Hard disk life prediction method and device, equipment and storage medium

Abstract

The invention provides a hard disk life prediction method and device, equipment and a storage medium, and the method comprises the steps: determining a sampling period information data set according to a historical hard disk information data set and a historical hard disk life prediction value, inputting the sampling period information data set into a first recurrent neural network model, and obtaining a sampling period; according to the sampling period, hard disk hardware information, hard disk warning information in the log and hard disk performance monitoring information are collected, and a hard disk information data set is obtained; preprocessing the hard disk information data set to obtain a target acquisition data matrix; according to the historical acquisition data matrix and the historical hard disk life prediction value, training a second recurrent neural network model to obtain a target recurrent neural network model; and inputting the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value, and displaying the hard disk life prediction value.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a hard disk life prediction method, device, equipment and storage medium. Background Art

[0002] In order to prevent the end of life due to hard drive failure, major data centers use different methods to monitor the status of hard drives so that data can be backed up and hard drives replaced before failure, thereby reducing the impact on the system.

[0003] In traditional data centers, the most common method for predicting hard disk life is to use the hard disk's SMART (Self-Monitoring, Analysis and Reporting Technology) monitoring information to predict the hard disk's life based on some or part of the monitoring items. This method is simple to use, but the disadvantage is that the error is relatively large.

[0004] In the prior art, the following methods are usually used to predict the life of hard disks: SMART monitoring information is used as input, and the result is calculated by a data model as the life prediction value. The data model includes traditional machine learning methods, such as decision trees, Bayesian classification, and ordinary neural network algorithms. Compared with directly using SMART monitoring information for judgment, the advantage of these data models is that they relate various SMART monitoring information to reduce the factors of human judgment. The disadvantages are as follows: the input only contains SMART monitoring information, which is not sufficient, and there may be scenarios where SMART monitoring information is falsely reported; the mathematical model is not consistent, and traditional machine learning calculates the life result based on isolated monitoring information at a single moment, ignoring the importance of historical information.

[0005] Therefore, in view of the shortcomings of the prior art solutions, the present invention provides a hard disk life prediction method. Summary of the invention

[0006] Based on this, it is necessary to provide a hard disk life prediction method, device, equipment and storage medium to address the above technical problems.

[0007] On the one hand, a hard disk life prediction method is provided, the method comprising: determining a sampling period information data set based on a historical hard disk information data set and a historical hard disk life prediction value, inputting the sampling period information data set into a first recurrent neural network model to obtain a sampling period; according to the sampling period, collecting hard disk hardware information, hard disk warning information in a log, and hard disk performance monitoring information to obtain a hard disk information data set; preprocessing the hard disk information data set to obtain a target collection data matrix; training a second recurrent neural network model based on the historical collection data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; inputting the target collection data matrix into the target recurrent neural network model to obtain a hard disk life prediction value, and displaying the hard disk life prediction value.

[0008] Optionally, the method of determining a sampling period information data set based on a historical hard disk information data set and a historical hard disk life prediction value, and inputting the sampling period information data set into a first recurrent neural network model to obtain a sampling period includes: acquiring the historical hard disk information data set and the historical hard disk life prediction value to obtain a sampling period information data set; preprocessing the sampling period information data set to determine a sampling period data matrix; and inputting the sampling period data matrix into the first recurrent neural network model to obtain a sampling period.

[0009] Optionally, the preprocessing of the hard disk information data set to obtain a target acquisition data matrix includes: filtering each data in the hard disk information data set to obtain a filtered data set; and normalizing the filtered data set to obtain the target acquisition data matrix.

[0010] Optionally, the step of inputting the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value includes: obtaining a historical hidden layer data matrix through the target recurrent neural network model, wherein the target recurrent neural network model includes a first parameter matrix, a second parameter matrix and a third parameter matrix; multiplying the target acquisition data matrix with the first parameter matrix to obtain a first input data matrix; multiplying the historical hidden layer data matrix with the second parameter matrix to obtain a second input data matrix; adding the first input data matrix and the second input data matrix to obtain a third input data matrix; regularizing the third input data matrix through a first processing function to obtain a target input data matrix; processing the target input data matrix and the third parameter matrix through a second processing function to obtain an output data matrix; and determining the hard disk life prediction value according to the output data matrix.

