One-to-many server testing method based on detection technology integration

Abstract

The invention discloses a one-to-many server testing method based on detection technology integration, and relates to the field of server performance testing, and the method comprises the steps: classifying historical fault data to obtain at least one data detection set; obtaining at least one difference feature set; calculating a correlation coefficient between the fault reason and difference features in the difference feature set, and determining a correlation feature set in the difference feature set; forming an identification mechanism of fault causes of the server types, and obtaining a detection technical scheme of the server types; obtaining an integrated detection packet, and obtaining a fault reason of the to-be-detected server; and using a fault maintenance scheme corresponding to the fault reason of the to-be-tested server to perform maintenance. By forming the difference feature set, calculating the correlation coefficient between the fault reason and the difference feature in the difference feature set and forming the recognition mechanism of the fault reason of the server type, parallel detection of a plurality of servers can be performed, and the detection precision can also be ensured.

Description

Technical Field

[0001] The present invention relates to the field of server performance testing, and more particularly to a one-to-many server testing method based on the integration of detection technologies. Background Art

[0002] A server is a specific IT device that provides computing power and runs software applications in a network environment. It provides computing or application services for other client machines in the network. Generally, a server has the ability to undertake response service requests, undertake services, and ensure services. As an electronic device, the internal structure of a server is very complex. The main components of a server are: CPU, memory, chipset, I / O devices, memory, peripheral devices, voltage regulator, power supply, and cooling system.

[0003] In order to enable the server to work properly in various daily states, the server needs to be tested. Traditional server testing is carried out one by one manually, with low testing efficiency. Moreover, there are different models of servers, and the testing methods and testing standards for different models of servers will all change. Using the one-to-one method for testing is prone to errors and the efficiency needs to be improved. Summary of the Invention

[0004] To solve the above technical problems, a one-to-many server testing method based on the integration of detection technologies is provided. This technical solution solves the problems in the above background art that traditional server testing is carried out one by one manually, with low testing efficiency, and there are different models of servers, and the testing methods and testing standards for different models of servers will all change. Using the one-to-one method for testing is prone to errors and the efficiency needs to be improved.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A one-to-many server testing method based on the integration of detection technologies, comprising:

[0007] Obtaining historical failure data of at least one server, and classifying the historical failure data based on the server types to obtain at least one data detection set, where the data detection set corresponds to the server types;

[0008] Performing differential feature extraction on the data detection set to obtain at least one differential feature set, where the differential feature set corresponds to the data detection set and the differential feature set is composed of at least one differential feature;

[0009] Obtain at least one failure cause of the server, calculate the correlation coefficient between the failure cause and the differential features in the differential feature set, and based on the correlation coefficient, determine the associated feature set in the differential feature set. The associated feature set corresponds to the failure cause and is composed of differential features;

[0010] Based on the associated feature set, form an identification mechanism for the failure causes of the server types, and summarize the identification mechanisms for at least one failure cause of the server types to obtain a detection technical solution for the server types;

[0011] Obtain the detection technical solution corresponding to the server type, integrate at least one detection technical solution to obtain an integrated detection package, and use the integrated detection package to detect at least one server to be tested to obtain the failure cause of the server to be tested;

[0012] Perform repairs using the failure repair plan corresponding to the failure cause of the server to be tested.

[0013] Preferably, the step of classifying historical failure data into at least one data detection set based on the server type includes the following steps:

[0014] Summarize the historical failure data belonging to the same server type into a data detection set.

[0015] Preferably, the step of extracting differential features from the data detection set to obtain at least one differential feature set includes the following steps:

[0016] Obtain the generation location of the historical failure data in the server, and summarize the historical failure data with the same generation location in the data detection set into a display feature data set, where the display feature data set corresponds to the generation location;

[0017] Use the maximum and minimum values of the display feature data set to form a display interval;

[0018] Use the display intervals of the remaining data detection sets other than the data detection set as non-target display intervals;

[0019] Merge the non-target display intervals corresponding to the same generation location into a display summary interval;

[0020] Obtain the intersection of the display interval and the display summary interval corresponding to the same generation location to obtain an unnecessary interval;

[0021] Delete the part occupied by the unnecessary interval corresponding to the same generation location in the display interval to obtain a differential interval;

[0022] Pair the differential interval with the generation location as a differential feature;

[0023] Summarize the differential features in the data detection set to obtain a differential feature set.

