Computer cluster maintenance task management method and system

Abstract

The invention relates to the technical field of maintenance task management, and particularly discloses a computer cluster maintenance task management method and system. The method comprises the following steps of S1, obtaining the number of computing nodes in a cluster, S2, collecting data of all the computing nodes in the cluster, S3, analyzing cluster operation state data, S4, judging and managing cluster operation health indexes, S5, analyzing cluster task execution data, and S6, analyzing cluster task execution data. S7, a user interface module is established, and the execution efficiency index of the cluster task is judged and managed; according to the method, efficient management and optimization of the computer cluster are achieved by calculating node division and collecting data, analyzing the cluster operation health index and the cluster task execution efficiency index and establishing the optimization management model, the performance and reliability of the cluster are improved through the improvements, the management efficiency and the user experience are remarkably improved, and the user experience is improved. The method has important practical application value.

Description

Technical Field

[0001] The present invention relates to the technical field of maintenance task management, and in particular to a computer cluster maintenance task management method and system. Background Art

[0002] With the rapid development of information technology, computer clusters have been widely used in various large-scale data processing and high-performance computing scenarios. A computer cluster is a system consisting of multiple independent computers (nodes) connected together through a high-speed network to form a unified computing resource pool to jointly complete computing tasks or provide services. Each node usually has its own processor, memory, and storage device, and exchanges data and coordinates work through network communication protocols. In order to ensure the efficient operation and high reliability of the cluster, regular maintenance task management is very necessary.

[0003] In traditional computer cluster maintenance task management, administrators usually rely on a series of manual operations to ensure the stability and performance of the cluster. These operations include manually assigning maintenance tasks, regularly checking the health of the cluster, performing software updates and hardware checks on each node, analyzing system logs in detail to identify potential problems, and responding when failures occur. These methods often rely on the personal experience and intuition of administrators, who perform various tasks according to preset maintenance plans to ensure that the cluster can continue to operate reliably.

[0004] However, traditional computer cluster maintenance task management methods have some disadvantages. First, due to the high reliance on manual operations, efficiency is often low and prone to errors. In large clusters, manual maintenance tasks are not only time-consuming but also prone to missing key details, making it difficult to ensure maintenance quality. Secondly, traditional management methods cannot analyze the operating status of computer clusters and the performance of cluster task execution in real time, nor can they timely warn and optimize management of abnormal data, reducing management efficiency. Finally, high maintenance costs are also a major disadvantage of traditional methods because they require a lot of manpower and time investment, and these resources could have been used for more value-added activities. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a computer cluster maintenance task management method and system to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a computer cluster maintenance task management method, comprising the following steps: step S1: obtaining the number of computing nodes inside the cluster, step S2: collecting data of each computing node in the cluster, step S3: analyzing the cluster operation status data, step S4: judging and managing the cluster operation health indicators, step S5: analyzing the cluster task execution data, step S6: judging and managing the cluster task execution performance indicators and step S7: establishing a user interface module.

[0007] Step S1: Obtain the number of computing nodes in the cluster: determine the computer cluster as the target area, obtain the number of computing nodes in the cluster, and mark each computing node as 1, 2, ..., i, ..., n in sequence, where i is the number of each computing node and n represents the total number of computing nodes in the cluster;

[0008] Step S2: Collecting data from each computing node in the cluster: monitoring and collecting cluster task comprehensive data of each computing node in the cluster, wherein the cluster task comprehensive data includes cluster operation status data and cluster task execution data;

[0009] Step S3: Analyze the cluster operation status data: Analyze the cluster operation status data collected in step S2, and calculate the cluster operation health index of each computing node through the operation status mathematical model;

[0010] Step S4: judging and managing the cluster operation health index: judging the cluster operation health index, optimizing and managing the cluster operation health index through the cluster operation optimization model, and obtaining the optimized cluster operation health index;

[0011] Step S5: Analyze the cluster task execution data: Analyze the cluster task execution data collected in step S2, and calculate the cluster task execution efficiency index of each computing node through the task execution mathematical model;

[0012] Step S6: judging and managing the cluster task execution efficiency index: judging the cluster task execution efficiency index, optimizing and managing the cluster task execution efficiency index through the cluster task scheduling optimization model, and obtaining the optimized cluster task execution efficiency index;

[0013] Step S7: Establishing a user interface module: Displaying the computer cluster operation status data and the data after the cluster maintenance task optimization management through the user interface module for operation and maintenance personnel to analyze and make decisions.

