Hardware fault detection method and device, equipment and storage medium

Abstract

The invention relates to a hardware fault detection method and device, equipment and a storage medium. The method comprises the steps of obtaining hardware attribute information and hardware use information corresponding to each piece of hardware in a server; for each piece of hardware in the server, the following operations are executed: determining a performance index model corresponding to the hardware; determining a target performance index corresponding to the hardware according to the hardware use information corresponding to the hardware and a performance index model; according to the hardware attribute information, the hardware use information and the target performance index corresponding to the hardware, a first failure probability corresponding to the hardware is determined, and the first failure probability represents the failure probability of the hardware under the influence of the hardware attribute information, the hardware use information and the target performance index; and performing fault detection on the hardware according to the target performance index and the first failure probability. According to the method, manual troubleshooting is not needed when the server hardware fails, the troubleshooting efficiency of the server hardware faults is improved, and the running reliability of the server is ensured.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a hardware fault detection method, device, equipment and storage medium. Background Art

[0002] The business processing of cloud technology companies involves a large number of servers, but with the increasing number of servers, server hardware failures are inevitable. When server hardware fails, it will affect the normal processing of cloud technology companies' business.

[0003] Currently, when a server hardware failure occurs, there may be no log error, but the server performance has been affected. At this time, relevant personnel are required to locate the server hardware failure to determine the cause of the failure. However, there are many hardware in the server, and manual fault location of the server hardware is inefficient and has low accuracy. Summary of the invention

[0004] In view of this, in order to solve the above-mentioned technical problems or part of the technical problems, the embodiments of the present application provide a hardware fault detection method, device, equipment and storage medium.

[0005] In a first aspect, the present application provides a hardware fault detection method, comprising:

[0006] Obtain hardware attribute information and hardware usage information corresponding to each hardware in the server;

[0007] For each of the hardware in the server, perform the following operations: determine a performance indicator model corresponding to the hardware, where the performance indicator model is used to determine the performance indicator corresponding to the hardware;

[0008] Determining a target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model;

[0009] Determine, according to the hardware attribute information, the hardware usage information, and the target performance indicator corresponding to the hardware, a first failure probability corresponding to the hardware, wherein the first failure probability represents a probability that the hardware fails under the influence of the hardware attribute information, the hardware usage information, and the target performance indicator;

[0010] Performing fault detection on the hardware according to the target performance indicator and the first failure probability.

[0011] In an optional implementation, the hardware usage information includes the hardware usage time. When the hardware is a mechanical hard disk, the performance indicator model corresponding to the hardware includes a first indicator model and a second indicator model. The target performance indicator corresponding to the hardware includes a cumulative failure rate and an instantaneous failure rate.

[0012] Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes:

[0013] The hardware usage time is input into the first indicator model to obtain the cumulative failure rate corresponding to the hardware, and the first indicator model includes:

[0014]

[0015] The hardware usage time is input into the second indicator model to obtain the instantaneous failure rate corresponding to the hardware, and the second indicator model includes:

[0016]

[0017] In the above formula, F(t) represents the cumulative failure rate, t represents the usage time of the hardware, β1 represents the shape parameter, η1 represents the scale parameter, and e () represents an exponential function, h(t) represents the instantaneous failure rate, β1 and η1 are determined by optimizing the log-likelihood function of the first indicator model or the second indicator model using the hardware usage information through the maximum likelihood estimation method.

[0018] In an optional implementation, when the hardware is a solid-state hard disk, the performance indicator model corresponding to the hardware includes a third indicator model, and the target performance indicator corresponding to the hardware includes a non-failure rate;

[0019] Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes:

[0020] The hardware usage time is input into the third indicator model to obtain the non-failure rate corresponding to the hardware, and the third indicator model includes:

[0021] P(t)=e -λt

[0022] In the above formula, P(t) represents the failure rate, t represents the usage time of the hardware, and e () represents an exponential function, λ represents a failure rate parameter, and λ is determined by optimizing the log-likelihood function of the third indicator model using the hardware usage information through a maximum likelihood estimation method.

[0023] In an optional implementation, when the hardware is a central processing unit or a memory, the performance indicator model corresponding to the hardware includes a fourth indicator model, and the target performance indicator corresponding to the hardware includes a degradation amount;

[0024] Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes:

[0025] The hardware usage time is input into the fourth indicator model to obtain the degradation amount corresponding to the hardware, and the fourth indicator model includes:

[0026] D(t)=αt 2 +εt+γ

[0027] In the above formula, D(t) represents the degradation amount, t represents the usage time of the hardware, α represents the first regression parameter, ε represents the second regression parameter, γ represents the third regression parameter, and α, ε and γ are determined by optimizing the log-likelihood function of the fourth indicator model using the hardware usage information through the maximum likelihood estimation method.

[0028] In an optional implementation, determining the performance indicator model corresponding to the hardware includes:

[0029] When the hardware is a disk array card or a graphics card, obtaining a first quantity of the hardware in the server;

[0030] When the first number is at least two, determining a target connection relationship of the first number of hardware in the server;

[0031] The performance indicator model corresponding to the hardware is determined according to the target connection relationship.

[0032] In an optional implementation, when the hardware is a disk array card or a graphics card, the performance indicator model corresponding to the hardware includes a fifth indicator model, a sixth indicator model and a seventh indicator model, and the target performance indicator corresponding to the hardware includes system reliability;

[0033] Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes:

[0034] When the hardware is a disk array card or a graphics card, the hardware duration corresponding to each of the hardware in the server is input into the fifth indicator model to obtain the hardware reliability corresponding to the hardware;

[0035] When the target connection relationship is a series connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the sixth indicator model to obtain the system reliability corresponding to the hardware, and when the target connection relationship is a parallel connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the seventh indicator model to obtain the system reliability corresponding to the hardware;

[0036] Determining the system reliability as a target performance indicator corresponding to the hardware;

[0037] Wherein, the fifth indicator model includes:

[0038]

[0039] The sixth indicator model includes:

[0040]

[0041] The seventh indicator model includes:

[0042]

[0043] In the above formula, R(t) represents the hardware reliability, R 系统 (t) represents the system reliability, i represents the i-th hardware in the server, n represents the first number of the hardware in the server, R i (t) represents the hardware reliability of the i-th hardware, e () represents an exponential function, β2 represents a shape parameter, η2 represents a scale parameter, and β2 and η2 are determined by optimizing the log-likelihood function of the fifth indicator model using the hardware usage information through the maximum likelihood estimation method.

[0044] In an optional implementation, the hardware usage information includes hardware usage time, hardware damage parameters, hardware workload, hardware environment parameters and hardware maintenance parameters;

[0045] The determining, according to the hardware attribute information, the hardware usage information and the target performance indicator corresponding to the hardware, a first failure probability corresponding to the hardware includes:

[0046] Determine a first weight of the hardware attribute information corresponding to the hardware, a second weight corresponding to each parameter in the hardware usage information, and a third weight corresponding to the target performance indicator;

[0047] Determine a target weighted sum among the hardware attribute information, the first weight, each parameter in the hardware usage information, the second weight, the target performance indicator, and the third weight;

[0048] The weighted sum of the targets is determined as a first failure probability corresponding to the hardware.