[0024] Preferably, calculating the correlation coefficient between the cause of the fault and the differential features in the differential feature set includes the following steps:

[0025] Obtain at least one triggering location of the cause of the fault, and obtain the triggering probability of the cause of the fault triggered by the triggering location. The triggering probability is the probability of the cause of the fault when the triggering location is abnormal;

[0026] Pair the triggering location with the generation location whose distance is less than the preset distance, and match the triggering probability corresponding to the triggering location to the generation location corresponding to the triggering location;

[0027] Uniformly take at least one recognition point in the differential interval. Under the condition that the data at the generation location of the server type corresponding to the differential interval is equal to the value at the recognition point, obtain the probability of the cause of the fault, which is used as the conditional probability;

[0028] Take the mean of at least one conditional probability in the differential interval to obtain the conditional average probability;

[0029] Use the correlation formula to calculate the correlation coefficient between the cause of the fault and the differential features in the differential feature set;

[0030] The correlation formula is as follows:

[0031] A = ab

[0032] Wherein, A is the correlation coefficient between the cause of the fault and the differential features in the differential feature set, a is the triggering probability corresponding to the generation location in the differential feature, and b is the conditional average probability of the differential interval in the differential feature.

[0033] Preferably, determining the associated feature set in the differential feature set based on the correlation coefficient includes the following steps:

[0034] Form an identification critical value for the correlation coefficient between the cause of the fault and the differential features;

[0035] When the correlation coefficient is greater than the identification critical value, the differential features corresponding to the correlation coefficient are incorporated into the associated feature set, and the initial state of the associated feature set is an empty set.

[0036] Preferably, forming an identification critical value for the correlation coefficient between the cause of the fault and the differential features includes the following steps:

[0037] Regard the differential features corresponding to the generation location with a triggering probability less than the preset value as non-essential features;

[0038] Take the maximum value of the non-essential features as the identification critical value.

[0039] Preferably, the recognition mechanism for the cause of failure of server types based on the associated feature set includes the following steps:

[0040] Use the first comprehensive failure formula to calculate the historical parameters of the cause of failure, and take the minimum value of the historical parameters as the failure critical value;

[0041] Based on historical data, obtain the detection method for the server type, and use the detection method for the server type to test the server consistent with the server type to obtain the actual test data generated at the location where the difference feature occurs;

[0042] Use the second comprehensive failure formula to calculate the actual parameters of the cause of failure;

[0043] When the actual parameter of the cause of failure is less than the failure critical value of the cause of failure, the cause of failure does not exist; otherwise, the cause of failure exists;

[0044] The first comprehensive failure formula is as follows:

[0045]

[0046] Among them, B is the historical parameter of the cause of failure, i is the subscript, n is the total number of difference features in the associated feature set, c i is the value of the left endpoint of the difference interval in the i-th difference feature in the associated feature set, and d i is the correlation coefficient between the i-th difference feature in the associated feature set and the cause of failure;

[0047] The second comprehensive failure formula is as follows:

[0048]

[0049] Among them, C is the actual parameter of the cause of failure, and e i is the actual test data generated at the location where the i-th difference feature in the associated feature set occurs.

[0050] Preferably, the integration of at least one detection technical solution to obtain an integrated detection package includes the following steps:

[0051] Obtain at least one historical detection data of the server consistent with the server type using the detection technical solution of the server type, and obtain the value range of the historical detection data collected at the generation location as the value interval;

[0052] For the generation mechanism of the matching coefficient of the detection technical solution, the generation mechanism of the matching coefficient is as follows: Use the detection technical solution to detect the server to obtain the detection data of at least one generation location. When the detection data of the generation location all belong to the value range corresponding to the generation location, the matching coefficient of the detection technical solution is 1; otherwise, the matching coefficient of the detection technical solution is 0.

[0053] Pair and summarize the detection technical solution with the matching coefficient to obtain an integrated detection package.

[0054] Preferably, the use of the integrated detection package to detect at least one server to be tested includes the following steps:

[0055] Use at least one detection technical solution in the integrated detection package to perform parallel detection on at least one server to be tested;

[0056] Pair the detection technical solution with the server to be tested, and satisfy that the matching coefficient generated by using the detection technical solution to detect the server to be tested is 1;

[0057] Use the identification mechanism of the fault cause in the detection technical solution corresponding to the server to be tested to determine the existing fault cause in the server to be tested.

