frequent defect AI items of concern included in the original frequent defect AI items. Among the y frequent local defect AI items corresponding to the R frequent defect AI items of concern, the frequent local defect AI items under the same frequent AI item distribution label are integrated into the frequent AI items to obtain the target frequent defect AI items corresponding to the openGauss database running text records to be analyzed; R is a positive integer and y is a positive integer.

6. The method according to claim 1, characterized in that, The process involves adjusting the frequent AI items of the target database operating status using the frequent AI items of the target defect to obtain frequent database operating status upgrade items. These frequent database operating status upgrade items are then decoded into OpenGauss database performance optimization records corresponding to the OpenGauss database operating text records to be analyzed. This includes: The frequent items of the target defect AI and the frequent items of the target database operation status are input into the frequent items of status adjustment network, and the defect heat map of the frequent items of the target defect AI is obtained through the frequent items of status adjustment network. In the labeling model of the state frequent item adjustment network, the database operation state label frequent item of the target database operation state is obtained based on the defect heatmap, and the openGauss database operation text record adjustment AI frequent item of the target database operation state is generated based on the database operation state label frequent item. Based on the text records of the openGauss database, the AI ​​frequent items for adjusting the running status of the target database are adjusted to obtain the frequent items for database running status upgrades. The frequently upgraded database running status items are decoded into the OpenGauss database running performance optimization records corresponding to the OpenGauss database running text records to be analyzed.

7. The method according to claim 1, characterized in that, The process involves adjusting the frequent AI items of the target database operating status using the frequent AI items of the target defect to obtain frequent database operating status upgrade items. These frequent database operating status upgrade items are then decoded into OpenGauss database performance optimization records corresponding to the OpenGauss database operating text records to be analyzed. This includes: The frequent items of the target defect AI and the frequent items of the target database operation status are input into the frequent item adjustment network; the frequent item adjustment network includes p local adjustment models, the p local adjustment models include a first local adjustment model and a second local adjustment model, where p is a positive integer; In the first local adjustment model, the AI ​​frequent item of the target database running status is adjusted based on the target defect AI frequent item to obtain the first adjusted AI frequent item. Based on the AI ​​frequent item adjustment result, the target defect AI frequent item is changed to obtain the first defect AI frequent item. In the second local adjustment model, the first adjustment AI frequent item is adjusted based on the first defect AI frequent item to obtain the database running status upgrade frequent item; The frequently upgraded database running status items are decoded into the OpenGauss database running performance optimization records corresponding to the OpenGauss database running text records to be analyzed.

8. A database performance optimization system, characterized in that, The system includes a processor, a memory, and a bus connected to the processor; wherein the processor and the memory communicate with each other via the bus; the processor is used to call program instructions in the memory to execute the openGauss database performance optimization method for NUMA CPU architecture as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, It stores a program that, when executed by a processor, implements the openGauss database performance optimization method for NUMA CPU architecture as described in any one of claims 1-7.

Images