Cluster fault repair method and model training method, device and server

Abstract

The embodiments of this specification relate to a cluster fault repair method, device and server. The method includes: obtaining fault log data; the fault log data includes historical fault data and historical fault repair strategies; forming a fault text according to the historical fault data; representing the fault text as a structured text representation vector; inputting the text representation vector into a multi-layer classifier, outputting the association between the text representation vector and the historical fault repair strategy, and obtaining the cluster fault repair model. Thus, a cluster fault repair model with higher efficiency and accuracy is obtained, thereby realizing automatic repair of cluster faults with high efficiency and high accuracy.

Description

Technical Field

[0001] The embodiments of this specification relate to the technical field of automatic cluster fault processing, and more specifically, to a cluster fault repair model training method, a cluster fault repair method, a cluster fault repair model training device, a cluster fault repair device, a server, and a computer-readable storage medium. Background Art

[0002] In the operation and maintenance of large-scale cluster servers, intelligent machine repair is very important. Due to the large number of servers in the cluster, there are also many servers that fail at the same time. In the existing operation and maintenance system, a software and hardware automatic repair system (auto-healing) is deployed, and its repair strategy is based on rule matching. It has a fast response speed and high interpretability.

[0003] Although this solution realizes the automatic repair of server failures in the cluster, the repair strategy is configured by experts and triggered by a single rule. For example, only one monitor is paid attention to among multiple alarm monitoring modules (monitors), or only a certain state of the server is paid attention to, such as state = os_error, etc. This will lead to low accuracy of fault repair and incomplete rule coverage.

[0004] Therefore, it is necessary to provide a high-efficiency and high-accuracy cluster fault repair method. Summary of the invention

[0005] One purpose of the embodiments of this specification is to provide a new technical solution for cluster fault recovery.

[0006] According to a first aspect of an embodiment of this specification, a cluster fault recovery model training method is provided, including:

[0007] Acquire fault log data; the fault log data includes historical fault data and historical fault repair strategies;

[0008] forming a fault text according to the historical fault data;

[0009] Representing the fault text as a structured text representation vector;

[0010] The text representation vector is input into a multi-layer classifier, and the association relationship between the text representation vector and the historical fault repair strategy is output to obtain the cluster fault repair model.

[0011] Optionally, the forming of a fault text according to the historical fault data includes:

[0012] Extracting the fault monitoring module name and the corresponding fault level from the historical fault data;

[0013] The fault monitoring module name and the corresponding fault level form the fault text.

[0014] Optionally, the step of representing the fault text as a structured text representation vector includes:

[0015] According to a preset fault level weight value, the fault text is converted into a corresponding word frequency vector;

[0016] The word frequency vector is input into a language model to obtain the structured text representation vector.

[0017] Optionally, the language model is a probabilistic statistical model or a deep learning language model; wherein the probabilistic statistical model includes a term frequency-inverse document frequency tf-idf model and a document topic generation model LDA; and the deep learning language model is a word2vec model.

[0018] Optionally, the historical fault repair strategy includes a successful repair strategy and an unsuccessful repair strategy;

[0019] The step of inputting the text representation vector into a multi-layer classifier and outputting an association relationship between the text representation vector and the historical fault repair strategy includes:

[0020] Inputting the text representation vector into a multi-classifier, and outputting the association relationship between the text representation vector and the successful repair strategy;

[0021] And, the text representation vector is input into a binary classifier, and the correlation relationship between the text representation vector and the successful repair strategy and the unsuccessful repair strategy is output.

[0022] According to a second aspect of an embodiment of this specification, a cluster fault recovery method is provided, including:

[0023] Obtain the fault data to be repaired;

[0024] Input the fault data to be repaired into the cluster fault repair model, and output a corresponding fault repair strategy;

[0025] The cluster fault repair model is formed by forming a fault text according to historical fault data, representing the fault text as a structured text representation vector, inputting the text representation vector into a multi-layer classifier for training, and outputting the correlation relationship between the text representation vector and the historical fault repair strategy.