[0058] Preferably, the use of the fault repair solution corresponding to the fault cause of the server to be tested for repair includes the following steps:

[0059] Based on historical data, summarize the fault repair data, construct a repair call model, and pair the fault cause of the server type with the fault repair solution in the repair call model;

[0060] Call the fault repair solution corresponding to the fault cause of the server type that is the same as the fault cause of the server to be tested for repair.

[0061] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0062] By forming a difference feature set, calculating the correlation coefficient between the fault cause and the difference features in the difference feature set, and forming an identification mechanism for the fault cause of the server type, different models of servers can be identified, and corresponding detection technical solutions can be matched for them, so as to achieve detection and fault identification. In the whole process, algorithms are used for identification without using manpower. At the same time, parallel detection of multiple servers can be carried out, which can effectively improve the detection efficiency and ensure the detection accuracy. Description of the Drawings

[0063] Figure 1 It is a schematic flowchart of the one-to-many server testing method based on detection technology integration of the present invention;

[0064] Figure 2 It is a schematic flow chart for extracting differential features from a data detection set in the present invention to obtain at least one differential feature set;

[0065] Figure 3 It is a schematic flow chart for calculating the correlation coefficient between the cause of a fault and the differential features in the differential feature set in the present invention;

[0066] Figure 4 It is a schematic flow chart for determining an associated feature set in the differential feature set based on the correlation coefficient in the present invention;

[0067] Figure 5 It is a schematic flow chart for forming an identification threshold value for the correlation coefficient between the cause of a fault and the differential features in the present invention;

[0068] Figure 6 It is a schematic flow chart for forming an identification mechanism for the cause of a fault of a server type based on the associated feature set in the present invention;

[0069] Figure 7 It is a schematic flow chart for integrating at least one detection technical solution in the present invention to obtain an integrated detection package;

[0070] Figure 8 It is a schematic flow chart for using the integrated detection package to detect at least one server to be tested in the present invention;

[0071] Figure 9 It is a schematic flow chart for performing maintenance using a fault repair plan corresponding to the cause of a fault of a server to be tested in the present invention. Detailed implementation manners

[0072] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations.

[0073] Referring to Figure 1 As shown, a one-to-many server testing method based on detection technology integration includes:

[0074] Obtain historical fault data of at least one server, and classify the historical fault data based on the server type to obtain at least one data detection set, where the data detection set corresponds to the server type;

[0075] Extract differential features from the data detection set to obtain at least one differential feature set, where the differential feature set corresponds to the data detection set and the differential feature set is composed of at least one differential feature;

[0076] Obtain at least one fault cause of the server, calculate the correlation coefficient between the fault cause and the differential features in the differential feature set, and based on the correlation coefficient, determine an associated feature set in the differential feature set. The associated feature set corresponds to the fault cause and is composed of differential features;

[0077] Based on the associated feature set, form an identification mechanism for the fault causes of server types, and summarize the identification mechanisms for at least one fault cause of server types to obtain a detection technical solution for server types;

[0078] Obtain the detection technical solutions corresponding to server types, integrate at least one detection technical solution to obtain an integrated detection package, and use the integrated detection package to detect at least one server to be tested to obtain the fault causes of the servers to be tested;

[0079] Perform repairs using the fault repair plan corresponding to the fault cause of the server to be tested.

[0080] When performing server detection, since there are different server models, the detection and identification schemes adopted by different models of servers are different, and the faults of servers usually do not exist singly. Therefore, it is difficult to separate the detection data for each fault one by one. Therefore, it is easy to make abnormal judgments based on the detection data, but it is more difficult to determine the fault causes. Moreover, since the faults are unknown, it will further increase the difficulty of accurate judgment. In this solution, corresponding algorithms are set for the various detection difficulties mentioned above, and targeted detection and identification are carried out.

[0081] Based on server types, classifying historical fault data to obtain at least one data detection set includes the following steps:

[0082] Summarize the historical fault data belonging to the same server type into a data detection set.

