Hard disk abnormity real-time monitoring method and system

Abstract

The invention provides a hard disk exception real-time monitoring method and system, and relates to the technical field of storage monitoring, and the method comprises the steps: obtaining the operation data of a hard disk in real time through a preset monitoring system, and outputting the original data; preprocessing the original data by using a preset preprocessing method, and outputting to-be-analyzed data; performing feature extraction on the to-be-analyzed data, selecting historical operation data matched with the to-be-analyzed data from a historical database based on the extracted features, and training and optimizing a preset model based on the historical operation data to obtain an anomaly monitoring model; and performing exception analysis and prediction analysis on the to-be-analyzed data by using the exception monitoring model, outputting an exception-prediction analysis result, and obtaining an exception coping strategy matched with the exception-prediction analysis result in combination with a preset result-strategy comparison table. According to the method, the running state of the hard disk can be monitored and predicted in real time, potential hard disk problems can be found and processed in time, and the stability of the system and the safety of data are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of storage monitoring, and in particular to a method and system for real-time monitoring of hard disk abnormalities. Background Art

[0002] Hard disk abnormality monitoring is of great significance for maintaining system stability, ensuring data security, saving maintenance costs and extending hard disk life. Regularly monitoring hard disk operation data and handling abnormal situations in a timely manner are key steps to maintain the healthy operation of computer systems.

[0003] At present, hard disk monitoring usually adopts methods such as interface monitoring or sensor monitoring to monitor the hard disk. Although this traditional monitoring method can evaluate the hard disk status intuitively and quickly, its monitoring accuracy is low and it is difficult to find potential abnormal problems in the hard disk. As a result, it can often only find hard disk failures that have occurred, which may affect the normal use of the hard disk.

[0004] Therefore, the present invention provides a method and system for real-time monitoring of hard disk abnormalities. Summary of the invention

[0005] The present invention provides a hard disk abnormality real-time monitoring method and system, which are used to monitor the operating parameters of the hard disk in real time. Not only can the abnormal state of the hard disk be discovered in real time, but also the hard disk can be predicted and analyzed to timely discover potential abnormal conditions of the hard disk and ensure the stable operation of the hard disk.

[0006] In one aspect, the present invention provides a method for real-time monitoring of hard disk abnormalities, comprising: Step 1: Obtain the hard disk operation data in real time through the preset monitoring system and output the original data; Step 2: preprocessing the raw data using a preset preprocessing method, and outputting the data to be analyzed; Step 3: extracting features from the data to be analyzed, selecting historical operation data matching the data to be analyzed from a historical database based on the extracted features, and training and optimizing a preset model based on the historical operation data to obtain an abnormality monitoring model; Step 4: Use the anomaly monitoring model to perform anomaly analysis and predictive analysis on the data to be analyzed, output anomaly-prediction analysis results, and obtain anomaly response strategies that match the anomaly-prediction analysis results in combination with a preset result-strategy comparison table.

[0007] Preferably, step 1 includes: Obtain monitoring demand information of the current hard disk, and in combination with a preset demand-data comparison table, obtain real-time operation data in the current hard disk that matches the monitoring demand information; At the same time, the original data is output in combination with the device information of the current hard disk and the timestamp corresponding to the real-time operation data of the current hard disk.

[0008] Preferably, in step 2, before preprocessing the raw data using a preset preprocessing method, the method includes: Extracting features from the monitoring demand information, and constructing a demand feature set based on the extracted features; Based on the demand feature set and in combination with a preset feature-method matching table, a preset preprocessing method matching the monitoring demand information is selected from a method database.

[0009] Preferably, step 2 further includes: Matching the raw data with the preset preprocessing method, binding the raw data that meets the preset matching conditions with the preset preprocessing method, and outputting a data-method matching table; Based on the data-method matching table, the corresponding raw data is preprocessed using the preset preprocessing method to output the data to be analyzed.

[0010] Preferably, feature extraction is performed on the data to be analyzed, and historical operation data matching the data to be analyzed is selected from a historical database based on the extracted features, including: Extracting features from the data to be analyzed, and constructing a data feature set based on the extracted features; Based on the data feature set, historical operation data that meets preset screening conditions is selected from the historical database, and the hard disk historical operation data is output.