[0026] Optionally, after outputting the corresponding fault repair strategy, the method further includes:

[0027] The to-be-repaired fault data and the corresponding fault repair strategy are stored in a preset fault knowledge base.

[0028] According to a third aspect of an embodiment of this specification, a training device for a cluster fault recovery model is provided, including:

[0029] An acquisition module, used to acquire fault log data; the fault log data includes historical fault data and historical fault repair strategies;

[0030] A forming module, used for forming a fault text according to the historical fault data;

[0031] A representation module, used for representing the fault text as a structured text representation vector;

[0032] The training module is used to input the text representation vector into a multi-layer classifier, output the association relationship between the text representation vector and the historical fault repair strategy, and obtain the cluster fault repair model.

[0033] According to a fourth aspect of an embodiment of this specification, a cluster fault recovery device is provided, including:

[0034] An acquisition module is used to acquire the fault data to be repaired;

[0035] An output module, used for inputting the fault data to be repaired into the cluster fault repair model and outputting a corresponding fault repair strategy;

[0036] The cluster fault repair model is formed by forming a fault text according to historical fault data, representing the fault text as a structured text representation vector, inputting the text representation vector into a multi-layer classifier for training, and outputting the correlation relationship between the text representation vector and the historical fault repair strategy.

[0037] According to a fifth aspect of an embodiment of this specification, a server is provided, including the training device of the cluster fault recovery model as described in the third aspect of the embodiment of this specification, or the server includes:

[0038] A memory for storing executable commands;

[0039] The processor is used to execute the cluster fault repair model training method as described in any one of the first aspects of the embodiments of this specification under the control of the executable command.

[0040] According to a sixth aspect of an embodiment of this specification, a server is provided, including the cluster fault recovery device according to the fourth aspect of the embodiment of this specification, or the server includes:

[0041] A memory for storing executable commands;

[0042] The processor is used to execute the cluster fault repair method as described in any one of the second aspects of the embodiments of this specification under the control of the executable command.

[0043] According to a seventh aspect of the embodiments of this specification, a computer-readable storage medium is further provided, storing executable instructions, which, when executed by a processor, execute the cluster fault repair model training method as described in the first aspect of the embodiments of this specification.

[0044] According to an eighth aspect of the embodiments of this specification, a computer-readable storage medium is further provided, storing executable instructions, and when the executable instructions are executed by a processor, the cluster fault repair method as described in the second aspect of the embodiments of this specification is executed.

[0045] One beneficial effect of the embodiment of this specification is that the method of this embodiment obtains fault log data; the fault log data includes historical fault data and historical fault repair strategies; forms a fault text according to the historical fault data; represents the fault text as a structured text representation vector; inputs the text representation vector into a multi-layer classifier, and outputs the association relationship between the text representation vector and the historical fault repair strategy to obtain the cluster fault repair model. Thus, a cluster fault repair model with higher efficiency and accuracy is obtained, thereby realizing automatic repair of cluster faults with high efficiency and high accuracy.

[0046] Other features and advantages of the embodiments of the present specification will become apparent from the following detailed description of exemplary embodiments of the embodiments of the present specification with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the embodiments of the present specification and, together with the description, serve to explain the principles of the embodiments of the present specification.

[0048] Figure 1 A schematic diagram of a system architecture for implementing a cluster fault recovery method according to an embodiment of this specification;

[0049] Figure 2 is a block diagram showing a hardware configuration of a server 1000 that can implement an embodiment of the present invention;

[0050] Figure 3is a flow chart of a cluster fault repair method according to an embodiment of this specification;

[0051] Figure 4 is a principle block diagram of a cluster fault recovery device according to an embodiment of this specification;

[0052] Figure 5 A functional block diagram of a server according to an embodiment of this specification is shown;

[0053] Figure 6 is a flow chart of a cluster fault recovery method according to the second embodiment of this specification;

[0054] Figure 7 is a principle block diagram of a cluster fault recovery device according to the second embodiment of this specification;

[0055] Figure 8 It is a principle block diagram showing a server according to the second embodiment of this specification. DETAILED DESCRIPTION

[0056] Various exemplary embodiments of the present specification will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present specification embodiments.