[0083] Refer to Figure 2 As shown, extracting differential features from the data detection set to obtain at least one differential feature set includes the following steps:

[0084] Obtain the generation location of the historical fault data in the server, and summarize the historical fault data with the same generation location in the data detection set into a display feature data set. The display feature data set corresponds to the generation location;

[0085] Use the maximum and minimum values of the display feature data set to form a display interval;

[0086] Take the display interval of the remaining data detection sets outside the data detection set as the non-target display interval;

[0087] Merge non-target display intervals corresponding to the same generation location into a display summary interval;

[0088] Obtain the intersection of the display interval and the display summary interval corresponding to the same generation location to obtain an unnecessary interval;

[0089] Delete the part occupied by the unnecessary interval corresponding to the same generation location in the display interval to obtain a difference interval;

[0090] Pair the difference interval with the generation location as a difference feature;

[0091] Summarize the difference features in the data detection set to obtain a difference feature set.

[0092] Although the types of servers are different, the basic structures of the servers are the same. Therefore, data can be classified according to location, and historical fault data located at approximate locations can be regarded as data of the same indicator. However, due to different server models, the detection data at the generation location of different models of servers is different, and the judgment for different models of servers also needs to be distinguished. Therefore, a difference feature set is formed, and the difference feature sets of each type of server are different. Thus, different judgment criteria can be formed for each type of server.

[0093] Refer to Figure 3 As shown, calculating the correlation coefficient between the cause of the fault and the difference features in the difference feature set includes the following steps:

[0094] Obtain at least one triggering location of the cause of the fault, and obtain the triggering probability of the triggering location triggering the cause of the fault. The triggering probability is the probability of the cause of the fault being triggered when the triggering location has an abnormality;

[0095] Pair the triggering location with the generation location whose distance is less than a preset distance, and match the triggering probability corresponding to the triggering location to the generation location corresponding to the triggering location;

[0096] Uniformly take at least one identification point in the difference interval. Under the condition that the data at the generation location of the server type corresponding to the difference interval is equal to the value at the identification point, obtain the probability of triggering the cause of the fault as the conditional probability;

[0097] Take the mean of at least one conditional probability of the difference interval to obtain the conditional average probability;

[0098] Use the correlation formula to calculate the correlation coefficient between the cause of the fault and the difference features in the difference feature set;

[0099] The correlation formula is as follows:

[0100] A = ab

[0101] Among them, A is the correlation coefficient between the cause of the fault and the differential feature in the set of differential features, a is the triggering probability corresponding to the generation position in the differential feature, and b is the conditional average probability of the differential interval in the differential feature.

[0102] There are two determinants of the correlation coefficient, namely the cause of the fault and the differential feature. When the two are determined, the correlation coefficient is determined. Since the detection data of the server is the superposition effect of several actual causes of the fault, therefore, when performing identification, it is necessary to decompose the detection data of the server to determine the comprehensive result of the data regarding a single cause of the fault in the detection data, and then make a judgment based on the comprehensive result. Therefore, it is necessary to obtain the correlation coefficient between the cause of the fault and the differential feature, which can be used as a weight to extract and synthesize the data related to a single cause of the fault in the detection data of the server.

[0103] Refer to Figure 4 As shown, based on the correlation coefficient, determining the associated feature set in the set of differential features includes the following steps:

[0104] Form an identification threshold for the correlation coefficient between the cause of the fault and the differential feature;

[0105] When the correlation coefficient is greater than the identification threshold, the differential feature corresponding to the correlation coefficient is incorporated into the associated feature set, and the initial state of the associated feature set is an empty set.

[0106] Since not all of the differential features in the set of differential features are features with a key role, in order to control the workload of subsequent detection, therefore, only select the differential features in the set of differential features that have a greater association with the cause of the fault to form an associated feature set. Thus, the detection range can be reduced, and the detection efficiency can be improved.

[0107] Refer to Figure 5 As shown, forming an identification threshold for the correlation coefficient between the cause of the fault and the differential feature includes the following steps:

[0108] Regard the differential feature corresponding to the generation position where the triggering probability is less than the preset value as a non-essential feature;

[0109] Take the maximum value of the non-essential features as the identification threshold.