[0011] Optionally, the target acquisition data matrix is ​​input into the target recurrent neural network model to obtain a hard disk life prediction value, including: creating a gated recurrent unit, training the gated recurrent unit, and obtaining parameters corresponding to the gated recurrent unit; and processing the historical hidden layer data matrix and the target acquisition data matrix through the gated recurrent unit to obtain a target input data matrix.

[0012] Optionally, determining the hard disk life prediction value based on the output data matrix includes: determining a maximum value of the output data matrix based on the output data matrix; and determining data corresponding to a position of the maximum value in the output data matrix as the hard disk life prediction value.

[0013] Optionally, the method further includes: presetting a life warning value, comparing the hard disk life prediction value with the life warning value; when the hard disk life prediction value is smaller than the life warning value, generating an alarm message, and sending the alarm message to a user.

[0014] On the other hand, a hard disk life prediction device is provided, which includes: a first processing module, which is used to determine a sampling period information data set based on a historical hard disk information data set and a historical hard disk life prediction value, and input the sampling period information data set into a first recurrent neural network model to obtain a sampling period; a collection module, which is used to collect hard disk hardware information, hard disk warning information in the log, and hard disk performance monitoring information according to the sampling period to obtain a hard disk information data set; a second processing module, which is used to pre-process the hard disk information data set to obtain a target collection data matrix; a training module, which is used to train a second recurrent neural network model based on the historical collection data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; a third processing module, which is used to input the target collection data matrix into the target recurrent neural network model to obtain a hard disk life prediction value, and display the hard disk life prediction value.

[0015] On the other hand, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: according to a sampling period, hard disk hardware information, hard disk warning information in a log, and hard disk performance monitoring information are collected to obtain a hard disk information data set; the hard disk information data set is preprocessed to obtain a target collection data matrix; a recurrent neural network model is trained according to a historical collection data matrix and historical hard disk life prediction values ​​to obtain a target recurrent neural network model; the target collection data matrix is ​​input into the target recurrent neural network model to obtain a hard disk life prediction value, and the hard disk life prediction value is displayed.

[0016] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: a sampling period information data set is determined based on a historical hard disk information data set and a historical hard disk life prediction value, and the sampling period information data set is input into a first recurrent neural network model to obtain a sampling period; based on the sampling period, hard disk hardware information, hard disk warning information in the log, and hard disk performance monitoring information are collected to obtain a hard disk information data set; the hard disk information data set is preprocessed to obtain a target collection data matrix; a second recurrent neural network model is trained based on the historical collection data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; the target collection data matrix is ​​input into the target recurrent neural network model to obtain a hard disk life prediction value, and the hard disk life prediction value is displayed.

[0017] The above-mentioned disk life prediction method, device, equipment and storage medium, the method includes: determining a sampling period information data set based on a historical hard disk information data set and a historical hard disk life prediction value, inputting the sampling period information data set into a first recurrent neural network model to obtain a sampling period; according to the sampling period, collecting hard disk hardware information, hard disk warning information in the log and hard disk performance monitoring information to obtain a hard disk information data set; preprocessing the hard disk information data set to obtain a target collection data matrix; training a second recurrent neural network model based on the historical collection data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; inputting the target collection data matrix into the target recurrent neural network model to obtain a hard disk life prediction value, and displaying the hard disk life prediction value; in this way, the data sampling period is dynamically adjusted through an adaptive algorithm to achieve a balance between reducing system pressure and improving result accuracy; using a recurrent neural network RNN ​​to predict the life of a hard disk can combine the current monitoring data and historical monitoring data of the hard disk to more accurately and comprehensively predict the life of the hard disk and improve the accuracy of the prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A diagram showing an application environment of a disk life prediction method in one embodiment;

[0019] Figure 2 A schematic diagram of a process flow of a disk life prediction method in one embodiment;

[0020] Figure 3 A schematic diagram of a recurrent neural network of a disk life prediction method in one embodiment;

[0021] Figure 4 A schematic diagram of a module of a disk life prediction method in one embodiment;

[0022] Figure 5A structural block diagram of a disk life prediction device in one embodiment;

[0023] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0025] It should be understood that in the description of the present application, unless the context clearly requires otherwise, words such as "include", "comprises", and the like throughout the specification should be interpreted as including rather than being exclusive or exhaustive; that is, as including but not limited to.