[0057] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the embodiments of the present specification and its application or uses.

[0058] Technologies, methods and equipment known to persons of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such technologies, methods and equipment should be considered part of the specification.

[0059] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0060] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0061] <Hardware Configuration>

[0062] Figure 1 This is a schematic diagram of the system architecture for implementing the cluster fault recovery method of the embodiment of this specification.

[0063] Reference Figure 1As shown, the implementation system of the cluster fault repair method includes a data center and multiple servers, such as server 1, server 2, ... server n. When any server in the cluster fails, the failure is reported to the data center, and the data center provides a corresponding fault repair strategy.

[0064] Specifically, the data center may include a preprocessing module, a language module, a multi-layer classification module, and a fault knowledge base.

[0065] During the training process of the cluster fault repair model, for the semi-structured fault log data of server faults, the fault log data generated by expert experience is first collected, and the fault log data includes historical fault data and historical fault repair strategies; the fault log data is input into the preprocessing module, and then the preprocessing module is used to extract the server's fault monitoring module name (monitor name) and the corresponding fault level (level), such as fatal level (fatal), error level (error), etc., to form a fault text containing historical fault data.

[0066] Then the fault text is input into the language module, in which reasonable weight values ​​are pre-set for the fault levels, such as fatal = 5 and error = 4. According to the preset fault level weight values, the fault text is converted into the corresponding word frequency vector, and then the word frequency vector is input into the language model, such as the probability statistics model (tf-idf, lda) or the deep language model (word2vec), to obtain a structured text representation vector.

[0067] Finally, the text representation vector is input into the multi-layer classifier to obtain the correlation between the text representation vector and the historical fault repair strategy. The multi-layer classifier includes a multi-classifier and a binary classifier. The multi-classifier is used to learn the successful repair strategy, while the binary classifier is used to learn both successful and unsuccessful repair strategies, thereby completing the training of the cluster fault repair model.

[0068] After a server fails and reports the fault data to be repaired to the data center, the data center inputs the fault data to be repaired into the cluster fault repair model and outputs the corresponding fault repair strategy. This avoids the need for expert intervention for vector interpretation in the existing technology, thereby improving the efficiency and accuracy of fault repair.

[0069] At the same time, knowledge abstraction and expert confirmation are performed on the successfully repaired fault log data to form new repair rules, which are stored in the fault knowledge base to facilitate the root cause analysis of the fault.

[0070] In this way, the entire data center is composed of a preprocessing module, a language module, a multi-layer classifier and a fault knowledge base, which forms an automated processing of server fault data. It can also make full use of a large amount of successful and unsuccessful repair fault data for learning, further improving the accuracy and efficiency of cluster repair model fault repair, and can perform knowledge abstraction and expert confirmation on the successfully repaired fault log data to generate new repair rules, further increasing interpretability.

[0071] Figure 2 is a block diagram showing a hardware configuration of a server 1000 that can implement an embodiment of the present invention.

[0072] The server 1000 may be, for example, a blade server.

[0073] In one example, server 1000 may be a computer.

[0074] In another example, the server 1000 may be Figure 2 As shown, the server includes a processor 1100, a memory 1200, an interface device 1300, a communication device 1400, a display device 1500, and an input device 1600. Although the server may also include a speaker, a microphone, etc., these components are irrelevant to the present invention and are therefore omitted here.

[0075] The processor 1100 may be, for example, a central processing unit (CPU), a microprocessor (MCU), etc. The memory 1200 may include, for example, a ROM (read-only memory), a RAM (random access memory), a non-volatile memory such as a hard disk, etc. The interface device 1300 may include, for example, a USB interface, a serial interface, etc. The communication device 1400 may be capable of wired or wireless communication, for example. The display device 1500 may be, for example, a liquid crystal display screen. The input device 1600 may include, for example, a touch screen, a keyboard, etc.