[0110] The role of the identification threshold is to screen the correlation coefficient. Therefore, it regards the differential feature corresponding to the generation position where the triggering probability is less than the preset value as a non-essential feature. The possibility of the differential feature at these positions causing the cause of the fault is small. Therefore, it can be regarded as an irrelevant factor and can be ignored when making a selection.

[0111] Refer to Figure 6 As shown, forming an identification mechanism for the cause of the fault of the server type based on the associated feature set includes the following steps:

[0112] Using the first fault synthesis formula, calculate the historical parameters of the cause of the fault, and take the minimum value of the historical parameters as the fault critical value;

[0113] Based on historical data, obtain the detection method for the type of server. Use the detection method for the type of server to test the servers consistent with the type of server, and obtain the actual test data generated at the location where the difference feature occurs;

[0114] Using the second fault synthesis formula, calculate the actual parameters of the cause of the fault;

[0115] When the actual parameter of the cause of the fault is less than the fault critical value of the cause of the fault, then the cause of the fault does not exist; otherwise, the cause of the fault exists;

[0116] The first fault synthesis formula is as follows:

[0117]

[0118] Where B is the historical parameter of the cause of the fault, i is the subscript, n is the total number of difference features in the associated feature set, c i is the value of the left end point of the difference interval in the i-th difference feature in the associated feature set, and d i is the correlation coefficient between the i-th difference feature in the associated feature set and the cause of the fault;

[0119] The second fault synthesis formula is as follows:

[0120]

[0121] Where C is the actual parameter of the cause of the fault, and e i is the actual test data generated at the location where the i-th difference feature in the associated feature set occurs.

[0122] Although the types of servers are different, their basic structures are similar. Therefore, the causes of their faults are the same, but the criteria for identifying the causes of faults are different. When conducting identification, since it is unknown whether each cause of the fault exists, it is necessary to judge each cause of the fault and determine its existence or not. Therefore, first use the first fault synthesis formula to separate the part associated with the cause of the fault in the historical data and use the correlation coefficient for weighted synthesis to form the fault critical value. At the same time, also use the second fault synthesis formula to separate the part associated with the cause of the fault in the actual data and use the correlation coefficient for weighted synthesis to obtain the actual parameter of the cause of the fault. Furthermore, by comparing the two, it can be determined whether the cause of the fault exists.

[0123] Refer to Figure 7As shown, the steps for integrating at least one detection technical solution to obtain an integrated detection package are as follows:

[0124] Obtain at least one historical detection data of a server consistent with the server type using the detection technical solution for the server type, and obtain the value range of the historical detection data collected at the generation location as the value interval;

[0125] Generate a generation mechanism for the matching coefficient of the detection technical solution. The generation mechanism of the matching coefficient is as follows: Use the detection technical solution to detect the server to obtain the detection data of at least one generation location. When the detection data of the generation location all belong to the value interval corresponding to the generation location, the matching coefficient of the detection technical solution is 1; otherwise, the matching coefficient of the detection technical solution is 0;

[0126] Pair and summarize the detection technical solution with the matching coefficient to obtain an integrated detection package.

[0127] When detecting a to-be-detected server, since the model of the to-be-detected server is unknown, it is necessary to determine the detection technical solution corresponding to its model. The data detected using the detection technical solution corresponding to the server will surely form a value interval at each generation location. When using a detection technical solution that does not correspond to the server type for detection, the data generated at the generation location has a high probability of being outside the value interval. Since there are multiple generation locations, based on the basic knowledge of probability, when the detection data of the generation location all belong to the value interval corresponding to the generation location, the matching coefficient of the detection technical solution is 1. The possibility of a matching error occurring with such a setting is very small. For example, when using a detection technical solution that does not correspond to the server type for detection, assume that the probability of the data generated at the generation location being outside the value interval is 0.6, and assume that the number of generation locations is 10. In this case, the probability that the matching coefficient of the detection technical solution that does not correspond to the server type is set to 1 is the probability that all the data generated at the generation locations are within the value interval, that is, the tenth power of 0.4, which is almost 0. Therefore, it is reasonable to set the matching coefficient in this way.

[0128] Refer to Figure 8 As shown, using the integrated detection package to detect at least one to-be-detected server includes the following steps:

[0129] Use at least one detection technical solution in the integrated detection package to perform parallel detection on at least one to-be-detected server;

[0130] Pair the detection technical solution with the to-be-detected server so that the matching coefficient generated by using the detection technical solution to detect the to-be-detected server is 1;

[0131] Use the fault cause identification mechanism in the detection technical solution corresponding to the server under test to determine the existing fault causes in the server under test.