[0049] In an optional implementation, the performing fault detection on the hardware according to the target performance indicator and the first failure probability includes:

[0050] Determine a first threshold corresponding to the target performance indicator and a second threshold corresponding to the first failure probability;

[0051] Determining a first comparison result between the target performance indicator and the first threshold and determining a second comparison result between the first failure probability and the second threshold;

[0052] Fault detection is performed on the hardware according to the first comparison result and the second comparison result.

[0053] In a second aspect, the present application provides a hardware fault detection device, comprising:

[0054] An acquisition module is used to acquire hardware attribute information and hardware usage information corresponding to each hardware in the server;

[0055] A determination module, used to determine, for each hardware in the server, a performance indicator model corresponding to the hardware, wherein the performance indicator model is used to determine the performance indicator corresponding to the hardware;

[0056] The determination module is further used to determine the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model;

[0057] The determination module is further configured to determine a first failure probability corresponding to the hardware according to the hardware attribute information, the hardware usage information, and the target performance indicator corresponding to the hardware, wherein the first failure probability represents the probability of failure of the hardware under the influence of the hardware attribute information, the hardware usage information, and the target performance indicator;

[0058] A detection module is used to perform fault detection on the hardware according to the target performance indicator and the first failure probability.

[0059] In a third aspect, the present application provides an electronic device, comprising: a processor and a memory, wherein the processor is used to execute a hardware fault detection program stored in the memory to implement the hardware fault detection method as described above.

[0060] In a fourth aspect, the present application further provides a storage medium, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the hardware fault detection method as described above.

[0061] The above-mentioned technical scheme provided by the embodiment of the present application has the following advantages compared with the prior art. The method provided by the embodiment of the present application includes: obtaining hardware attribute information and hardware usage information corresponding to each hardware in the server; for each hardware in the server, performing the following operations: determining a performance indicator model corresponding to the hardware, the performance indicator model is used to determine the performance indicator corresponding to the hardware; determining the target performance indicator corresponding to the hardware according to the hardware usage information and the performance indicator model corresponding to the hardware; determining the first failure probability corresponding to the hardware according to the hardware attribute information, hardware usage information and target performance indicator corresponding to the hardware, the first failure probability characterizing the probability of failure of the hardware under the influence of the hardware attribute information, hardware usage information and target performance indicator; performing fault detection on the hardware according to the target performance indicator and the first failure probability. Through the above method, this embodiment obtains the hardware attribute information and hardware usage information corresponding to each hardware in the server during the operation of the server, so as to determine the target performance indicator of the current hardware according to the hardware usage information and the performance indicator model corresponding to the hardware, thereby determining the failure probability of the hardware under the influence of the hardware attribute information, hardware usage information and the target performance indicator, and then realizing hardware fault detection according to the target performance indicator and failure probability, so that the server hardware can accurately locate the hardware fault when it is about to fail, and there is no need to manually troubleshoot when the server hardware fails, thereby improving the troubleshooting efficiency of the server hardware fault and ensuring the reliability of the server operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0064] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0065] Figure 1A flowchart of a hardware fault detection method provided in an embodiment of the present application;

[0066] Figure 2 A flowchart of another hardware fault detection method provided in an embodiment of the present application;

[0067] Figure 3 A flowchart of another hardware fault detection method provided in an embodiment of the present application;

[0068] Figure 4 A schematic diagram of the structure of a hardware fault detection device provided in an embodiment of the present application;

[0069] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0070] In the above attached figure:

[0071] 10. Acquisition module; 20. Determination module; 30. Detection module;

[0072] 500, electronic device; 501, processor; 502, memory; 5021, operating system; 5022, application; 503, user interface; 504, network interface; 505, bus system. DETAILED DESCRIPTION

[0073] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0074] The disclosure below provides many different embodiments or examples to implement different structures of the present invention. In order to simplify the disclosure of the present invention, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present invention. In addition, the present invention can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.

[0075] refer to Figure 1 , Figure 1 A schematic diagram of a hardware fault detection method provided in an embodiment of the present application. A hardware fault detection method provided in this embodiment includes the following steps:

[0076] S101: Obtain hardware attribute information and hardware usage information corresponding to each hardware in the server.

[0077] In this embodiment, the hardware in the server includes a mechanical hard disk (HDD), a solid-state disk (SSD), a central processing unit (CPU), a memory, a disk array card (RAID card) and a graphics card. The hardware attribute information includes the hardware brand, the hardware product model and the total number of hardware purchased, and the hardware usage information includes the hardware usage time, hardware damage parameters, hardware workload, hardware environment parameters and hardware maintenance parameters. The hardware damage parameters include the current total number of hardware damages and the current hardware damage time, the hardware environment parameters include the current hardware environment temperature and the current hardware environment humidity, the hardware maintenance parameters include the current average number of repairs per hardware, the hardware workload can be understood as the number of read and write operations of the current hardware per day, and the hardware usage time mentioned above is the current hardware usage time.

[0078] Among them, when it is necessary to perform fault detection on each hardware in the server, the relevant information of each hardware in the server is obtained from the server cluster where the server is located. After obtaining the relevant information of each hardware, the relevant information of each hardware obtained is cleaned up to obtain the hardware attribute information and hardware usage information corresponding to the hardware.

[0079] Specifically, the hardware attribute information also includes the slot where the hardware is located. When the hardware attribute information and hardware usage information corresponding to each hardware in the server are obtained in step S101, a fault log can also be obtained from the server cluster where the server is located to compare the fault log with the hardware attribute information corresponding to each hardware to determine the slot where the faulty hardware in the server is located.

[0080] More specifically, dmesg logs, message logs, and mcelog logs may be extracted from the server cluster where the server is located, so as to extract fault logs from the above logs. dmidecode logs may be extracted from the server cluster where the server is located, so as to extract hardware attribute information corresponding to each hardware from the above logs.

[0081] S102: For each hardware in the server, perform the following operations: determine a performance indicator model corresponding to the hardware.

[0082] In this embodiment, the performance indicator model is used to determine the performance indicators corresponding to the hardware. After obtaining the hardware attribute information and hardware usage information corresponding to each hardware in the server, the performance indicator model corresponding to each hardware can be determined from the first association relationship. The first association relationship stores multiple groups of correspondences between hardware and performance indicators. After obtaining the performance indicator model corresponding to the hardware, the performance indicator model can be used to determine the performance indicators of the corresponding hardware, thereby determining the operating status of the current hardware based on the determined performance indicators, and then determining whether the hardware is about to fail based on the operating status of the hardware, thereby realizing hardware fault detection. It should be noted that the performance indicator model corresponding to the hardware is selected based on the relevant characteristics of the hardware.