[0076] In this embodiment, the server 1000 obtains fault log data; the fault log data includes historical fault data and historical fault repair strategies; a fault text is formed according to the historical fault data; the fault text is represented as a structured text representation vector; the text representation vector is input into a multi-layer classifier, and the association relationship between the text representation vector and the historical fault repair strategy is output to obtain the cluster fault repair model.

[0077] Figure 2The server shown is merely illustrative and does not in any way imply any limitation on the present invention, its application or use. In the embodiment of the present invention, the memory 1200 of the server 1000 is used to store instructions for controlling the processor 1100 to operate to perform any cluster fault recovery model training method provided in the embodiment of the present invention.

[0078] It should be understood by those skilled in the art that although Figure 2 , multiple devices are shown for the server 1000 ; however, the present invention may only involve some of the devices, for example, the server 1000 only involves the processor 1100 and the storage device 1200 .

[0079] A technician can design instructions according to the scheme disclosed in the present invention. How instructions control the processor to operate is well known in the art, so it will not be described in detail here.

[0080] <First Embodiment>

[0081] <Method>

[0082] This embodiment provides a cluster fault recovery model training method, which can be performed as follows: Figure 2 The server 1000 shown executes.

[0083] like Figure 3 As shown, the method includes the following steps 3100-3400:

[0084] Step 3100, obtaining fault log data; the fault log data includes historical fault data and historical fault repair strategies.

[0085] Specifically, the fault log data may be generated by expert experience. The historical fault repair strategy includes a successful repair strategy and an unsuccessful repair strategy. The historical fault data may include information such as the fault monitoring module name, fault level, and repair strategy.

[0086] Step 3200: generating a fault text based on the historical fault data.

[0087] In this step, the server 1000 may extract the fault monitoring module name and the corresponding fault level from the historical fault data, and make the fault monitoring module name and the corresponding fault level form the fault text.

[0088] Step 3300: represent the fault text as a structured text representation vector.

[0089] In practical applications, it is necessary to pre-set weight values ​​for different fault levels. For example, when the fault level is fatal, its preset fault level weight value can be set to 5, and when the fault level is error, its preset fault level weight value can be set to 4.

[0090] In this step, the server 1000 can convert the fault text into a corresponding word frequency vector according to a preset fault level weight value; and then input the word frequency vector into a language model to obtain the structured text representation vector.

[0091] The language model may be a probabilistic statistical model or a deep learning language model. The probabilistic statistical model includes a term frequency-inverse document frequency (tf-idf) model or a latent Dirichlet Allocation (LDA) model; the deep learning language model may be, for example, a word2vec model. No specific limitation is given here.

[0092] In one example, the language module uses the LDA model, and during the model training process, Gibbs sampling is used for iteration. The probability distribution calculation formula can be

[0093] From the probability distribution calculation formula, we can obtain the probability that the fault text (alarm information of the monitor combination) belongs to each potential fault mode. Each potential fault mode corresponds to a corresponding fault repair strategy. These probabilities constitute the "repair features" of the fault text, and we have completed the feature extraction of each document.

[0094] Step 3400: input the text representation vector into a multi-layer classifier, output the association relationship between the text representation vector and the historical fault repair strategy, and obtain the cluster fault repair model.

[0095] Among them, the multi-layer classifier includes a multi-classifier and a binary classifier. The multi-classifier is used to learn the successful repair strategy, and the binary classifier is used to learn both successful and unsuccessful repair strategies. Specifically, the server 1000 inputs the text representation vector into the multi-classifier, and outputs the association relationship between the text representation vector and the successful repair strategy; and, inputs the text representation vector into the binary classifier, and outputs the association relationship between the text representation vector and the successful repair strategy and the unsuccessful repair strategy.

[0096] In this way, when the cluster repair model is applied, the multi-classifier first outputs the fault repair strategy corresponding to the fault data to be repaired. In order to further confirm the accuracy of the fault repair strategy and ensure the stability of the system, the fault repair strategy is then input into the binary classifier to determine whether the fault repair strategy is successful. That is, the multi-layer classifier is used to output the fault repair strategy, and the binary classifier is used to determine whether the fault repair strategy is successful.