[0011] Preferably, the training and optimization of the preset model based on the historical operation data to obtain the abnormality monitoring model includes: Extracting features from the historical operation data, and constructing a historical data feature set based on the extracted features; Based on the historical data feature set and in combination with the feature-model mapping table, a matching first model is selected in the model database, and a model candidate pool is constructed based on at least two of the first models; The first model selected by the system according to the preset recommendation algorithm is obtained, and the system self-selection instruction is output. At the same time, the model selection instruction input manually is obtained through the preset IO port, and the manual selection instruction is output; Obtaining priority information corresponding to the system self-selected instructions and the manually selected instructions, and performing comparative analysis, and outputting the first model selected from the model candidate pool based on the comparative analysis result as a preset model; Obtain the data format requirements corresponding to the preset model, output the model format requirement information, and at the same time, obtain the data formats corresponding to all the data in the historical operation data, and output the data format information; Based on the model format requirement information and the data format information, a format comparison table is constructed, and a format conversion method matching the format comparison table is selected from a method database; Performing format conversion on the historical operation data based on the format conversion method to obtain first data matching the model format requirement information; Selecting a data partitioning method adapted to the monitoring demand information from a method database, and performing data partitioning on the first data using the data partitioning method to obtain a historical data training data set, a historical data test data set, and a historical data verification data set; The preset model is trained and optimized based on the historical data training data set, the historical data test data set and the historical data verification data set, and the preset model that meets the preset performance conditions is output as an abnormality monitoring model.

[0012] Preferably, step 4 includes: Based on the monitoring requirement information, corresponding evaluation indicators are selected from an indicator database to construct a first indicator set for real-time analysis of the hard disk and a second indicator set for predictive analysis of the hard disk; Based on the first indicator set, and using the anomaly monitoring model, the data to be analyzed is analyzed in real time, and a first analysis result is output; At the same time, based on the second indicator set, the data to be analyzed is predicted and analyzed by the anomaly monitoring model, and a second analysis result is output; Based on the first analysis result and according to a preset first coordinate system, a first change curve corresponding to each indicator in the first indicator set is constructed, and a first curve graph is constructed based on each of the first change curves; At the same time, based on the second analysis result and according to a preset second coordinate system, a second change curve corresponding to each indicator in the second indicator set is constructed, and a second curve graph is constructed based on each of the second change curves; Time-aligning the first curve graph and the second curve graph, and performing comparative analysis to obtain a real-time-prediction curve comparison graph; Based on the real-time-prediction curve comparison diagram, the curves under the same indicator in the first indicator set and the second indicator set are analyzed to obtain a first sub-result under each indicator. At the same time, combined with the preset indicator relationship diagram, the curves under different indicators in the first indicator set and the second indicator set are subjected to correlation analysis to obtain a second sub-result under each indicator. Performing an abnormal analysis on the first sub-result and the second sub-result in each time period in the real-time-prediction curve comparison diagram through the abnormal monitoring model, and outputting the abnormal analysis result;

[0013] in, Indicates the abnormal analysis result value corresponding to the tth period in the real-time-forecast curve comparison chart; It represents the total number of the same indicators in the first indicator set and the second indicator set in the tth period; It represents the total number of indicators that appear in the first indicator set and the second indicator set in the tth period; Indicates the data offset corresponding to the i-th indicator in the t-th time period; It represents the deviation standard control value corresponding to the ith indicator in the tth time period; represents the abnormal impact factor corresponding to the i-th indicator in the t-th period; Indicates the data offset corresponding to the jth indicator in the tth period; It represents the deviation standard control value corresponding to the jth indicator in the tth time period; Indicates the jth indicator in the tth period and other The correlation coefficients between the indicators; In combination with the preset result-strategy comparison table, an abnormality response strategy matching the abnormality analysis result is obtained, and the hard disk is regulated based on the abnormality response strategy.

[0014] On the other hand, the present invention also provides a hard disk abnormality real-time monitoring system, comprising: The data acquisition module is used to obtain the operation data of the hard disk in real time through a preset monitoring system and output the original data; A preprocessing module, used to preprocess the raw data using a preset preprocessing method and output the data to be analyzed; A model training module is used to extract features from the data to be analyzed, select historical operation data matching the data to be analyzed from a historical database based on the extracted features, and train and optimize a preset model based on the historical operation data to obtain an abnormality monitoring model; The anomaly analysis module is used to use the anomaly monitoring model to perform anomaly analysis and predictive analysis on the data to be analyzed, output anomaly-prediction analysis results, and obtain anomaly response strategies that match the anomaly-prediction analysis results in combination with a preset result-strategy comparison table.