[0083] S103: Determine the target performance index corresponding to the hardware according to the hardware usage information and the performance index model corresponding to the hardware.

[0084] In this embodiment, the performance indicator model corresponding to the hardware is the relationship between the performance indicator corresponding to the hardware and the hardware usage information corresponding to the hardware. Therefore, after determining the performance indicator model corresponding to the hardware, the hardware usage information corresponding to the hardware can be input into the performance indicator model corresponding to the hardware, so that the performance indicator model outputs the target performance indicator corresponding to the hardware. After obtaining the target performance indicator corresponding to the hardware, it can be determined whether the hardware is about to fail according to the target performance indicator corresponding to the hardware, thereby realizing hardware fault detection.

[0085] Among them, in step S103, according to the hardware usage information and performance indicator model corresponding to the hardware, the target performance indicator corresponding to the hardware is determined, including:

[0086] Construct the log-likelihood function of the performance indicator model;

[0087] According to the hardware usage information corresponding to the hardware and the maximum likelihood estimation method, the likelihood logarithm function is optimized to determine the various parameters in the performance indicator model;

[0088] According to the hardware usage information, various parameters in the performance indicator model and the performance indicator model, the target performance indicator corresponding to the hardware is determined.

[0089] Specifically, in order to determine the target performance indicator corresponding to the hardware, there are parameters in the performance indicator model. Therefore, in order to be able to determine the target performance indicator based on the performance indicator model, it is necessary to determine the various parameters in the performance indicator model. When determining the various parameters in the performance indicator model, the likelihood logarithm function corresponding to the performance indicator model is constructed, and the likelihood logarithm function is optimized using the maximum likelihood estimation method and the hardware usage information corresponding to the hardware to determine the various parameters in the performance indicator model. The various parameters in the performance indicator model are substituted into the performance indicator model to complete the update of the performance indicator model, so that the hardware usage information is input into the updated performance indicator model, so that the performance indicator model outputs the target performance indicator corresponding to the hardware.

[0090] In one embodiment, when the hardware is a mechanical hard disk, the performance indicator model corresponding to the hardware includes a first indicator model and a second indicator model, and the target performance indicator corresponding to the hardware includes a cumulative failure rate and an instantaneous failure rate. When the hardware is a mechanical hard disk, the target performance indicator corresponding to the hardware is determined in the above step S103 according to the hardware usage information and the performance indicator model corresponding to the hardware, including:

[0091] The hardware usage time corresponding to the hardware is input into the first indicator model to obtain the cumulative failure rate corresponding to the hardware, and the hardware usage time corresponding to the hardware is input into the second indicator model to obtain the instantaneous failure rate corresponding to the hardware.

[0092] Among them, the first indicator model includes:

[0093]

[0094] The second indicator model includes:

[0095]

[0096] In the above formula, F(t) represents the cumulative failure rate, t represents the hardware usage time, β1 represents the shape parameter, η1 represents the scale parameter, and e () represents an exponential function, h(t) represents the instantaneous failure rate, and β1 and η1 are determined by optimizing the log-likelihood function of the first indicator model or the second indicator model using the hardware usage information through the maximum likelihood estimation method. Among them, β1 and η1 are determined by optimizing the log-likelihood function of the first indicator model or the second indicator model using the historical hardware damage time included in the hardware usage information through the maximum likelihood estimation method.

[0097] Specifically, the Weibull distribution can be used to establish a failure rate indicator model for mechanical hard disks, which can be specifically referred to as shown in the first indicator model and the second indicator model mentioned above. After obtaining the hardware attribute information and hardware usage information corresponding to the mechanical hard disk, the hardware attribute information and the hardware usage information are cleaned to check the missing values ​​and abnormal values ​​in the hardware attribute information and the hardware usage information, and perform corresponding processing (such as filling or deletion). After the hardware attribute information and the hardware usage information are cleaned, the hardware attribute information and the hardware usage information are normalized to standardize the numerical data to a specified range (such as 0-1) to eliminate the impact of different dimensions. After normalizing the hardware attribute information and the hardware usage information, the classification features (such as hardware brand and hardware product model) are converted into numerical codes, or new features are generated to improve model performance.

[0098] In order to determine the cumulative failure rate and instantaneous failure rate corresponding to the hardware, a likelihood logarithmic function for the first indicator model can be constructed, and then the likelihood logarithmic function can be optimized according to the maximum likelihood estimation method and the historical hardware damage times included in the hardware usage information to obtain the shape parameter β1 and scale parameter η1 in the first indicator model and the second indicator model.

[0099] The log-likelihood function in the above example can be expressed as:

[0100]

[0101] The objective function used by the maximum likelihood estimation method can be expressed as:

[0102]

[0103] In the above formula, n represents the number of historical hardware failure times, t i represents the i-th historical hardware damage time, β1 represents the shape parameter, η1 represents the scale parameter, and argmax() represents the function for finding the parameters of the logarithmic function of the likelihood.

[0104] It should be noted that the shape parameter and scale parameter that maximize the likelihood function can be determined by the gradient descent method and each historical hardware failure time, and the determined shape parameter and scale parameter are used as the shape parameter and scale parameter in the first indicator model and the second indicator model. Among them, the maximum likelihood estimation of each parameter in the Weibull distribution can refer to the existing technology, which will not be repeated in this embodiment.

[0105] More specifically, when determining the first indicator model and the second indicator model corresponding to the mechanical hard disk, the hardware usage time can be input into the first indicator model and the second indicator model respectively, so that the first indicator model outputs the cumulative failure rate corresponding to the hardware usage time and the second indicator model outputs the instantaneous failure rate corresponding to the hardware usage time.

[0106] In another embodiment, when the hardware is a solid-state drive, the performance indicator model corresponding to the hardware includes a third indicator model, and the target performance indicator corresponding to the hardware includes the non-failure rate. When the hardware is a solid-state drive, the target performance indicator corresponding to the hardware is determined in the above step S103 according to the hardware usage information and the performance indicator model corresponding to the hardware, including:

[0107] Input the hardware usage time into the third indicator model to obtain the non-failure rate corresponding to the hardware. The third indicator model includes:

[0108] P(t)=e -λt

[0109] In the above formula, P(t) represents the failure rate, t represents the hardware usage time, and e () represents an exponential function, λ represents a failure rate parameter, and λ is determined by optimizing the log-likelihood function of the third indicator model using the hardware usage information through the maximum likelihood estimation method. Among them, λ is determined by optimizing the log-likelihood function of the first indicator model or the second indicator model using the historical hardware failure time included in the hardware usage information through the maximum likelihood estimation method.