[0097] In addition, if the binary classifier determines that the operation corresponding to the fault repair strategy will fail, in actual applications, upgrade recommendations can be made based on the cost of the operation corresponding to the fault repair strategy, thereby eliminating invalid fault repair strategies, thereby further ensuring the stability and efficiency of cluster fault repair.

[0098] The method of this embodiment obtains fault log data; the fault log data includes historical fault data and historical fault repair strategies; forms a fault text according to the historical fault data; represents the fault text as a structured text representation vector; inputs the text representation vector into a multi-layer classifier, and outputs the association relationship between the text representation vector and the historical fault repair strategy to obtain the cluster fault repair model. Thus, a cluster fault repair model with higher efficiency and accuracy is obtained, thereby realizing automatic repair of cluster faults with high efficiency and high accuracy.

[0099] <Device Example>

[0100] This embodiment provides a training device for a cluster fault recovery model. The device is, for example, Figure 4 The cluster fault repair model training device 4000 shown in the figure may include an acquisition module 4100 , a formation module 4200 , a representation module 4300 and a training module 4400 .

[0101] The acquisition module 4100 is used to acquire fault log data; the fault log data includes historical fault data and historical fault repair strategies.

[0102] The forming module 4200 is used to form a fault text according to the historical fault data.

[0103] The representation module 4300 is used to represent the fault text as a structured text representation vector.

[0104] The training module 4400 is used to input the text representation vector into a multi-layer classifier, output the association relationship between the text representation vector and the historical fault repair strategy, and obtain the cluster fault repair model.

[0105] Optionally, the forming module 4200 may be specifically used to extract the fault monitoring module name and the corresponding fault level from the historical fault data; and form the fault text with the fault monitoring module name and the corresponding fault level.

[0106] Optionally, the representation module 4300 can be specifically used to convert the fault text into a corresponding word frequency vector according to a preset fault level weight value; input the word frequency vector into a language model to obtain the structured text representation vector. The language model is a probabilistic statistical model or a deep learning language model; the probabilistic statistical model includes a tf-idf model and an LDA model; the deep learning language model is a word2vec language model.

[0107] In one example, the historical fault repair strategy includes a successful repair strategy and an unsuccessful repair strategy. The training module 4400 is specifically used to input the text representation vector into a multi-classifier, output the association relationship between the text representation vector and the successful repair strategy; and input the text representation vector into a binary classifier, output the association relationship between the text representation vector and the successful repair strategy and the unsuccessful repair strategy.

[0108] In an example, the training module 4400 may also be used to input the unsuccessful fault repair strategies into the binary classifier, so that the binary classifier learns the success mode and the unsuccessful mode of different fault repair strategies.

[0109] The training device of the cluster fault repair model of this embodiment can be used to execute the technical solution of the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here.

[0110] <Server Implementation>

[0111] In this embodiment, a server is also provided, which includes the training device 4000 of the cluster fault recovery model described in the device embodiment of this specification; or, the server is Figure 5 The server 5000 shown includes:

[0112] The memory 5100 is used to store executable commands.

[0113] The processor 5200 is used to execute the method described in any method embodiment of this specification under the control of the executable instructions stored in the memory 5100.

[0114] <Computer Readable Storage Medium Embodiment>

[0115] This embodiment provides a computer-readable storage medium, in which executable commands are stored. When the executable commands are executed by a processor, the method described in any method embodiment of this specification is executed.

[0116] <Second Embodiment>

[0117] <Method>

[0118] This embodiment provides a cluster fault repair method. For example, the method can be performed as follows: Figure 2 The server 1000 shown executes.

[0119] like Figure 6 As shown, the method includes the following steps 6100-6200:

[0120] Step 6100, obtaining the fault data to be repaired.

[0121] Specifically, the to-be-repaired fault data may include information such as the name of the fault monitoring module, the fault level, and the repair strategy.