[0026] It should also be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, "plurality" means two or more.

[0027] It should be noted that the terms "S1", "S2", etc. are only used for the purpose of describing the steps, and do not specifically refer to the order or sequence, nor are they used to limit the present application. They are only for the convenience of describing the method of the present application, and cannot be understood as indicating the order of the steps. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present application.

[0028] The disk life prediction method provided in this application can be applied to Figure 1 In the application environment shown, the terminal 102 communicates with the server 104 through a network. The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices, and the server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0029] In one embodiment, Figure 2 As shown, a disk life prediction method is provided, comprising the following steps:

[0030] S201: Determine a sampling period information data set according to a historical hard disk information data set and a historical hard disk life prediction value, and input the sampling period information data set into a first recurrent neural network model to obtain a sampling period.

[0031] Here, the Recurrent Neural Network (RNN) model is a neural network model that takes sequence data as input, circulates in the direction of sequence evolution, and all nodes are connected in a chain, which can effectively process time-based state sequences.

[0032] The historical hard disk information data set and the historical hard disk life prediction value are the data collected last time and the life prediction value calculated last time.

[0033] S202: According to a sampling period, hard disk hardware information, hard disk warning information in the log, and hard disk performance monitoring information are collected to obtain a hard disk information data set.

[0034] Here, the hard disk hardware information is SMART monitoring data, which may include hard disk temperature, error rate, startup times, shutdown times, power-on time, total write volume, etc. The hard disk errors in the log may include read and write errors, media errors, etc. The hard disk performance monitoring information may include hard disk usage rate, hard disk busyness, etc.

[0035] S203: Preprocess the hard disk information data set to obtain a target acquisition data matrix.

[0036] Here, the target acquisition data matrix is ​​a 1*m dimensional matrix.

[0037] Specifically, the data in the hard disk information data set includes data of different dimensions. Through preprocessing, the data is unified into data of the same dimension to obtain a target acquisition data matrix.

[0038] S204: training a second recurrent neural network model according to the historically collected data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model.

[0039] Here, the historical collection data matrix is ​​the preprocessed hard disk information data set collected last time according to the sampling period, and the historical hard disk life prediction value is the life value predicted by the RNN model last time.

[0040] Furthermore, a training data set is prepared in advance, including a collection data matrix and life prediction values, and the RNN model is trained using BPTT (Back Propagation Through Time) to obtain an RNN model for predicting hard disk life.

[0041] Specifically, the neural network is relearned based on historical input data and historical output data to improve the adaptability of the neural network model.

[0042] S205: Input the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value, and display the hard disk life prediction value.

[0043] In one embodiment, the update frequency of the hard disk life prediction value can be determined according to the sampling period, and the displayed hard disk life prediction value can be updated according to the update frequency.

[0044] Specifically, exemplary, Figure 3 is a schematic diagram of a recurrent neural network in an embodiment of the present application, such as Figure 3 As shown in the figure, it consists of three parts, namely the input layer X, the hidden layer h and the output layer L. There is an arrow in the hidden layer to indicate the cyclic update of the data. The input layer has two parts: the input parameter X at time t and the hidden layer output h at time t-1. The input X at time t is the monitoring data collected at each moment (i.e., the target collection data matrix), and the output h at time t-1 represents the monitoring data in the historical process.

[0045] It should be noted that this application is based on the recurrent neural network (RNN) model. By regularly sampling the monitoring data indicators of the hard disk and combining the current monitoring data and historical monitoring data of the hard disk, it can more accurately and comprehensively predict the life of the physical hard disk, improve the accuracy of the life prediction results, improve the operation and maintenance efficiency of the data center, reduce the cost of the data center, and improve the stability of the system storage service and the security of the data.

[0046] In some specific implementations, determining a sampling period information data set based on a historical hard disk information data set and a historical hard disk life prediction value, and inputting the sampling period information data set into a first recurrent neural network model to obtain a sampling period includes:

[0047] Acquire the historical hard disk information data set and the historical hard disk life prediction value to obtain a sampling period information data set;

[0048] Preprocessing the sampling period information data set to determine a sampling period data matrix;

[0049] The sampling period data matrix is ​​input into the first recurrent neural network model to obtain the sampling period.