[0110] Specifically, the life of the solid-state drive is mainly limited by the number of writes, so an exponential decay model is used to determine the non-failure rate of the solid-state drive, so as to determine whether the solid-state drive is about to fail based on the non-failure rate. After obtaining the hardware attribute information and hardware usage information corresponding to the solid-state drive, the hardware attribute information and hardware usage information are cleaned, and the missing values ​​and abnormal values ​​in the hardware attribute information and hardware usage information are checked, and corresponding processing (such as filling or deletion) is performed. After the hardware attribute information and hardware usage information are cleaned, the hardware attribute information and hardware usage information are normalized to standardize the numerical data to a specified range (such as 0-1) to eliminate the impact of different dimensions. After the hardware attribute information and hardware usage information are normalized, the classification features (such as hardware brand and hardware product model) are converted into numerical codes, or new features are generated to improve model performance.

[0111] In order to determine the non-failure rate corresponding to the hardware, a likelihood logarithm function for the third indicator model may be constructed, and the likelihood logarithm function may be optimized according to each historical hardware damage time included in the hardware usage information to obtain the failure rate parameter λ in the third indicator model.

[0112] The log-likelihood function in the above example can be expressed as:

[0113]

[0114] The failure rate parameter λ formula can be expressed as:

[0115]

[0116] In the above formula, n represents the number of historical hardware failure times, t i represents the i-th historical hardware damage time, and λ represents the failure rate parameter.

[0117] It should be noted that the likelihood function can be differentiated and the reciprocal can be set equal to 0 to obtain the above failure rate parameter λ formula, and then all historical hardware damage times are substituted into the above failure rate parameter λ formula to obtain the failure rate parameter λ. Among them, the maximum likelihood estimation of the parameters in the exponential function can refer to the existing technology, which will not be repeated in this embodiment.

[0118] In another embodiment, when the hardware is a central processing unit or a memory, the performance indicator model corresponding to the hardware includes a fourth indicator model, and the target performance indicator corresponding to the hardware includes the degradation amount. When the hardware is a central processing unit or a memory, the target performance indicator corresponding to the hardware is determined in the above step S103 according to the hardware usage information and the performance indicator model corresponding to the hardware, including:

[0119] The hardware usage time is input into the fourth indicator model to obtain the corresponding degradation amount of the hardware. The fourth indicator model includes:

[0120] D(t)=αt 2 +εt+γ

[0121] In the above formula, D(t) represents the degradation amount, t represents the hardware usage time, α represents the first regression parameter, ε represents the second regression parameter, γ represents the third regression parameter, and α, ε and γ are determined by optimizing the log-likelihood function of the fourth indicator model using the hardware usage information through the maximum likelihood estimation method. Among them, α, ε and γ are determined by optimizing the log-likelihood function of the first indicator model or the second indicator model using the historical hardware damage time included in the hardware usage information through the maximum likelihood estimation method.

[0122] Specifically, the performance degradation of the central processing unit and memory is usually slow, and the quadratic regression model can be used to represent the amount of hardware degradation. When the hardware is a central processing unit or memory, after obtaining the hardware attribute information and hardware usage information corresponding to the mechanical hard disk, the hardware attribute information and hardware usage information are cleaned to check the missing values ​​and outliers in the hardware attribute information and hardware usage information, and perform corresponding processing (such as filling or deletion). After the hardware attribute information and hardware usage information are cleaned, the hardware attribute information and hardware usage information are normalized to standardize the numerical data to a specified range (such as 0-1) to eliminate the impact of different dimensions. After normalizing the hardware attribute information and hardware usage information, convert the classification features (such as hardware brand and hardware product model) into numerical codes, or generate new features to improve model performance.

[0123] In order to determine the degradation amount corresponding to the hardware, a likelihood logarithm function of the fourth indicator model can be constructed, and then the likelihood logarithm function can be optimized according to the maximum likelihood estimation method and the historical hardware damage times included in the hardware usage information to obtain the first regression parameter α, the second regression parameter ε and the third regression parameter γ in the fourth indicator model.

[0124] The log-likelihood function in the above example can be expressed as:

[0125]

[0126] The objective function used by the maximum likelihood estimation method can be expressed as:

[0127]

[0128] In the above formula, n represents the number of historical hardware failure times, t i represents the i-th historical hardware failure time, D i represents the historical degradation amount corresponding to the i-th historical hardware damage time, α represents the first regression parameter, ε represents the second regression parameter, γ represents the third regression parameter, argmax() represents the function for finding the parameters of the likelihood logarithm function, and σ represents the standard deviation.

[0129] It should be noted that the first regression parameter α, the second regression parameter ε, and the third regression parameter γ that maximize the likelihood function can be determined by the gradient descent method and each historical hardware damage time, and the determined first regression parameter, second regression parameter, and third regression parameter are used as the first regression parameter, second regression parameter, and third regression parameter in the fourth indicator model. Among them, the maximum likelihood estimation of each parameter in the quadratic regression model can refer to the existing technology, which is not repeated in this embodiment.

[0130] In another embodiment, when the hardware is a disk array or a graphics card, since there may be multiple disk arrays or graphics cards in the server, in order to improve the accuracy of hardware fault detection, hardware fault detection is performed through the connection relationship between each hardware.

[0131] Determining the performance indicator model corresponding to the hardware in the above step S102 includes:

[0132] When the hardware is a disk array card or a graphics card, obtaining a first quantity of the hardware in the server;

[0133] When the first number is at least two, determining a target connection relationship of the first number of hardware in the server;

[0134] According to the target connection relationship, determine the performance indicator model corresponding to the hardware.

[0135] The target connection relationship includes a series connection and a parallel connection. A second association relationship between the connection relationship and the performance indicator model can be preset, so that when the hardware is a disk array card or a graphics card, the performance indicator model corresponding to the target connection relationship can be determined based on the second association relationship, thereby realizing accurate fault detection of the hardware according to the determined performance indicator model.

[0136] In this embodiment, when the hardware is a disk array card or a graphics card, the performance indicator model corresponding to the hardware includes a fifth indicator model, a sixth indicator model and a seventh indicator model, and the target performance indicator corresponding to the hardware includes system reliability. In the above step S103, the target performance indicator corresponding to the hardware is determined according to the hardware usage information and the performance indicator model corresponding to the hardware, including:

[0137] When the hardware is a disk array card or a graphics card, the hardware duration corresponding to each hardware in the server is input into the fifth indicator model to obtain the hardware reliability corresponding to the hardware;

[0138] When the target connection relationship is a series connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the sixth indicator model to obtain the system reliability corresponding to the hardware, and when the target connection relationship is a parallel connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the seventh indicator model to obtain the system reliability corresponding to the hardware;

[0139] Determine system reliability as the target performance indicator corresponding to the hardware;

[0140] Among them, the fifth indicator model includes:

[0141]

[0142] The sixth indicator model includes:

[0143]

[0144] The seventh indicator model includes:

[0145]

[0146] In the above formula, R(t) represents the hardware reliability, R 系统 (t) represents the system reliability, i represents the i-th hardware in the server, n represents the first number of hardware in the server, R i (t) represents the hardware reliability of the i-th hardware, e () represents an exponential function, β2 represents a shape parameter, η2 represents a scale parameter, and β2 and η2 are determined by optimizing the log-likelihood function of the fifth indicator model using the hardware usage information through the maximum likelihood estimation method. β2 and η2 are determined by optimizing the log-likelihood function of the fifth indicator model using the historical hardware damage time included in the hardware usage information through the maximum likelihood estimation method.