[0015] The present invention provides a method and system for real-time monitoring of hard disk anomalies. The present invention obtains the operation data of the hard disk in real time through a pre-set monitoring system. After pre-processing and feature extraction, the data in the historical database is used to train an abnormal monitoring model, and abnormal analysis and predictive analysis are performed on the data to be analyzed. Finally, the abnormality-prediction analysis results are output, and the corresponding abnormal response strategy is determined in combination with the preset result-strategy comparison table. The present invention can realize real-time monitoring and prediction of the operation status of the hard disk, timely discover and deal with potential hard disk problems, and ensure the stability of the system and the security of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a flowchart of a method for real-time monitoring of hard disk abnormalities provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of the framework of a real-time monitoring system for hard disk abnormalities provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] like Figure 1 As shown, a method for real-time monitoring of hard disk abnormalities provided by an embodiment of the present invention includes: Step 1: Obtain the hard disk operation data in real time through the preset monitoring system and output the original data; Step 2: Use the preset preprocessing method to preprocess the raw data and output the data to be analyzed; Step 3: Extract features from the data to be analyzed, select historical operation data that matches the data to be analyzed from the historical database based on the extracted features, and train and optimize the preset model based on the historical operation data to obtain an abnormal monitoring model; Step 4: Use the anomaly monitoring model to perform anomaly analysis and predictive analysis on the data to be analyzed, output the anomaly-prediction analysis results, and combine the preset result-strategy comparison table to obtain the anomaly response strategy that matches the anomaly-prediction analysis results.

[0020] In this embodiment, the preset monitoring system: a system for monitoring hard disk operation data is preset; In this embodiment, raw data refers to unprocessed data obtained from the hard disk in real time, including temperature, read / write speed, SMART parameters, etc.; In this embodiment, the pre-processing method is preset: a method for pre-processing raw data to prepare the data for subsequent analysis, including data cleaning, data conversion, data normalization, etc.; In this embodiment, the data to be analyzed: data prepared for feature extraction and anomaly analysis after preprocessing; In this embodiment, historical database: a database storing past hard disk operation data, used for comparison and analysis with current data; In this embodiment, historical operation data: data selected from a historical database that matches the data to be analyzed, and used to train and optimize the anomaly monitoring model; In this embodiment, the preset model: a model trained and optimized based on historical operation data, used for anomaly monitoring; In this embodiment, the anomaly monitoring model: a model obtained by training according to a preset model, used for performing anomaly analysis and predictive analysis on the data to be analyzed; In this embodiment, abnormality-prediction analysis result: the result obtained after abnormality analysis and prediction of the data to be analyzed; In this embodiment, the preset result-strategy comparison table: a table including the mapping relationship between the anomaly-prediction analysis results and the anomaly response strategies; In this embodiment, abnormality response strategy refers to specific response measures or strategies formulated for different abnormal situations, which are used to handle hard disk abnormal situations.

[0021] The implementation principle and beneficial effects of this embodiment: The present invention obtains the operation data of the hard disk in real time through a pre-set monitoring system, and after pre-processing and feature extraction, uses the data in the historical database to train the abnormal monitoring model, performs abnormal analysis and predictive analysis on the data to be analyzed, and finally outputs the abnormal-predictive analysis results, and determines the corresponding abnormal response strategy in combination with the preset result-strategy comparison table. The present invention can realize real-time monitoring and prediction of the hard disk operation status, timely discover and deal with potential hard disk problems, and ensure the stability of the system and the security of data.

[0022] An embodiment of the present invention provides a method for real-time monitoring of hard disk abnormalities, wherein step 1 includes: Obtain the monitoring requirement information of the current hard disk, and combine it with the preset requirement-data comparison table to obtain the real-time operation data in the current hard disk that matches the monitoring requirement information; At the same time, the original data is output in combination with the device information of the current hard disk and the timestamp corresponding to the real-time operation data of the current hard disk.

[0023] In this embodiment, monitoring requirement information refers to specific information or parameters required for hard disk monitoring set by a user or system, such as temperature, read / write speed, SMART parameters, accuracy, response time, etc.; In this embodiment, the preset demand-data comparison table: a comparison table containing the corresponding relationship between monitoring demand information and data, used to determine the data to be acquired according to the monitoring demand information; In this embodiment, real-time operation data: unprocessed data obtained from the current hard disk, including data matching the monitoring requirement information; In this embodiment, the device information includes information about the hard disk device itself, such as the hard disk model, capacity, and interface type; In this embodiment, timestamp: a mark that records the time when data is generated or obtained, and is used to identify the time sequence or time series relationship of the data.