[0122] Step 6200: input the fault data to be repaired into the cluster fault repair model, and output the corresponding fault repair strategy.

[0123] Among them, the cluster fault repair model forms a fault text based on historical fault data, represents the fault text as a structured text representation vector, inputs the text representation vector into a multi-layer classifier for training, and obtains the correlation between the output and the historical fault repair strategy.

[0124] Furthermore, in one example, the server 1000 can also store the to-be-repaired fault data and the corresponding fault repair strategy in a preset fault knowledge base. Specifically, in this step, knowledge abstraction and expert confirmation are performed on the successfully repaired fault log data to form new repair rules, increase interpretability, and the accumulated fault knowledge base can also be used to analyze the root cause of the fault.

[0125] The method of this embodiment obtains the fault data to be repaired; inputs the fault data to be repaired into a pre-trained cluster fault repair model, and outputs the corresponding fault repair strategy. The cluster fault repair model forms a fault text based on historical fault data, represents the fault text as a structured text representation vector, and inputs the text representation vector into a multi-layer classifier for training, and outputs the correlation relationship between the text representation vector and the historical fault repair strategy. Thus, automatic repair of cluster faults with high efficiency and high accuracy is achieved.

[0126] <Device Example>

[0127] This embodiment provides a cluster fault recovery device, which is, for example, Figure 7 The cluster fault repairing device 7000 shown may include an acquisition module 7100 and an output module 7200 .

[0128] The acquisition module 7100 is used to acquire the fault data to be repaired.

[0129] The output module 7200 is used to input the fault data to be repaired into the cluster fault repair model and output a corresponding fault repair strategy.

[0130] In an example, the cluster fault repair device 7000 may further include a storage module, which is used to store the to-be-repaired fault data and the corresponding fault repair strategy in a preset fault knowledge base.

[0131] The cluster fault repair device of this embodiment can be used to execute the technical solution of the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here.

[0132] <Server Implementation>

[0133] In this embodiment, a server is further provided, the server comprising the cluster fault recovery device 7000 described in the device embodiment of this specification; or the server is Figure 8 The server 8000 shown includes:

[0134] Memory 8100 is used to store executable commands.

[0135] The processor 8200 is used to execute the method described in any method embodiment of this specification under the control of the executable instructions stored in the memory 8100.

[0136] <Computer Readable Storage Medium Embodiment>

[0137] This embodiment provides a computer-readable storage medium, in which executable commands are stored. When the executable commands are executed by a processor, the method described in any method embodiment of this specification is executed.

[0138] The embodiments of the present specification may be systems, methods and / or computer program products. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the embodiments of the present specification.

[0139] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0140] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0141] The computer program instructions for performing the operation of the embodiments of this specification may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, executed as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, by utilizing the state information of a computer-readable program instruction to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA) or a programmable logic array (PLA), the electronic circuit may execute a computer-readable program instruction, thereby realizing the various aspects of the embodiments of this specification.

[0142] Various aspects of the embodiments of the present specification are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present specification. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0143] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device for implementing the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0144] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0145] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the multiple embodiments of the embodiments of this specification. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of an instruction, and a part of a module, a program segment or an instruction contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart can be implemented by a dedicated hardware-based system that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that it is equivalent to implement it by hardware, implement it by software, and implement it by combining software and hardware.

[0146] The embodiments of the present specification have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the embodiments of the present specification is defined by the appended claims.

Claims

1. A training method for a cluster fault recovery model, comprising: Get fault log data; The fault log data includes historical fault data and historical fault repair strategies; Forming a fault text according to the historical fault data includes: extracting the fault monitoring module name and the corresponding fault level from the historical fault data; forming the fault text with the fault monitoring module name and the corresponding fault level; the historical fault repair strategy includes a successful repair strategy and an unsuccessful repair strategy; Representing the fault text as a structured text representation vector; Inputting the text representation vector into a multi-layer classifier, outputting the association relationship between the text representation vector and the historical fault repair strategy, and obtaining the cluster fault repair model, including: Inputting the text representation vector into a multi-classifier, and outputting the association relationship between the text representation vector and the successful repair strategy; And, the text representation vector is input into a binary classifier, and the correlation relationship between the text representation vector and the successful repair strategy and the unsuccessful repair strategy is output.