[0050] Furthermore, the RNN model corresponding to the sampling period can be continuously relearned according to the predicted output data and input data to improve the accuracy of the predicted sampling period.

[0051] In this way, the sampling period is dynamically adjusted according to the hard disk monitoring information and life prediction results. When the hard disk life is longer, the sampling period can be increased to reduce system pressure. When the hard disk life is shorter, the sampling period can be reduced to improve the accuracy of life prediction, thus achieving a balance between reducing system pressure and improving result accuracy.

[0052] In some specific implementations, the preprocessing of the hard disk information data set to obtain a target acquisition data matrix includes:

[0053] Filtering each data in the hard disk information data set to obtain a filtered data set;

[0054] The screened data set is normalized to obtain the target acquisition data matrix.

[0055] Here, filtering is used to filter out abnormal data, such as blank or garbled data; normalization is used to normalize data of different dimensions to data between [0,1].

[0056] Specifically, since the dimensions of the sampled data items are different, some monitoring items are particularly small, and some monitoring items are particularly large. Data with inconsistent dimensions will have negative consequences for the results of the calculation model. Therefore, the dimensions of the sampled data are unified and all normalized to values ​​between [0,1].

[0057] In this way, the dimensions of the sampling data are unified, improving the accuracy of the life prediction value.

[0058] In some specific implementations, inputting the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value includes:

[0059] Obtaining a historical hidden layer data matrix through the target recurrent neural network model, wherein the target recurrent neural network model includes a first parameter matrix, a second parameter matrix, and a third parameter matrix;

[0060] Multiplying the target acquisition data matrix by the first parameter matrix to obtain a first input data matrix;

[0061] Multiplying the historical hidden layer data matrix by the second parameter matrix to obtain a second input data matrix;

[0062] Adding the first input data matrix and the second input data matrix to obtain a third input data matrix;

[0063] Regularizing the third input data matrix by using a first processing function to obtain a target input data matrix;

[0064] Processing the target input data matrix and the third parameter matrix through a second processing function to obtain an output data matrix;

[0065] The hard disk life prediction value is determined according to the output data matrix.

[0066] Here, the target data matrix (i.e., input layer) is a 1*m-dimensional matrix Represents the input at time t.

[0067] Here, the hidden layer data matrix (i.e. the output of the hidden layer) is a 1*p-dimensional matrix It not only represents the output h at time t, but also represents the historical data of the hard disk at time t. It is used to memorize various parameters during the operation of the hardware, and is mainly used to reflect the impact of long-term operation on the life of the hard disk.

[0068] Here, the output data matrix (i.e., the output layer) is a 1*n dimensional matrix L 1,n , L 1,n =[l1,l2,…,l n ], indicating the service quality data at time t.

[0069] Among them, the first processing function is tanh, which is used to regularize the data to prevent overfitting, and the second processing function is the softmax function, which is used to convert a set of real numbers into a probability distribution so that each element of the output is between 0 and 1, and the sum of all elements is equal to 1.

[0070] Exemplarily, the RNN model calculation process includes: performing regularization processing on the input parameter X at time t and the hidden layer output h at time t-1, and the formula is as follows:

[0071]

[0072] Among them, tanh is the hyperbolic tangent function, which means that tanh calculation is performed for each data item of the matrix in brackets. It can regularize the output of the hidden layer. The formula is as follows:

[0073]

[0074] U is an m*p dimensional matrix, and the formula is as follows:

[0075]

[0076] V is a p*p dimensional matrix, and the formula is as follows:

[0077]

[0078] The hidden layer output at time t is calculated through the softmax function to obtain the output layer result. The formula is as follows:

[0079]

[0080] Among them, function f is the softmax function, parameter W is a p*n dimensional matrix, and the formula is as follows:

[0081]

[0082] Specifically, the input layer data at time t Multiply by the parameter matrix U m,p , the hidden layer data at time t-1 Multiply by the parameter matrix V p,p , combining the two and performing regularization, we get the hidden layer output at time t According to the parameter matrix W p,n And the softmax function, After processing, we get L 1,n , according to L 1,n Get the predicted value of hard disk life.

[0083] In this way, the hard disk life can be comprehensively judged and calculated by combining the monitoring information at the current moment and the monitoring information at the historical moment; by regularizing the data, the RNN model can be prevented from overfitting.