[0024] The implementation principle and beneficial effects of this embodiment: The present invention obtains monitoring demand information, obtains real-time operation data matching the monitoring demand according to the preset demand-data comparison table, and obtains the original data in combination with the device information and timestamp of the hard disk, thereby ensuring that data related to the monitoring demand is obtained from the hard disk, and the time information of the data can be tracked. The present invention can track and monitor the hard disk status in real time by monitoring the operation data of the hard disk in real time, combining the device information and timestamp, and can timely discover abnormal conditions of the hard disk, prevent data loss or hard disk failure, and improve the stability and reliability of the system.

[0025] In a method for real-time monitoring of hard disk abnormalities provided by an embodiment of the present invention, in step 2, before preprocessing the raw data using a preset preprocessing method, the method includes: Extract features from monitoring demand information, and construct a demand feature set based on the extracted features; Based on the demand feature set and in combination with the preset feature-method matching table, a preset preprocessing method matching the monitoring demand information is selected from the method database.

[0026] In this embodiment, the demand feature set is a feature set extracted from the monitoring demand information, used to describe the characteristics of the hard disk monitoring demand; In this embodiment, the preset feature-method matching table: a table containing mapping relationships between demand features and preset preprocessing methods, so as to select a preset preprocessing method that matches the monitoring demand information; In this embodiment, the method database includes a database of various preset preprocessing methods, and is used to store the processing logic, algorithms or processes of different methods.

[0027] The implementation principle and beneficial effects of this embodiment: The present invention first extracts features from the monitoring demand information and constructs a demand feature set. Then, based on the demand feature set and in combination with the preset feature-method matching table, a preset preprocessing method that matches the monitoring demand information is selected from the method database. In this way, according to the characteristics of the hard disk monitoring requirements, a suitable preprocessing method can be selected to process the raw data, providing preparation for subsequent analysis and modeling. By preprocessing the raw data using a preset preprocessing method, the present invention can improve the quality and availability of the data and provide a more reliable data basis for subsequent abnormal analysis and prediction.

[0028] The embodiment of the present invention provides a method for real-time monitoring of hard disk abnormalities, wherein step 2 further includes: Matching the original data with the preset preprocessing method, binding the original data that meets the preset matching conditions with the preset preprocessing method, and outputting a data-method matching table; Based on the data-method matching table, the corresponding raw data are preprocessed using the preset preprocessing method to output the data to be analyzed.

[0029] In this embodiment, the preset matching condition: a condition for determining whether the original data matches the preset preprocessing method, which may be a specific data attribute, feature or other rule for screening a suitable preprocessing method; In this embodiment, the data-method matching table is used to record the matching relationship between the original data and the preset preprocessing method, and to bind the original data and the preset preprocessing method that meet the preset matching conditions to facilitate subsequent data processing and analysis.

[0030] The implementation principle and beneficial effects of this embodiment: The present invention matches the original data with the preset preprocessing method, determines whether the matching condition is met according to the preset matching condition, binds the original data that meets the condition with the preset preprocessing method, and outputs the data-method matching table, and then preprocesses the corresponding original data using the preset preprocessing method according to the data-method matching table, and outputs the data to be analyzed. The present invention can ensure that the appropriate preprocessing method is used for the original data, making the data more analyzable and usable. At the same time, by matching the original data and the preset preprocessing method, the most appropriate processing method can be selected according to the actual situation, thereby improving the efficiency and accuracy of data processing.

[0031] An embodiment of the present invention provides a method for real-time monitoring of hard disk anomalies, which extracts features from data to be analyzed, and selects historical operation data matching the data to be analyzed from a historical database based on the extracted features, including: Extract features from the data to be analyzed, and construct a data feature set based on the extracted features; Based on the data feature set, historical operation data that meets the preset screening conditions is selected from the historical database, and the hard disk historical operation data is output.

[0032] In this embodiment, data feature set: a feature set extracted from the data to be analyzed, used to describe the features of the data to be analyzed, which may be the operating status, performance indicators or other relevant information of the hard disk; In this embodiment, the preset screening conditions: the conditions used to screen the historical operation data may be specific data features, attributes or other rules to ensure that the selected historical data matches the data to be analyzed; In this embodiment, the hard disk historical operation data refers to the operation data record of the hard disk in the past, including various operation information of the hard disk in different time periods, such as temperature, read / write speed, fault records, etc.

[0033] The implementation principle and beneficial effects of this embodiment: The present invention first extracts features from the data to be analyzed, constructs a data feature set, and then selects historical operation data that meets preset screening conditions from the historical database based on the data feature set, and outputs the hard disk historical operation data. By analyzing the historical operation data, the present invention can have a more comprehensive understanding of the operation of the hard disk, thereby better evaluating the status and potential problems of the current data to be analyzed. At the same time, selecting historical operation data that matches the data to be analyzed helps to establish a more accurate model and prediction, and improve the accuracy and efficiency of abnormal monitoring.