[0147] Specifically, the Weibull distribution can be used to establish a reliability model for a single disk array card or graphics card. For details, please refer to the fifth indicator model mentioned above. When the hardware is a disk array card or graphics card, after obtaining the hardware attribute information and hardware usage information corresponding to the hardware, the hardware attribute information and hardware usage information are cleaned to check the missing values ​​and abnormal values ​​in the hardware attribute information and hardware usage information, and perform corresponding processing (such as filling or deletion). After the hardware attribute information and hardware usage information are cleaned, the hardware attribute information and hardware usage information are normalized to standardize the numerical data to a specified range (such as 0-1) to eliminate the impact of different dimensions. When the hardware attribute information and hardware usage information are normalized, the classification features (such as hardware brand and hardware product model) are converted into numerical codes, or new features are generated to improve model performance.

[0148] When the hardware is a disk array card or a graphics card, in order to determine the hardware reliability corresponding to the hardware, a likelihood logarithm function of the fifth indicator model can be constructed, and then the likelihood logarithm function is optimized according to the maximum likelihood estimation method and the historical hardware damage times included in the hardware usage information to obtain the shape parameter β2 and scale parameter η2 in the fifth indicator model.

[0149] The log-likelihood function in the above example can be expressed as:

[0150]

[0151] The objective function used by the maximum likelihood estimation method can be expressed as:

[0152]

[0153] In the above formula, n represents the number of historical hardware failure times, t i represents the i-th historical hardware damage time, β2 represents the shape parameter, η2 represents the scale parameter, and argmax() represents the function for finding the parameters of the logarithmic function of the likelihood.

[0154] It should be noted that the shape parameter and scale parameter that maximize the likelihood function can be determined by the gradient descent method and each historical hardware failure time, and the determined shape parameter and scale parameter are used as the shape parameter and scale parameter in the fifth indicator model. Among them, the maximum likelihood estimation of each parameter in the Weibull distribution can refer to the existing technology, which will not be repeated in this embodiment.

[0155] In this embodiment, when the hardware is a disk array card or a graphics card, the hardware reliability corresponding to a single hardware in the server can be determined by the hardware usage time and the fifth indicator model. For the same type of hardware, when the connection relationship between all hardware is a series relationship, the hardware reliability corresponding to all hardware can be substituted into the sixth indicator model to obtain the target performance indicator corresponding to the hardware; when the connection relationship between all hardware is a parallel relationship, the hardware reliability corresponding to all hardware can be substituted into the seventh indicator model to obtain the target performance indicator corresponding to the hardware.

[0156] S104: Determine a first failure probability corresponding to the hardware according to hardware attribute information, hardware usage information, and target performance indicators corresponding to the hardware.

[0157] In this embodiment, the first failure probability represents the probability of hardware failure under the influence of hardware attribute information, hardware usage information and target performance indicators. Since the hardware attribute information, hardware usage information and target performance indicators will have a certain impact on the hardware failure, on the basis of determining the target performance indicator corresponding to the hardware, in order to more accurately determine whether the hardware is about to fail, the first failure probability corresponding to the hardware is determined according to the hardware attribute information, hardware usage information and target performance indicator corresponding to the hardware, so that the hardware failure is detected together according to the determined target performance indicator corresponding to the hardware and the first failure probability, thereby improving the accuracy of hardware failure detection.

[0158] S105: Performing hardware fault detection according to the target performance indicator and the first failure probability.

[0159] In this embodiment, after obtaining the target performance indicator and the first failure probability corresponding to each hardware in the server, for each hardware in the server, the target performance indicator corresponding to the hardware can be compared with the corresponding threshold, and the first failure probability corresponding to the hardware can be compared with the corresponding threshold, so as to jointly determine whether the hardware is about to fail based on the comparison result of the target performance indicator and the comparison result of the first failure probability, thereby realizing hardware fault detection.

[0160] A hardware fault detection method provided in this embodiment obtains hardware attribute information and hardware usage information corresponding to each hardware in the server during the operation of the server, so as to determine the target performance indicator of the current hardware according to the hardware usage information and the performance indicator model corresponding to the hardware, thereby determining the failure probability of the hardware under the influence of the hardware attribute information, the hardware usage information and the target performance indicator, and then realizing hardware fault detection according to the target performance indicator and the failure probability, so that the server hardware can accurately locate the hardware fault when it is about to fail, without the need for manual troubleshooting when the server hardware fails, thereby improving the troubleshooting efficiency of the server hardware fault and ensuring the reliability of the server operation.

[0161] refer to Figure 2 , Figure 2 A flowchart of another hardware fault detection method provided in an embodiment of the present application. A hardware fault detection method provided in this embodiment includes the following steps:

[0162] S201: Obtain hardware attribute information and hardware usage information corresponding to each hardware in the server.

[0163] S202: For each hardware in the server, perform the following operations: determine a performance indicator model corresponding to the hardware.

[0164] S203: Determine the target performance indicator corresponding to the hardware according to the hardware usage information and the performance indicator model corresponding to the hardware.

[0165] Regarding the above-mentioned steps S201 to S203, step S201 is consistent with the above-mentioned step S101, step S202 is consistent with the above-mentioned step S102, and step S203 is consistent with the above-mentioned step S103. For details, please refer to the above-mentioned steps S101 to S103, which will not be repeated in this embodiment.

[0166] S204: Determine a first weight corresponding to hardware attribute information corresponding to the hardware, a second weight corresponding to each parameter in the hardware usage information, and a third weight corresponding to the target performance indicator.

[0167] S205: Determine a target weighted sum among the hardware attribute information, the first weight, the hardware usage information, the second weight, the target performance indicator, and the third weight.

[0168] S206: Determine the weighted sum of the targets as the first failure probability corresponding to the hardware.

[0169] For the above-mentioned steps S204 to S206, the hardware usage information includes the hardware usage time, hardware damage parameters, hardware workload, hardware environment parameters and hardware maintenance parameters. The hardware damage parameters include the total number of current hardware damages and the current hardware damage time, the hardware environment parameters include the current hardware environment temperature and the current hardware environment humidity, and the hardware maintenance parameters include the current average number of maintenance times for each hardware.

[0170] Among them, the second weight corresponding to each parameter in the hardware usage information can be understood as the second weight corresponding to the hardware usage time, the second weight corresponding to the current total number of hardware damages, the second weight corresponding to the current hardware damage time, the second weight corresponding to the current hardware environment temperature, the second weight corresponding to the current hardware environment humidity, the second weight corresponding to the hardware workload, and the second weight corresponding to the current average number of repairs per hardware. The second weight corresponding to each parameter in the hardware usage information mentioned above can be set according to actual needs, and is not specifically limited in this embodiment. The first weight and the third weight can also be set according to actual needs, and are not limited in this embodiment.