[0084] In some specific implementations, inputting the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value includes:

[0085] Creating a gated recurrent unit, training the gated recurrent unit, and obtaining parameters corresponding to the gated recurrent unit;

[0086] The historical hidden layer data matrix and the target acquisition data matrix are processed by the gated recurrent unit to obtain a target input data matrix.

[0087] Here, the parameters corresponding to the gated recurrent unit (GRU) may include a weight vector of an update gate, a bias parameter corresponding to the update gate, a weight vector of a reset gate, a bias parameter corresponding to the reset gate, and the like.

[0088] In this way, by adding gated recurrent units, the problem of gradient disappearance and explosion in the RNN model can be solved.

[0089] In some specific implementations, determining the hard disk life prediction value according to the output data matrix includes:

[0090] According to the output data matrix, determining the maximum value of the output data matrix;

[0091] According to the position of the maximum value in the output data matrix, the data corresponding to the position is determined as the hard disk life prediction value.

[0092] Here, the output data matrix includes probability values, and the sum of all data in the matrix is ​​1.

[0093] Specifically, a life value matrix is ​​preset, which is [1, 3, 5, 7, 15, 30, 90, 180, 360, N], where the unit of each number is day, indicating how many days the hard disk life is left. N means more than 1 year. The output data matrix is ​​[0, 0, 0, 0, 0.3, 0.6, 0.1, 0, 0, 0], where the maximum value is 0.6. It is determined that 0.6 corresponds to the sixth bit in the L matrix. The life value matrix is ​​queried to determine that the life value corresponding to the sixth bit is 30, and the hard disk life prediction value is 30.

[0094] In some embodiments, the method further comprises:

[0095] Preset a life warning value, and compare the hard disk life prediction value with the life warning value;

[0096] When the hard disk life prediction value is less than the life warning value, an alarm message is generated and sent to a user.

[0097] Specifically, hard disks that do not meet a certain lifespan value will be given an alarm to remind operation and maintenance personnel.

[0098] In this way, users can be reminded to replace the hard disk in time to prevent system failure.

[0099] In some specific implementations, the gated recurrent unit is used to process the historical hidden layer data matrix and the target acquisition data matrix to obtain a target input data matrix, including:

[0100] Determine the reset gate and update gate of the gated recurrent unit according to the parameters of the gated recurrent unit, the historical hidden layer data matrix and the target acquisition data matrix;

[0101] Candidate hidden states are calculated based on the reset gate and the target acquisition data matrix;

[0102] The final hidden state is determined by combining the update gate, the historical hidden layer data matrix, and the candidate state.

[0103] Among them, the reset gate is used to decide how to combine the new input with the information of the previous moment, and the update gate is used to decide how much past information to retain and how much new information to add.

[0104] In this way, by adding gated recurrent units, the dependencies of long sequences can be captured, the structure is simple, and it is easy to implement and adjust parameters.

[0105] In one embodiment, Figure 4 It is a schematic diagram of the modules in the embodiment of the present application, such as Figure 4 As shown, the modules in the present application include: a data sampling module, an adaptive module, a data preprocessing module, a calculation module, a life display module and a life alarm module.

[0106] Specifically, the data sampling module is used to periodically sample the monitoring information of the hard disk, which mainly includes the following: smart monitoring information; hard disk-related alarms and error information in the system log; and performance monitoring information of the physical hard disk.

[0107] Specifically, the adaptive module is used to dynamically adjust the sampling period through the RNN algorithm according to the monitoring information and life prediction results.

[0108] Specifically, the data preprocessing module is used to normalize the data obtained by the data sampling module as the input of the calculation module. Since the dimensions of the data items sampled by the data sampling module are different, some monitoring items are particularly small, and some monitoring items are particularly large, data with inconsistent dimensions will bring negative results to the results of the calculation module. Therefore, the data preprocessing module is used to unify the dimensions of the sampled data and normalize them all to values ​​between [0,1].

[0109] Specifically, the core of the calculation module is the recurrent neural network (RNN) model, which inputs the preprocessed data into the recurrent neural network to obtain the predicted life value of the hard disk. In addition, the data processing module will relearn the neural network based on the current input data and output data to improve the adaptability of the neural network model.