[0034] An embodiment of the present invention provides a method for real-time monitoring of hard disk anomalies, which trains and optimizes a preset model based on historical operation data to obtain an anomaly monitoring model, including: Extract features from historical operation data, and construct a historical data feature set based on the extracted features; Based on the historical data feature set and in combination with the feature-model mapping table, a matching first model is selected in the model database, and a model candidate pool is constructed based on at least two first models; The first model selected by the system according to the preset recommendation algorithm is obtained, and the system self-selection instruction is output. At the same time, the model selection instruction input manually is obtained through the preset IO port, and the manual selection instruction is output; Obtaining priority information corresponding to the system self-selected instructions and the manually selected instructions, and performing comparative analysis, and outputting the first model selected from the model candidate pool based on the comparative analysis results as the preset model; Obtain the data format requirements corresponding to the preset model, output the model format requirement information, and at the same time, obtain the data format corresponding to all data in the historical operation data, and output the data format information; Based on the model format requirement information and the data format information, a format comparison table is constructed, and a format conversion method matching the format comparison table is selected from a method database; Performing format conversion on the historical operation data based on the format conversion method to obtain first data matching the model format requirement information; Selecting a data partitioning method adapted to the monitoring demand information in a method database, and performing data partitioning on the first data using the data partitioning method to obtain a historical data training data set, a historical data test data set, and a historical data verification data set; The preset model is trained and optimized based on the historical data training data set, the historical data test data set and the historical data verification data set, and the preset model that meets the preset performance conditions is output as an anomaly monitoring model.

[0035] In this embodiment, historical data feature set: a feature set extracted from historical operation data, used to describe the features of the historical data, which may be the operation status, performance indicators or other related information of the hard disk; In this embodiment, feature-model mapping table: a table containing mapping relationships between data features and models, so as to select a suitable model for training and optimization to obtain an abnormality monitoring model; In this embodiment, the model database: a database containing various models, used to store different types of models; In this embodiment, the first model: a model selected from a model database, used to construct a model candidate pool; In this embodiment, the model candidate pool is a set of candidate models consisting of at least two first models; In this embodiment, the preset recommendation algorithm is a recommendation algorithm used for the system to select a model automatically, and can be selected based on the historical data feature set and performance conditions; In this embodiment, the system self-selection instruction is an instruction of the first model selected by the system according to a preset recommendation algorithm; In this embodiment, the preset IO port is: an input and output port for receiving a model selection instruction input manually; In this embodiment, the manual selection instruction: the model selection instruction input manually obtained through the preset IO port; In this embodiment, priority information: priority information corresponding to the system self-selected instructions and the manual selected instructions, used to determine the model finally selected; In this embodiment, the data format requirement: is the data format requirement corresponding to the preset model; In this embodiment, the data format information: is the data format information corresponding to all data in the historical operation data; In this embodiment, the format comparison table is a table used to compare the format requirements of the preset model for data with the actual format of the historical data so as to perform format conversion; In this embodiment, the format conversion method: according to the information in the format comparison table, select an appropriate method to convert the data format to meet the data format requirements of the preset model; In this embodiment, the first data: data that meets the preset model data format requirements after format conversion; In this embodiment, the data partitioning method is used to divide the first data into a historical data training data set, a historical data test data set and a historical data verification data set, so as to be used for training, testing and verifying a preset model; In this embodiment, the historical data training data set, the historical data test data set, and the historical data verification data set are data sets used to train, test, and verify the preset model, respectively. In this embodiment, preset performance conditions: pre-set model performance indicators or conditions, used to evaluate whether the trained model meets the requirements, for example, response time, calculation accuracy, etc.

[0036] The implementation principle and beneficial effects of this embodiment: The present invention extracts features from historical operation data, constructs a historical data feature set based on the features, selects an adaptive model from the model database, constructs a model candidate pool, and selects the first model according to a preset recommendation algorithm. The system selects the first model according to the preset recommendation algorithm and outputs a self-selection instruction, while receiving a manual selection instruction, and determines the final preset model through comparative analysis. The present invention can select a suitable model for training and optimization based on the characteristics of historical data and model requirements, and finally output an abnormality monitoring model that meets performance conditions, which can improve the accuracy and efficiency of abnormality monitoring, thereby timely discovering abnormal conditions of the hard disk and ensuring the stability of the system and the security of the data.