[0132] Refer to Figure 9 As shown, the repair using the fault repair solution corresponding to the fault cause of the server under test includes the following steps:

[0133] Based on historical data, summarize the data of fault repair, construct a repair call model, and pair the fault causes of server types with fault repair solutions in the repair call model;

[0134] Call the fault repair solution corresponding to the fault cause of the server type that is the same as the fault cause of the server under test for repair.

[0135] Furthermore, this solution also proposes a storage medium on which a computer-readable program is stored. When the computer-readable program is called, it executes the above-mentioned one-to-many server testing method based on detection technology integration.

[0136] It can be understood that the storage medium can be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid-state drive (SSD).

[0137] In summary, the advantages of the present invention are as follows: By forming a set of differential features, calculating the correlation coefficient between the fault cause and the differential features in the set of differential features, and forming an identification mechanism for the fault causes of server types, different models of servers can be identified, and corresponding detection technical solutions can be matched for them, thereby realizing detection and fault identification. In the whole process, algorithms are used for identification without using human labor. At the same time, parallel detection of multiple servers can be carried out, which can effectively improve the detection efficiency and ensure the detection accuracy.

[0138] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A one-to-many server testing method based on the integration of detection technologies, characterized in that, Including: Obtain the historical failure data of at least one server, and classify the historical failure data based on the server type to obtain at least one data detection set, where the data detection set corresponds to the server type; Extract the differential features from the data detection set to obtain at least one differential feature set, where the differential feature set corresponds to the data detection set, and the differential feature set is composed of at least one differential feature; Obtain at least one failure cause of the server, calculate the correlation coefficient between the failure cause and the differential features in the differential feature set, and determine the associated feature set in the differential feature set based on the correlation coefficient. The associated feature set corresponds to the failure cause, and the associated feature set is composed of differential features; Form an identification mechanism for the failure cause of the server type based on the associated feature set, and summarize the identification mechanisms for at least one failure cause of the server type to obtain a detection technical solution for the server type; Obtain the detection technical solution corresponding to the server type, integrate at least one detection technical solution to obtain an integrated detection package, and use the integrated detection package to detect at least one server to be tested to obtain the failure cause of the server to be tested; Perform maintenance using the failure repair plan corresponding to the failure cause of the server to be tested.

2. The one-to-many server testing method based on the integration of detection technologies according to claim 1, wherein The step of classifying the historical failure data based on the server type to obtain at least one data detection set includes the following steps: Summarize the historical failure data belonging to the same server type into a data detection set.

3. The one-to-many server testing method based on detection technology integration according to claim 2, characterized in that The step of extracting the differential features from the data detection set to obtain at least one differential feature set includes the following steps: Obtain the generation location of the historical failure data in the server, and summarize the historical failure data with the same generation location in the data detection set into a display feature data set, where the display feature data set corresponds to the generation location; Use the maximum and minimum values of the display feature data set to form a display interval; Take the display intervals of the remaining data detection sets other than the data detection set as non-target display intervals; Merge the non-target display intervals corresponding to the same generation location into a display summary interval; Obtain the intersection of the display interval and the display summary interval corresponding to the same generation location to obtain an unnecessary interval; Delete the part occupied by the unnecessary interval corresponding to the same generation location in the display interval to obtain a differential interval; Pair the differential interval with the generation location as a differential feature; Summarize the differential features in the data detection set to obtain a differential feature set.

4. The one-to-many server testing method based on detection technology integration according to claim 3, characterized in that, The step of calculating the correlation coefficient between the failure cause and the differential features in the differential feature set includes the following steps: Obtain at least one triggering location of the failure cause, and obtain the triggering probability of the triggering location triggering the failure cause. The triggering probability is the probability of the failure cause being triggered when the triggering location has an abnormality; Pair the triggering location with the generation location whose distance is less than the preset distance, and match the triggering probability corresponding to the triggering location to the generation location corresponding to the triggering location; Uniformly take at least one identification point in the differential interval, and obtain the probability of triggering the failure cause under the condition that the data at the generation location of the server type corresponding to the differential interval is equal to the value at the identification point as the conditional probability; Take the mean of at least one conditional probability of the difference interval to obtain the conditional average probability; Use the correlation formula to calculate the correlation coefficient between the cause of the fault and the difference features in the set of difference features; The correlation formula is as follows: A = ab Where A is the correlation coefficient between the cause of the fault and the difference features in the set of difference features, a is the trigger probability corresponding to the generation position in the difference features, and b is the conditional average probability of the difference interval in the difference features.