[0110] Specifically, the life display module is used to display the remaining life of the physical hard disk in real time, that is, the data sampling device samples data every x seconds, and the life display device updates the life prediction value every x seconds.

[0111] Specifically, the life alarm module is used to issue an alarm to a hard disk that does not meet a certain life value, so as to remind the operation and maintenance personnel.

[0112] It should be understood that although Figure 1-4 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1-4At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0113] In one embodiment, Figure 5 As shown, a hard disk life prediction device is provided, and the device includes: a first processing module 501, which is used to determine a sampling period information data set based on a historical hard disk information data set and a historical hard disk life prediction value, and input the sampling period information data set into a first recurrent neural network model to obtain a sampling period; a collection module 502, which is used to collect hard disk hardware information, hard disk warning information in a log, and hard disk performance monitoring information according to the sampling period to obtain a hard disk information data set; a second processing module 503, which is used to pre-process the hard disk information data set to obtain a target collection data matrix; a training module 504, which is used to train a second recurrent neural network model based on the historical collection data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; a third processing module 505, which is used to input the target collection data matrix into the target recurrent neural network model to obtain a hard disk life prediction value, and display the hard disk life prediction value.

[0114] As a preferred implementation mode, in the embodiment of the present application, the first processing module 501 is specifically used to: obtain the historical hard disk information data set and the historical hard disk life prediction value to obtain a sampling period information data set; pre-process the sampling period information data set to determine a sampling period data matrix; input the sampling period data matrix into the first recurrent neural network model to obtain a sampling period.

[0115] As a preferred implementation, in the embodiment of the present application, the second processing module 503 is specifically used to: filter each data in the hard disk information data set to obtain a filtered data set; normalize the filtered data set to obtain the target acquisition data matrix.

[0116] As a preferred implementation, in the embodiment of the present application, the third processing module 505 is specifically used to: obtain a historical hidden layer data matrix through the target recurrent neural network model, wherein the target recurrent neural network model includes a first parameter matrix, a second parameter matrix and a third parameter matrix; multiply the target acquisition data matrix with the first parameter matrix to obtain a first input data matrix; multiply the historical hidden layer data matrix with the second parameter matrix to obtain a second input data matrix; add the first input data matrix and the second input data matrix to obtain a third input data matrix; regularize the third input data matrix through the first processing function to obtain a target input data matrix; process the target input data matrix and the third parameter matrix through the second processing function to obtain an output data matrix; determine the hard disk life prediction value according to the output data matrix.

[0117] As a preferred implementation manner, in an embodiment of the present application, the third processing module 505 is also specifically used to: create a gated recurrent unit, train the gated recurrent unit, and obtain parameters corresponding to the gated recurrent unit; through the gated recurrent unit, process the historical hidden layer data matrix and the target acquisition data matrix to obtain a target input data matrix.

[0118] As a preferred implementation, in the embodiment of the present application, the third processing module 505 is also specifically used to: determine the maximum value of the output data matrix based on the output data matrix; and determine the data corresponding to the position based on the position of the maximum value in the output data matrix as a predicted value of hard disk life.

[0119] As a preferred implementation mode, in the embodiment of the present application, the device also includes: an alarm module, which is specifically used to: preset a life alarm value, compare the hard disk life prediction value and the life alarm value; when the hard disk life prediction value is less than the life alarm value, generate an alarm message, and send the alarm message to the user.

[0120] For the specific definition of the hard disk life prediction device, please refer to the definition of the hard disk life prediction method in the above text, which will not be repeated here. Each module in the above-mentioned hard disk life prediction device can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0121] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a hard disk life prediction method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse, etc.

[0122] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0123] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: S1: determining a sampling period information data set based on a historical hard disk information data set and a historical hard disk life prediction value, inputting the sampling period information data set into a first recurrent neural network model to obtain a sampling period; S2: collecting hard disk hardware information, hard disk warning information in a log, and hard disk performance monitoring information based on the sampling period to obtain a hard disk information data set; S3: preprocessing the hard disk information data set to obtain a target collection data matrix; S4: training a second recurrent neural network model based on the historical collection data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; S5: inputting the target collection data matrix into the target recurrent neural network model to obtain a hard disk life prediction value, and displaying the hard disk life prediction value.