[0171] For example, let the hardware attribute information be A, the hardware usage information be B, and the target performance indicator be C, the first weight corresponding to A is w1, the second weight corresponding to B is w2, and the weight corresponding to C is w3, then the target weighted sum is equal to A*w1+B*w2+C*w3.

[0172] Specifically, after obtaining the target weighted sum, the target weighted sum is used as the first failure probability corresponding to the hardware, so as to perform hardware fault detection in combination with the target performance index corresponding to the hardware, so as to improve the accuracy of hardware fault detection.

[0173] S207: Determine a first threshold corresponding to the target performance indicator and a second threshold corresponding to the first failure probability.

[0174] S208: Determine a first comparison result between the target performance indicator and the first threshold value and determine a second comparison result between the first failure probability and the second threshold value.

[0175] S209: Perform hardware fault detection according to the first comparison result and the second comparison result.

[0176] For the above-mentioned steps S207 to S209, the first threshold value and the second threshold value can be set according to actual needs. In this embodiment, the specific values ​​of the first threshold value and the second threshold value are not limited. After obtaining the first comparison result and the second comparison result, if it is determined that the hardware meets the fault warning condition, a fault warning prompt is performed. Among them, the first comparison result includes meeting the fault warning condition and not meeting the fault warning condition, and the second comparison result includes meeting the fault warning condition and not meeting the fault warning condition.

[0177] Among them, when the hardware is a disk display card or a graphics card, after obtaining the target performance indicator, 1-target performance indicator is used to obtain the second failure probability corresponding to the hardware, and then the second failure probability is compared with the first threshold to obtain a first comparison result, and then according to the first comparison result and the second comparison result, the hardware is detected for faults.

[0178] Specifically, after performing fault detection on each hardware in the server to obtain the hardware detection results corresponding to each hardware in the server, each hardware detection result can be displayed on the target interface, and when a hardware detection result is a hardware abnormality, the hardware detection result of the hardware abnormality is highlighted in the target interface, so that relevant personnel can quickly locate the faulty hardware in the server and then quickly repair the hardware.

[0179] A hardware fault detection method provided in this embodiment obtains hardware attribute information and hardware usage information corresponding to each hardware in the server during the operation of the server, so as to determine the target performance indicator of the current hardware according to the hardware usage information and the performance indicator model corresponding to the hardware, thereby determining the failure probability of the hardware under the influence of the hardware attribute information, the hardware usage information and the target performance indicator, and then realizing hardware fault detection according to the target performance indicator and the failure probability, so that the server hardware can accurately locate the hardware fault when it is about to fail, without the need for manual troubleshooting when the server hardware fails, thereby improving the troubleshooting efficiency of the server hardware fault and ensuring the reliability of the server operation.

[0180] As an example, refer to Figure 4 , let’s take a closer look at the entire hardware fault detection process, as follows:

[0181] When it is necessary to detect the faults of various hardware in the server, extract the dmesg log, message log, mcelog log, and dmidecode log from the server cluster where the server is located, extract the current average number of repairs of each hardware from the equipment maintenance information library, and extract the hardware brand, hardware product model and total number of hardware purchased from the equipment production information. Extract the fault log from the dmesg log, message log and mcelog log, compare the fault log with the dmidecode log to determine the slot where the faulty hardware is located, and provide a fault prompt. The hardware usage time, the current total number of hardware damages, the current hardware damage time, the current hardware environment temperature, the current hardware environment humidity, and the number of read and write operations of the current hardware per day of each hardware are obtained from the dmidecode log. For each hardware, the hardware usage time corresponding to the hardware is input into the performance indicator model corresponding to the hardware, so that the performance indicator model outputs the target performance indicator corresponding to the hardware, and the first failure probability corresponding to the hardware is determined according to the hardware brand, hardware product model, the total number of hardware purchases, the hardware usage time, the current total number of hardware damages, the current hardware damage time, the current hardware environment temperature, the current hardware environment humidity, and the number of read and write operations of the current hardware per day, and then the hardware is fault detected according to the first failure probability and the target performance indicator. When it is determined that there is hardware that is about to fail in the server, it is displayed on the target interface of the corresponding maintenance platform.

[0182] Figure 5 A schematic diagram of the structure of a hardware fault detection device provided in an embodiment of the present application. A hardware fault detection device provided in an embodiment of the present application includes an acquisition module 10, a determination module 20 and a detection module 30. Among them, the acquisition module 10 is used to obtain hardware attribute information and hardware usage information corresponding to each hardware in the server; the determination module 20 is used to determine the performance indicator model corresponding to the hardware for each hardware in the server, and the performance indicator model is used to determine the performance indicator corresponding to the hardware; the determination module 20 is also used to determine the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model; the determination module 20 is also used to determine the first failure probability corresponding to the hardware according to the hardware attribute information corresponding to the hardware, the hardware usage information and the target performance indicator, and the first failure probability represents the probability of failure of the hardware under the influence of the hardware attribute information, the hardware usage information and the target performance indicator; the detection module 30 is used to perform fault detection on the hardware according to the target performance indicator and the first failure probability.

[0183] In this embodiment, the hardware usage information includes the hardware usage time. When the hardware is a mechanical hard disk, the performance indicator model corresponding to the hardware includes a first indicator model and a second indicator model. The target performance indicator corresponding to the hardware includes a cumulative failure rate and an instantaneous failure rate. The determination module 20 is also used to:

[0184] The hardware usage time is input into the first indicator model to obtain the cumulative failure rate corresponding to the hardware, and the first indicator model includes:

[0185]

[0186] The hardware usage time is input into the second indicator model to obtain the instantaneous failure rate corresponding to the hardware, and the second indicator model includes:

[0187]

[0188] In the above formula, F(t) represents the cumulative failure rate, t represents the usage time of the hardware, β1 represents the shape parameter, η1 represents the scale parameter, and e () represents an exponential function, h(t) represents the instantaneous failure rate, β1 and η1 are determined by optimizing the log-likelihood function of the first indicator model or the second indicator model using the hardware usage information through the maximum likelihood estimation method.

[0189] In this embodiment, when the hardware is a solid state drive, the performance indicator model corresponding to the hardware includes a third indicator model, and the target performance indicator corresponding to the hardware includes a non-failure rate; the determination module 20 is further used to:

[0190] The hardware usage time is input into the third indicator model to obtain the non-failure rate corresponding to the hardware, and the third indicator model includes:

[0191] P(t)=e -λt

[0192] In the above formula, P(t) represents the failure rate, t represents the usage time of the hardware, and e () represents an exponential function, λ represents a failure rate parameter, and λ is determined by optimizing the log-likelihood function of the third indicator model using the hardware usage information through a maximum likelihood estimation method.