[0037] An embodiment of the present invention provides a method for real-time monitoring of hard disk abnormalities, wherein step 4 includes: Based on the monitoring demand information, corresponding evaluation indicators are selected from the indicator database to construct a first indicator set for real-time analysis of the hard disk and a second indicator set for predictive analysis of the hard disk; Based on the first indicator set, the data to be analyzed is analyzed in real time through the abnormal monitoring model, and a first analysis result is output; At the same time, based on the second indicator set, the data to be analyzed is predicted and analyzed through the abnormal monitoring model, and the second analysis result is output; Based on the first analysis result and according to a preset first coordinate system, a first change curve corresponding to each indicator in the first indicator set is constructed, and a first curve graph is constructed based on each first change curve; At the same time, based on the second analysis result and according to the preset second coordinate system, a second change curve corresponding to each indicator in the second indicator set is constructed, and a second curve graph is constructed based on each second change curve; Time-aligning the first curve graph and the second curve graph, and performing comparative analysis to obtain a real-time-prediction curve comparison graph; Based on the real-time-prediction curve comparison chart, the curves under the same indicator in the first indicator set and the second indicator set are analyzed to obtain the first sub-result under each indicator. At the same time, combined with the preset indicator relationship chart, the curves under different indicators in the first indicator set and the second indicator set are subjected to correlation analysis to obtain the second sub-result under each indicator. The abnormal monitoring model is used to perform abnormal analysis on the first sub-result and the second sub-result in each time period in the real-time-prediction curve comparison diagram, and the abnormal analysis results are output;

[0038] in, Indicates the abnormal analysis result value corresponding to the tth period in the real-time-forecast curve comparison chart; It represents the total number of the same indicators in the first indicator set and the second indicator set in the tth period; It represents the total number of indicators that appear in the first indicator set and the second indicator set in the tth period; Indicates the data offset corresponding to the i-th indicator in the t-th time period; It represents the deviation standard control value corresponding to the ith indicator in the tth time period; represents the abnormal impact factor corresponding to the i-th indicator in the t-th period; Indicates the data offset corresponding to the jth indicator in the tth period; It represents the deviation standard control value corresponding to the jth indicator in the tth time period; Indicates the jth indicator in the tth period and other The correlation coefficients between the indicators; Combined with the preset result-strategy comparison table, obtain the abnormal response strategy that matches the abnormal analysis result, and adjust the hard disk based on the abnormal response strategy.

[0039] In this embodiment, the indicator database: includes a set of evaluation indicators for real-time analysis and predictive analysis of the hard disk; In this embodiment, the evaluation index: an index selected from the index database for real-time analysis and predictive analysis of the hard disk, such as transmission speed, seek time, bad track rate, read-write consistency, etc.; In this embodiment, the first indicator set: a set of evaluation indicators used for real-time analysis; In this embodiment, the second indicator set: a set of evaluation indicators used for predictive analysis; In this embodiment, the first analysis result: that is, the real-time analysis result; In this embodiment, the second analysis result: that is, the prediction analysis result; In this embodiment, a first coordinate system is preset: a coordinate system for constructing a first change curve; In this embodiment, a second coordinate system is preset: a coordinate system for constructing a second variation curve; In this embodiment, the first change curve is a curve constructed according to the data changes corresponding to each indicator in the first indicator set, for example, a real-time curve of hard disk reading speed; In this embodiment, the second change curve is a curve constructed according to the change of the predicted data corresponding to each indicator in the second indicator set, for example, a hard disk read speed prediction curve, etc.; In this embodiment, the first curve graph: that is, a curve graph obtained according to each first change curve; In this embodiment, the second curve graph: that is, a curve graph obtained according to each second variation curve; In this embodiment, time alignment: aligning the time axes of different data curves to facilitate effective comparative analysis; In this embodiment, the real-time-prediction curve comparison chart: displays a comparison chart of the real-time analysis and prediction analysis results to help users understand the real-time status and future prediction of the hard disk performance; In this embodiment, the first sub-result: the result obtained by analyzing the curve under the same indicator; In this embodiment, the second sub-result: the result obtained by correlation analysis under different indicators; In this embodiment, the abnormal analysis result: the result obtained by performing abnormal analysis on the first sub-result and the second sub-result in each time period in the real-time-prediction curve comparison diagram, is used to indicate possible abnormal conditions of the hard disk.