5. A one-to-many server testing method based on the integration of detection technologies according to claim 4, characterized in that The step of determining the associated feature set in the set of difference features based on the correlation coefficient includes the following steps: Form an identification threshold for the correlation coefficient between the cause of the fault and the difference features; When the correlation coefficient is greater than the identification threshold, the difference feature corresponding to the correlation coefficient is incorporated into the associated feature set, and the initial state of the associated feature set is an empty set.

6. The one-to-many server testing method based on detection technology integration according to claim 5, characterized in that The step of forming an identification threshold for the correlation coefficient between the cause of the fault and the difference features includes the following steps: Take the difference features corresponding to the generation positions with trigger probabilities less than the preset value as non-essential features; Take the maximum value of the non-essential features as the identification threshold.

7. A one-to-many server testing method based on the integration of detection technologies according to claim 6, characterized in that, The step of forming an identification mechanism for the cause of the fault of the server type based on the associated feature set includes the following steps: Use the first fault synthesis formula to calculate the historical parameters of the cause of the fault, and take the minimum value of the historical parameters as the fault threshold; Based on historical data, obtain the detection method of the server type, and use the detection method of the server type to test the server consistent with the server type to obtain the actual test data generated at the generation position in the difference features; Use the second fault synthesis formula to calculate the actual parameters of the cause of the fault; When the actual parameter of the cause of the fault is less than the fault threshold of the cause of the fault, the cause of the fault does not exist, otherwise, the cause of the fault exists; The first fault synthesis formula is as follows: Among them, B is the historical parameter of the cause of the fault, i is the subscript, n is the total number of differential features in the associated feature set, c i is the value of the left endpoint of the differential interval in the i-th differential feature in the associated feature set, d i is the correlation coefficient between the i-th differential feature in the associated feature set and the cause of the fault; The second fault synthesis formula is as follows: Among them, C is the actual parameter of the fault cause, and e i is the actual test data generated at the generation position of the i-th differential feature in the associated feature set.

8. A one-to-many server testing method based on the integration of detection technologies according to claim 7, characterized in that, The step of integrating at least one detection technical solution to obtain an integrated detection package includes the following steps: Obtain at least one historical detection data of the server consistent with the server type using the detection technical solution of the server type, and obtain the value range of the historical detection data collected at the generation position as the value interval; Generate a generation mechanism for the matching coefficient of the detection technical solution. The generation mechanism of the matching coefficient is as follows: Use the detection technical solution to detect the server to obtain the detection data of at least one generation position. When the detection data of the generation position all belong to the value interval corresponding to the generation position, the matching coefficient of the detection technical solution is 1, otherwise, the matching coefficient of the detection technical solution is 0; Pair and summarize the detection technical solution and the matching coefficient to obtain an integrated detection package.

9. A one-to-many server testing method based on the integration of detection technologies according to claim 8, characterized in that, The step of using the integrated detection package to detect at least one server to be tested includes the following steps: Use at least one detection technical solution in the integrated detection package to perform parallel detection on at least one server to be tested; Pair the detection technical solution with the server to be tested, and satisfy that the matching coefficient generated by using the detection technical solution to detect the server to be tested is 1; Use the identification mechanism of the cause of the fault in the detection technical solution corresponding to the server to be tested to determine the cause of the fault existing in the server to be tested.

10. The one-to-many server testing method based on detection technology integration according to claim 9, characterized in that Performing maintenance using the fault repair plan corresponding to the fault cause of the server to be tested includes the following steps: Based on historical data, summarize the data for fault repair, construct a maintenance call model, and pair the fault causes of server types with fault repair plans in the maintenance call model; Call the fault repair plan corresponding to the fault cause of the server type that is the same as the fault cause of the server to be tested for maintenance.

Images