[0193] In this embodiment, when the hardware is a central processing unit or a memory, the performance indicator model corresponding to the hardware includes a fourth indicator model, and the target performance indicator corresponding to the hardware includes a degradation amount; the determination module 20 is further used for:

[0194] The hardware usage time is input into the fourth indicator model to obtain the degradation amount corresponding to the hardware, and the fourth indicator model includes:

[0195] D(t)=αt 2 +εt+γ

[0196] In the above formula, D(t) represents the degradation amount, t represents the usage time of the hardware, α represents the first regression parameter, ε represents the second regression parameter, γ represents the third regression parameter, and α, ε and γ are determined by optimizing the log-likelihood function of the fourth indicator model using the hardware usage information through the maximum likelihood estimation method.

[0197] In this embodiment, the determination module 20 is further configured to:

[0198] When the hardware is a disk array card or a graphics card, obtaining a first quantity of the hardware in the server;

[0199] When the first number is at least two, determining a target connection relationship of the first number of hardware in the server;

[0200] The performance indicator model corresponding to the hardware is determined according to the target connection relationship.

[0201] In this embodiment, when the hardware is a disk array card or a graphics card, the performance indicator model corresponding to the hardware includes a fifth indicator model, a sixth indicator model and a seventh indicator model, and the target performance indicator corresponding to the hardware includes system reliability; the determination module 20 is further used to:

[0202] When the hardware is a disk array card or a graphics card, the hardware duration corresponding to each of the hardware in the server is input into the fifth indicator model to obtain the hardware reliability corresponding to the hardware;

[0203] When the target connection relationship is a series connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the sixth indicator model to obtain the system reliability corresponding to the hardware, and when the target connection relationship is a parallel connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the seventh indicator model to obtain the system reliability corresponding to the hardware;

[0204] Determining the system reliability as a target performance indicator corresponding to the hardware;

[0205] Wherein, the fifth indicator model includes:

[0206]

[0207] The sixth indicator model includes:

[0208]

[0209] The seventh indicator model includes:

[0210]

[0211] In the above formula, R(t) represents the hardware reliability, R 系统 (t) represents the system reliability, i represents the i-th hardware in the server, n represents the first number of the hardware in the server, R i (t) represents the hardware reliability of the i-th hardware, e () represents an exponential function, β2 represents a shape parameter, η2 represents a scale parameter, and β2 and η2 are determined by optimizing the log-likelihood function of the fifth indicator model using the hardware usage information through the maximum likelihood estimation method.

[0212] In this embodiment, the hardware usage information includes hardware usage time, hardware damage parameters, hardware workload, hardware environment parameters and hardware maintenance parameters; the determination module 20 is also used to:

[0213] Determine a first weight of the hardware attribute information corresponding to the hardware, a second weight corresponding to each parameter in the hardware usage information, and a third weight corresponding to the target performance indicator;

[0214] Determine a target weighted sum among the hardware attribute information, the first weight, each parameter in the hardware usage information, the second weight, the target performance indicator, and the third weight;

[0215] The weighted sum of the targets is determined as a first failure probability corresponding to the hardware.

[0216] In this embodiment, the detection module 30 is further used for:

[0217] Determine a first threshold corresponding to the target performance indicator and a second threshold corresponding to the first failure probability;

[0218] Determining a first comparison result between the target performance indicator and the first threshold and determining a second comparison result between the first failure probability and the second threshold;

[0219] Fault detection is performed on the hardware according to the first comparison result and the second comparison result.

[0220] The present embodiment provides a hardware fault detection device, which determines the target performance index of the current hardware according to the hardware usage information and the performance index model corresponding to the hardware by obtaining the hardware attribute information and the hardware usage information corresponding to each hardware in the server during the operation of the server, thereby determining the failure probability of the hardware under the influence of the hardware attribute information, the hardware usage information and the target performance index, and then realizing hardware fault detection according to the target performance index and the failure probability, so that the server hardware can accurately locate the hardware fault when it is about to fail, without the need for manual troubleshooting when the server hardware fails, thereby improving the troubleshooting efficiency of the server hardware fault and ensuring the reliability of the server operation.

[0221] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 5 The electronic device 500 shown includes: at least one processor 501, a memory 502, at least one network interface 504 and other user interfaces 503. The various components in the electronic device 500 are coupled together via a bus system 505. It is understood that the bus system 505 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 505 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, the bus system 505 is not shown in FIG. Figure 5 Various buses are labeled as bus system 505 .

[0222] The user interface 503 may include a display, a keyboard, or a pointing device (eg, a mouse, a trackball, a touch pad, or a touch screen).

[0223] It can be understood that the memory 502 in the embodiment of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct RAM bus random access memory (DRRAM). The memory 502 described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0224] In some implementations, the memory 502 stores the following elements, executable units or data structures, or a subset thereof, or an extended set thereof: an operating system 5021 and an application program 5022 .

[0225] Among them, the operating system 5021 includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic services and process hardware-based tasks. The application 5022 includes various application programs, such as a media player (Media Player), a browser (Browser), etc., which are used to implement various application services. The program for implementing the method of the embodiment of the present application can be included in the application 5022.

[0226] In the embodiment of the present application, by calling the program or instruction stored in the memory 502, specifically, the program or instruction stored in the application 5022, the processor 501 is used to execute the method steps provided by each method embodiment.

[0227] The method disclosed in the above embodiment of the present application can be applied to the processor 501, or implemented by the processor 501. The processor 501 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 501. The above processor 501 can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software units in the decoding processor can be executed. The software unit can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 502, and the processor 501 reads the information in the memory 502 and completes the steps of the above method in combination with its hardware.

[0228] It is understood that the embodiments described herein may be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit may be implemented in one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPDevice, DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), general purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in the present application, or a combination thereof.

[0229] For software implementation, the technology described herein can be implemented by a unit that performs the functions described herein. The software code can be stored in a memory and executed by a processor. The memory can be implemented in the processor or outside the processor.

[0230] The electronic device provided in this embodiment may be Figure 5 The electronic device shown in FIG. 1 may perform the following steps: Figure 1 to Figure 3 All steps of the hardware fault detection method in Figure 1 to Figure 3 For details, please refer to Figure 1 to Figure 3 For the sake of brevity, the relevant description is not repeated here.

[0231] The embodiment of the present application also provides a storage medium (computer-readable storage medium). The storage medium here stores one or more programs. The storage medium may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a read-only memory, a flash memory, a hard disk or a solid-state drive; the memory may also include a combination of the above-mentioned types of memory.

[0232] When one or more programs in the storage medium can be executed by one or more processors, the hardware fault detection method executed on the hardware fault detection device side can be implemented.

[0233] The processor is used to execute the hardware fault detection program stored in the memory to implement the following steps of the hardware fault detection method executed on the hardware fault device side.

[0234] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0235] It should be noted that the phrases "one implementation", "an embodiment", "an exemplary embodiment", "some embodiments", etc. mentioned in the specification indicate that the described embodiments may include certain features, structures or characteristics, but not every embodiment may include the certain features, structures or characteristics. In addition, such phrases do not necessarily refer to the same embodiment. In addition, when describing certain features, structures or characteristics in conjunction with an embodiment, it is within the knowledge of those skilled in the art to implement such features, structures or characteristics in conjunction with other embodiments, whether explicitly or not explicitly described.