[0040] The implementation principle and beneficial effects of this embodiment: The present invention selects evaluation indicators according to monitoring demand information, constructs the first and second indicator sets, performs real-time analysis and predictive analysis, and generates corresponding analysis results. According to these results, a change curve is constructed, and time alignment and comparative analysis are performed to obtain a real-time-predictive curve comparison chart. The abnormal analysis results are calculated by formula, and the abnormal response strategy is obtained according to the preset result-strategy comparison table, and finally the hard disk is regulated. The present invention can timely detect hard disk abnormalities and take corresponding measures by integrating real-time analysis and predictive analysis results, combining abnormal analysis and abnormal response strategies, thereby improving the accuracy and real-time performance of hard disk abnormality monitoring, thereby ensuring system stability and data security.

[0041] like Figure 2As shown, a hard disk abnormality real-time monitoring system provided by an embodiment of the present invention includes: The data acquisition module is used to obtain the operation data of the hard disk in real time through a preset monitoring system and output the original data; A preprocessing module is used to preprocess the raw data using a preset preprocessing method and output the data to be analyzed; The model training module is used to extract features from the data to be analyzed, select historical operation data that matches the data to be analyzed from the historical database based on the extracted features, and train and optimize the preset model based on the historical operation data to obtain an abnormal monitoring model; The anomaly analysis module is used to use the anomaly monitoring model to perform anomaly analysis and predictive analysis on the data to be analyzed, output the anomaly-prediction analysis results, and obtain the anomaly response strategy that matches the anomaly-prediction analysis results in combination with the preset result-strategy comparison table.

[0042] The implementation principle and beneficial effects of this embodiment: The present invention obtains the operation data of the hard disk in real time through a pre-set monitoring system, and after pre-processing and feature extraction, uses the data in the historical database to train the abnormal monitoring model, performs abnormal analysis and predictive analysis on the data to be analyzed, and finally outputs the abnormal-predictive analysis results, and determines the corresponding abnormal response strategy in combination with the preset result-strategy comparison table. The present invention can realize real-time monitoring and prediction of the hard disk operation status, timely discover and deal with potential hard disk problems, and ensure the stability of the system and the security of data.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention 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 make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for real-time monitoring of hard disk abnormalities, characterized in that: include: Step 1: Obtain the hard disk operation data in real time through the preset monitoring system and output the original data; Step 2: preprocessing the raw data using a preset preprocessing method, and outputting the data to be analyzed; Step 3: extracting features from the data to be analyzed, selecting historical operation data matching the data to be analyzed from a historical database based on the extracted features, and training and optimizing a preset model based on the historical operation data to obtain an abnormality monitoring model; Step 4: Use the anomaly monitoring model to perform anomaly analysis and predictive analysis on the data to be analyzed, output anomaly-prediction analysis results, and obtain anomaly response strategies that match the anomaly-prediction analysis results in combination with a preset result-strategy comparison table.

2. The method for real-time monitoring of hard disk abnormalities according to claim 1, characterized in that: Step 1 includes: Obtain monitoring demand information of the current hard disk, and in combination with a preset demand-data comparison table, obtain real-time operation data in the current hard disk that matches the monitoring demand information; At the same time, the original data is output in combination with the device information of the current hard disk and the timestamp corresponding to the real-time operation data of the current hard disk.

3. The method for real-time monitoring of hard disk abnormality according to claim 2, characterized in that: In step 2, before the raw data is preprocessed using a preset preprocessing method, the following steps are included: Extracting features from the monitoring demand information, and constructing a demand feature set based on the extracted features; Based on the demand feature set and in combination with a preset feature-method matching table, a preset preprocessing method matching the monitoring demand information is selected from a method database.

4. The method for real-time monitoring of hard disk abnormalities according to claim 1, characterized in that: Step 2 also includes: Matching the raw data with the preset preprocessing method, binding the raw data that meets the preset matching conditions with the preset preprocessing method, and outputting a data-method matching table; Based on the data-method matching table, the corresponding raw data is preprocessed using the preset preprocessing method to output the data to be analyzed.

5. The method for real-time monitoring of hard disk abnormalities according to claim 1, characterized in that: Extracting features from the data to be analyzed, and selecting historical operation data matching the data to be analyzed from a historical database based on the extracted features, including: Extracting features from the data to be analyzed, and constructing a data feature set based on the extracted features; Based on the data feature set, historical operation data that meets preset screening conditions is selected from the historical database, and the hard disk historical operation data is output.