[0124] In one embodiment, when the processor executes the computer program, the following steps are also implemented: acquiring the historical hard disk information data set and the historical hard disk life prediction value to obtain a sampling period information data set; preprocessing the sampling period information data set to determine a sampling period data matrix; inputting the sampling period data matrix into the first recurrent neural network model to obtain a sampling period.

[0125] In one embodiment, when the processor executes the computer program, the following steps are also implemented: filtering each data in the hard disk information data set to obtain a filtered data set; normalizing the filtered data set to obtain the target acquisition data matrix.

[0126] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining a historical hidden layer data matrix through the target recurrent neural network model, wherein the target recurrent neural network model includes a first parameter matrix, a second parameter matrix and a third parameter matrix; multiplying the target acquisition data matrix with the first parameter matrix to obtain a first input data matrix; multiplying the historical hidden layer data matrix with the second parameter matrix to obtain a second input data matrix; adding the first input data matrix and the second input data matrix to obtain a third input data matrix; regularizing the third input data matrix through a first processing function to obtain a target input data matrix; processing the target input data matrix and the third parameter matrix through a second processing function to obtain an output data matrix; and determining the hard disk life prediction value according to the output data matrix.

[0127] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining a historical hidden layer data matrix through the target recurrent neural network model, wherein the target recurrent neural network model includes a first parameter matrix, a second parameter matrix and a third parameter matrix; multiplying the target acquisition data matrix with the first parameter matrix to obtain a first input data matrix; multiplying the historical hidden layer data matrix with the second parameter matrix to obtain a second input data matrix; adding the first input data matrix and the second input data matrix to obtain a third input data matrix; regularizing the third input data matrix through a first processing function to obtain a target input data matrix; processing the target input data matrix and the third parameter matrix through a second processing function to obtain an output data matrix; and determining the hard disk life prediction value according to the output data matrix.

[0128] In one embodiment, when the processor executes the computer program, the following steps are also implemented: creating a gated recurrent unit, training the gated recurrent unit, and obtaining parameters corresponding to the gated recurrent unit; and processing the historical hidden layer data matrix and the target acquisition data matrix through the gated recurrent unit to obtain a target input data matrix.

[0129] In one embodiment, when the processor executes the computer program, the following steps are also implemented: based on the output data matrix, determining the maximum value of the output data matrix; based on the position of the maximum value in the output data matrix, determining the data corresponding to the position as the hard disk life prediction value.

[0130] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: S1: determining a sampling period information data set according to a historical hard disk information data set and a historical hard disk life prediction value, and inputting the sampling period information data set into a first recurrent neural network model to obtain a sampling period; S2: collecting hard disk hardware information, hard disk warning information in a log, and hard disk performance monitoring information according to the sampling period to obtain a hard disk information data set; S3: preprocessing the hard disk information data set to obtain a target collection data matrix;

[0131] S4: Based on the historical collection data matrix and the historical hard disk life prediction value, train a second recurrent neural network model to obtain a target recurrent neural network model; S5: input the target collection data matrix into the target recurrent neural network model to obtain the hard disk life prediction value, and display the hard disk life prediction value.

[0132] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: acquiring the historical hard disk information data set and the historical hard disk life prediction value to obtain a sampling period information data set; preprocessing the sampling period information data set to determine a sampling period data matrix; inputting the sampling period data matrix into the first recurrent neural network model to obtain a sampling period.

[0133] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: filtering each data in the hard disk information data set to obtain a filtered data set; normalizing the filtered data set to obtain the target acquisition data matrix.

[0134] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining a historical hidden layer data matrix through the target recurrent neural network model, wherein the target recurrent neural network model includes a first parameter matrix, a second parameter matrix and a third parameter matrix; multiplying the target acquisition data matrix with the first parameter matrix to obtain a first input data matrix; multiplying the historical hidden layer data matrix with the second parameter matrix to obtain a second input data matrix; adding the first input data matrix and the second input data matrix to obtain a third input data matrix; regularizing the third input data matrix through a first processing function to obtain a target input data matrix; processing the target input data matrix and the third parameter matrix through a second processing function to obtain an output data matrix; and determining the hard disk life prediction value according to the output data matrix.

[0135] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: creating a gated recurrent unit, training the gated recurrent unit, and obtaining parameters corresponding to the gated recurrent unit; and processing the historical hidden layer data matrix and the target acquisition data matrix through the gated recurrent unit to obtain a target input data matrix.