[0236] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0237] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A hardware fault detection method, characterized in that: include: Obtain hardware attribute information and hardware usage information corresponding to each hardware in the server; For each of the hardware in the server, perform the following operations: determine a performance indicator model corresponding to the hardware, where the performance indicator model is used to determine the performance indicator corresponding to the hardware; Determining a target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model; Determine, according to the hardware attribute information, the hardware usage information, and the target performance indicator corresponding to the hardware, a first failure probability corresponding to the hardware, wherein the first failure probability represents a probability that the hardware fails under the influence of the hardware attribute information, the hardware usage information, and the target performance indicator; Performing fault detection on the hardware according to the target performance indicator and the first failure probability.

2. The method according to claim 1, characterized in that The hardware usage information includes the hardware usage time. When the hardware is a mechanical hard disk, the performance indicator model corresponding to the hardware includes a first indicator model and a second indicator model. The target performance indicator corresponding to the hardware includes a cumulative failure rate and an instantaneous failure rate. Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes: The hardware usage time is input into the first indicator model to obtain the cumulative failure rate corresponding to the hardware, and the first indicator model includes: The hardware usage time is input into the second indicator model to obtain the instantaneous failure rate corresponding to the hardware, and the second indicator model includes: In the above formula, F(t) represents the cumulative failure rate, t represents the usage time of the hardware, β1 represents the shape parameter, η1 represents the scale parameter, and e () represents an exponential function, h(t) represents the instantaneous failure rate, β1 and η1 are determined by optimizing the log-likelihood function of the first indicator model or the second indicator model using the hardware usage information through the maximum likelihood estimation method.

3. The method according to claim 2, characterized in that When the hardware is a solid state drive, the performance indicator model corresponding to the hardware includes a third indicator model, and the target performance indicator corresponding to the hardware includes a non-failure rate; Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes: The hardware usage time is input into the third indicator model to obtain the non-failure rate corresponding to the hardware, and the third indicator model includes: P(t)=e -λt In the above formula, P(t) represents the failure rate, t represents the usage time of the hardware, and e () represents an exponential function, λ represents a failure rate parameter, and λ is determined by optimizing the log-likelihood function of the third indicator model using the hardware usage information through a maximum likelihood estimation method.

4. The method according to claim 2, characterized in that: When the hardware is a central processing unit or a memory, the performance indicator model corresponding to the hardware includes a fourth indicator model, and the target performance indicator corresponding to the hardware includes a degradation amount; Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes: The hardware usage time is input into the fourth indicator model to obtain the degradation amount corresponding to the hardware, and the fourth indicator model includes: D(t)=αt 2 +εt+γ In the above formula, D(t) represents the degradation amount, t represents the usage time of the hardware, α represents the first regression parameter, ε represents the second regression parameter, γ represents the third regression parameter, and α, ε and γ are determined by optimizing the log-likelihood function of the fourth indicator model using the hardware usage information through the maximum likelihood estimation method.

5. The method according to claim 2, characterized in that: Determining the performance indicator model corresponding to the hardware includes: When the hardware is a disk array card or a graphics card, obtaining a first quantity of the hardware in the server; When the first number is at least two, determining a target connection relationship of the first number of hardware in the server; The performance indicator model corresponding to the hardware is determined according to the target connection relationship.

6. The method according to claim 5, characterized in that When the hardware is a disk array card or a graphics card, the performance indicator model corresponding to the hardware includes a fifth indicator model, a sixth indicator model and a seventh indicator model, and the target performance indicator corresponding to the hardware includes system reliability; Determining the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model includes: When the hardware is a disk array card or a graphics card, the hardware duration corresponding to each of the hardware in the server is input into the fifth indicator model to obtain the hardware reliability corresponding to the hardware; When the target connection relationship is a series connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the sixth indicator model to obtain the system reliability corresponding to the hardware, and when the target connection relationship is a parallel connection relationship, the hardware reliability corresponding to all the hardware in the server is input into the seventh indicator model to obtain the system reliability corresponding to the hardware; Determining the system reliability as a target performance indicator corresponding to the hardware; Wherein, the fifth indicator model includes: The sixth indicator model includes: The seventh indicator model includes: In the above formula, R(t) represents the hardware reliability, R 系统 (t) represents the system reliability, i represents the i-th hardware in the server, n represents the first number of the hardware in the server, R i (t) represents the hardware reliability of the i-th hardware, e () represents an exponential function, β2 represents a shape parameter, η2 represents a scale parameter, and β2 and η2 are determined by optimizing the log-likelihood function of the fifth indicator model using the hardware usage information through the maximum likelihood estimation method.

7. The method according to claim 1, characterized in that The hardware usage information includes hardware usage time, hardware damage parameters, hardware workload, hardware environment parameters and hardware maintenance parameters; The determining, according to the hardware attribute information, the hardware usage information and the target performance indicator corresponding to the hardware, a first failure probability corresponding to the hardware includes: Determine a first weight of the hardware attribute information corresponding to the hardware, a second weight corresponding to each parameter in the hardware usage information, and a third weight corresponding to the target performance indicator; Determine a target weighted sum among the hardware attribute information, the first weight, each parameter in the hardware usage information, the second weight, the target performance indicator, and the third weight; The weighted sum of the targets is determined as a first failure probability corresponding to the hardware.

8. The method according to claim 1, characterized in that The performing fault detection on the hardware according to the target performance indicator and the first failure probability includes: Determine a first threshold corresponding to the target performance indicator and a second threshold corresponding to the first failure probability; Determining a first comparison result between the target performance indicator and the first threshold and determining a second comparison result between the first failure probability and the second threshold; Fault detection is performed on the hardware according to the first comparison result and the second comparison result.

9. A hardware fault detection device, characterized in that: include: An acquisition module is used to acquire hardware attribute information and hardware usage information corresponding to each hardware in the server; A determination module, used to determine, for each hardware in the server, a performance indicator model corresponding to the hardware, wherein the performance indicator model is used to determine the performance indicator corresponding to the hardware; The determination module is further used to determine the target performance indicator corresponding to the hardware according to the hardware usage information corresponding to the hardware and the performance indicator model; The determination module is further configured to determine a first failure probability corresponding to the hardware according to the hardware attribute information, the hardware usage information, and the target performance indicator corresponding to the hardware, wherein the first failure probability represents the probability of failure of the hardware under the influence of the hardware attribute information, the hardware usage information, and the target performance indicator; A detection module is used to perform fault detection on the hardware according to the target performance indicator and the first failure probability.

10. An electronic device, characterized in that: include: A processor and a memory, wherein the processor is used to execute a hardware fault detection program stored in the memory to implement the hardware fault detection method according to any one of claims 1 to 8.

11. A storage medium, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the hardware fault detection method according to any one of claims 1 to 8.

Images