6. The method for real-time monitoring of hard disk abnormality according to claim 2, characterized in that: The training and optimization of the preset model based on the historical operation data to obtain the abnormality monitoring model includes: Extracting features from the historical operation data, and constructing a historical data feature set based on the extracted features; Based on the historical data feature set and in combination with the feature-model mapping table, a matching first model is selected in the model database, and a model candidate pool is constructed based on at least two of the first models; The first model selected by the system according to the preset recommendation algorithm is obtained, and the system self-selection instruction is output. At the same time, the model selection instruction input manually is obtained through the preset IO port, and the manual selection instruction is output; Obtaining priority information corresponding to the system self-selected instructions and the manually selected instructions, and performing comparative analysis, and outputting the first model selected from the model candidate pool based on the comparative analysis result as a preset model; Obtain the data format requirements corresponding to the preset model, output the model format requirement information, and at the same time, obtain the data formats corresponding to all the data in the historical operation data, and output the data format information; Based on the model format requirement information and the data format information, a format comparison table is constructed, and a format conversion method matching the format comparison table is selected from a method database; Performing format conversion on the historical operation data based on the format conversion method to obtain first data matching the model format requirement information; Selecting a data partitioning method adapted to the monitoring demand information from a method database, and performing data partitioning on the first data using the data partitioning method to obtain a historical data training data set, a historical data test data set, and a historical data verification data set; The preset model is trained and optimized based on the historical data training data set, the historical data test data set and the historical data verification data set, and the preset model that meets the preset performance conditions is output as an abnormality monitoring model.

7. The method for real-time monitoring of hard disk abnormality according to claim 2, characterized in that: Step 4 includes: Based on the monitoring requirement information, corresponding evaluation indicators are selected from an indicator database to construct a first indicator set for real-time analysis of the hard disk and a second indicator set for predictive analysis of the hard disk; Based on the first indicator set, and using the anomaly monitoring model, the data to be analyzed is analyzed in real time, and a first analysis result is output; At the same time, based on the second indicator set, the data to be analyzed is predicted and analyzed by the anomaly monitoring model, and a second analysis result is output; Based on the first analysis result and according to a preset first coordinate system, a first change curve corresponding to each indicator in the first indicator set is constructed, and a first curve graph is constructed based on each of the first change curves; At the same time, based on the second analysis result and according to a preset second coordinate system, a second change curve corresponding to each indicator in the second indicator set is constructed, and a second curve graph is constructed based on each of the second change curves; Time-aligning the first curve graph and the second curve graph, and performing comparative analysis to obtain a real-time-prediction curve comparison graph; Based on the real-time-prediction curve comparison diagram, the curves under the same indicator in the first indicator set and the second indicator set are analyzed to obtain a first sub-result under each indicator. At the same time, combined with the preset indicator relationship diagram, the curves under different indicators in the first indicator set and the second indicator set are subjected to correlation analysis to obtain a second sub-result under each indicator. Performing an abnormal analysis on the first sub-result and the second sub-result in each time period in the real-time-prediction curve comparison diagram through the abnormal monitoring model, and outputting the abnormal analysis result; in, Indicates the abnormal analysis result value corresponding to the tth period in the real-time-forecast curve comparison chart; It represents the total number of the same indicators in the first indicator set and the second indicator set in the tth period; It represents the total number of indicators that appear in the first indicator set and the second indicator set in the tth period; Indicates the data offset corresponding to the i-th indicator in the t-th time period; It represents the deviation standard control value corresponding to the ith indicator in the tth time period; represents the abnormal impact factor corresponding to the i-th indicator in the t-th period; Indicates the data offset corresponding to the jth indicator in the tth period; It represents the deviation standard control value corresponding to the jth indicator in the tth time period; Indicates the jth indicator in the tth period and other The correlation coefficients between the indicators; In combination with the preset result-strategy comparison table, an abnormality response strategy matching the abnormality analysis result is obtained, and the hard disk is regulated based on the abnormality response strategy.

8. A real-time monitoring system for hard disk abnormality, characterized in that: include: The data acquisition module is used to obtain the operation data of the hard disk in real time through a preset monitoring system and output the original data; A preprocessing module, used to preprocess the raw data using a preset preprocessing method and output the data to be analyzed; A model training module is used to extract features from the data to be analyzed, select historical operation data matching the data to be analyzed from a historical database based on the extracted features, and train and optimize a preset model based on the historical operation data to obtain an abnormality monitoring model; The anomaly analysis module is used to use the anomaly monitoring model to perform anomaly analysis and predictive analysis on the data to be analyzed, output anomaly-prediction analysis results, and obtain anomaly response strategies that match the anomaly-prediction analysis results in combination with a preset result-strategy comparison table.

Images