2. The method according to claim 1, wherein: The step of representing the fault text as a structured text representation vector comprises: According to a preset fault level weight value, the fault text is converted into a corresponding word frequency vector; The word frequency vector is input into a language model to obtain the structured text representation vector.

3. The method according to claim 2, wherein: The language model is a probabilistic statistical model or a deep learning language model; wherein the probabilistic statistical model includes a term frequency-inverse document frequency tf-idf model and a document topic generation model LDA; and the deep learning language model is a word2vec model.

4. A cluster fault repair method, comprising: Obtain the fault data to be repaired; Input the fault data to be repaired into the cluster fault repair model, and output a corresponding fault repair strategy; The cluster fault repair model extracts the fault monitoring module name and the corresponding fault level from the historical fault data; the fault monitoring module name and the corresponding fault level are formed into a fault text, the fault text is represented as a structured text representation vector, the text representation vector is input into a multi-classifier, and the correlation relationship between the text representation vector and the successful repair strategy is output; And, the text representation vector is input into a binary classifier, and the correlation relationship between the text representation vector and the successful repair strategy and the unsuccessful repair strategy is output.

5. The method according to claim 4, wherein: After the outputting of the corresponding fault repair strategy, the method further includes: The to-be-repaired fault data and the corresponding fault repair strategy are stored in a preset fault knowledge base.

6. A training device for a cluster fault recovery model, comprising: An acquisition module is used to obtain fault log data; The fault log data includes historical fault data and historical fault repair strategies; A forming module, used for forming a fault text according to the historical fault data, including: extracting the fault monitoring module name and the corresponding fault level from the historical fault data; forming the fault text with the fault monitoring module name and the corresponding fault level; the historical fault repair strategy includes a successful repair strategy and an unsuccessful repair strategy; A representation module, used for representing the fault text as a structured text representation vector; A training module is used to input the text representation vector into a multi-layer classifier, output the association relationship between the text representation vector and the historical fault repair strategy, and obtain the cluster fault repair model, including: inputting the text representation vector into a multi-classifier, outputting the association relationship between the text representation vector and the successful repair strategy; and inputting the text representation vector into a binary classifier, outputting the association relationship between the text representation vector and the successful repair strategy and the unsuccessful repair strategy.

7. A cluster fault recovery device, comprising: An acquisition module is used to acquire the fault data to be repaired; An output module, used for inputting the fault data to be repaired into the cluster fault repair model and outputting a corresponding fault repair strategy; The cluster fault repair model extracts the fault monitoring module name and the corresponding fault level from the historical fault data; the fault monitoring module name and the corresponding fault level are formed into a fault text, the fault text is represented as a structured text representation vector, the text representation vector is input into a multi-classifier, and the correlation relationship between the text representation vector and the successful repair strategy is output; And, the text representation vector is input into a binary classifier, and the correlation relationship between the text representation vector and the successful repair strategy and the unsuccessful repair strategy is output.

8. A server, comprising the training device for the cluster fault recovery model according to claim 6, or the server comprising: A memory for storing executable commands; A processor is used to execute the cluster fault repair model training method according to any one of claims 1 to 3 under the control of the executable command.

9. A server, comprising the cluster fault recovery device according to claim 7, or the server comprising: A memory for storing executable commands; The processor is used to execute the cluster fault repair method as claimed in claim 4 or 5 under the control of the executable command.

10. A computer-readable storage medium storing executable instructions, wherein when the executable instructions are executed by a processor, the cluster fault repair model training method according to any one of claims 1 to 3 is executed.

11. A computer-readable storage medium storing executable instructions, wherein when the executable instructions are executed by a processor, the cluster fault repair method according to claim 4 or 5 is executed.

Images