[0136] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: based on the output data matrix, determining the maximum value of the output data matrix; based on the position of the maximum value in the output data matrix, determining the data corresponding to the position as the hard disk life prediction value.

[0137] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: presetting a life warning value, comparing the hard disk life prediction value and the life warning value; when the hard disk life prediction value is less than the life warning value, generating an alarm message, and sending the alarm message to the user.

[0138] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0139] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0140] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A hard disk life prediction method, characterized in that: The method comprises: Determine a sampling period information data set according to a historical hard disk information data set and a historical hard disk life prediction value, and input the sampling period information data set into a first recurrent neural network model to obtain a sampling period; According to the sampling period, the hard disk hardware information, the hard disk warning information in the log and the hard disk performance monitoring information are collected to obtain the hard disk information data set; Preprocessing the hard disk information data set to obtain a target acquisition data matrix; Training a second recurrent neural network model according to the historically collected data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; The target acquisition data matrix is ​​input into the target recurrent neural network model to obtain a hard disk life prediction value, and the hard disk life prediction value is displayed.

2. The hard disk life prediction method according to claim 1, characterized in that: The method of determining a sampling period information data set according to a historical hard disk information data set and a historical hard disk life prediction value, and inputting the sampling period information data set into a first recurrent neural network model to obtain a sampling period includes: Acquire the historical hard disk information data set and the historical hard disk life prediction value to obtain a sampling period information data set; Preprocessing the sampling period information data set to determine a sampling period data matrix; The sampling period data matrix is ​​input into the first recurrent neural network model to obtain the sampling period.

3. The hard disk life prediction method according to claim 1, characterized in that: The preprocessing of the hard disk information data set to obtain a target acquisition data matrix includes: Filtering each data in the hard disk information data set to obtain a filtered data set; The screened data set is normalized to obtain the target acquisition data matrix.

4. The hard disk life prediction method according to claim 1, characterized in that: The step of inputting the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value includes: Obtaining a historical hidden layer data matrix through the target recurrent neural network model, wherein the target recurrent neural network model includes a first parameter matrix, a second parameter matrix, and a third parameter matrix; Multiplying the target acquisition data matrix by the first parameter matrix to obtain a first input data matrix; Multiplying the historical hidden layer data matrix by the second parameter matrix to obtain a second input data matrix; Adding the first input data matrix and the second input data matrix to obtain a third input data matrix; Regularizing the third input data matrix by using a first processing function to obtain a target input data matrix; Processing the target input data matrix and the third parameter matrix through a second processing function to obtain an output data matrix; The hard disk life prediction value is determined according to the output data matrix.

5. The hard disk life prediction method according to claim 4, characterized in that: The step of inputting the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value includes: Creating a gated recurrent unit, training the gated recurrent unit, and obtaining parameters corresponding to the gated recurrent unit; The historical hidden layer data matrix and the target acquisition data matrix are processed by the gated recurrent unit to obtain a target input data matrix.

6. The hard disk life prediction method according to claim 4, characterized in that: Determining the hard disk life prediction value according to the output data matrix includes: According to the output data matrix, determining the maximum value of the output data matrix; According to the position of the maximum value in the output data matrix, the data corresponding to the position is determined as the hard disk life prediction value.

7. The hard disk life prediction method according to claim 1, characterized in that: The method further comprises: Preset a life warning value, and compare the hard disk life prediction value with the life warning value; When the hard disk life prediction value is less than the life warning value, an alarm message is generated and sent to a user.

8. A hard disk life prediction device, characterized in that: The device comprises: A first processing module is used to determine a sampling period information data set according to a historical hard disk information data set and a historical hard disk life prediction value, and input the sampling period information data set into a first recurrent neural network model to obtain a sampling period; The acquisition module is used to collect hard disk hardware information, hard disk warning information in the log, and hard disk performance monitoring information according to the sampling period to obtain a hard disk information data set; A second processing module is used to pre-process the hard disk information data set to obtain a target acquisition data matrix; A training module, used for training a second recurrent neural network model according to the historical collection data matrix and the historical hard disk life prediction value to obtain a target recurrent neural network model; The third processing module is used to input the target acquisition data matrix into the target recurrent neural network model to obtain a hard disk life prediction value and display the hard disk life prediction value.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images