[0014] Preferably, the cluster operation status data includes CPU resource utilization, memory resource utilization, disk resource utilization, network bandwidth resource utilization, the number of tasks of computing nodes, the normal operation time of the computer cluster, the total operation time of the computer cluster and the total number of cluster failures; the cluster task execution data includes the actual execution time of the cluster task, the number of successfully completed cluster tasks, the number of cluster task timeouts, the total number of cluster tasks and the cluster task response time.

[0015] Preferably, the steps for calculating the cluster operation health index of each computing node are specifically as follows:

[0016] Step S31: Calculate the cluster resource utilization RU of each node i , the calculation model is as follows:

[0017] Among them, CPU i It is expressed as the CPU resource utilization of the i-th computing node, Memu i It is expressed as the memory resource utilization of the i-th computing node, Disu i Expressed as the disk resource utilization of the i-th computing node, Bwu i It is expressed as the network bandwidth resource utilization rate of the i-th computing node, where i is the number of each computing node;

[0018] Step S32: Calculate the cluster node load balancing degree NLB of each node i , the calculation model is as follows:

[0019] Among them, TN i It is represented by the number of tasks of the i-th computing node, n is represented by the total number of computing nodes in the cluster, TN max It is expressed as the maximum number of tasks, that is, the number of tasks of the computing node with the largest number of tasks among all computing nodes, TN min It is expressed as the minimum number of tasks, that is, the number of tasks of the computing node with the least number of tasks among all computing nodes;

[0020] Step S33: Calculate the cluster operation health index OHI of each node i , the calculation model is as follows:

[0021] Among them, OHI i Represented as the cluster operation health indicator of the i-th computing node, RU i It is expressed as the cluster resource utilization of the i-th computing node, NLB i It is represented as the cluster node load balancing degree of the i-th computing node, Ut i It is represented as the normal operation time of the computer cluster of the i-th computing node, Trti Ff is the total running time of the computer cluster of the i-th computing node. i It is represented as the total number of cluster failures of the i-th computing node, n is represented as the total number of computing nodes in the cluster, and e is represented as a natural constant.

[0022] Preferably, the specific content of judging the cluster operation health index in step S4 is:

[0023] The cluster operation health index of each computing node is extracted and compared with the preset health threshold. If the cluster operation health index of the ith computing node is greater than or equal to the preset health threshold, the cluster operation health index of the ith computing node is judged to be normal. If the cluster operation health index of the ith computing node is less than the preset health threshold, the cluster operation health index of the ith computing node is judged to be abnormal. An early warning signal is issued for the abnormal cluster operation result and the abnormal node number is sent to the corresponding operation and maintenance management personnel for review. The cluster operation abnormal data is optimized and managed through the cluster operation optimization model.

[0024] Preferably, the cluster operation optimization model is as follows:

[0025] maxY=W1×RU i +W2×NLB i +W3×(1-Fr i ), where maxY represents

[0026] The output value of the cluster running optimization model, RU i It is expressed as the cluster resource utilization of the i-th computing node, NLB i It is represented as the cluster node load balancing degree of the i-th computing node, Fr i It is represented as the cluster operation failure rate of the i-th computing node, W1, W2, and W3 are respectively represented as the weight factors of cluster resource utilization, cluster node load balance, and cluster operation failure rate, and W1+W2+W3=1.

[0027] Preferably, the calculation steps of the cluster task execution efficiency index of each computing node are specifically as follows:

[0028] Step S51: Calculate the cluster task execution deviation TED of each node i , the calculation model is as follows:

[0029] Among them, Aet i It is represented by the actual execution time of the cluster task of the i-th computing node, Ed 预 represents the ideal execution time preset for the cluster task of the i-th computing node, Ctn iIt is represented as the number of cluster tasks successfully completed by the i-th computing node;

[0030] Step S52: Calculate the cluster task timeout rate TTR of each node i , the calculation model is as follows:

[0031] Among them, Ton i Represented as the number of cluster task timeouts of the i-th computing node, ttn i It is represented as the total number of cluster tasks of the i-th computing node;

[0032] Step S53: Calculate the cluster task execution efficiency index EEI of each computing node i , the calculation model is as follows:

[0033] Among them, EEI i It is represented as the cluster task execution efficiency index of the i-th computing node, Ctn i It is expressed as the number of cluster tasks successfully completed by the i-th computing node, rt i Represented as the cluster task response time of the i-th computing node, ttn i It is represented as the total number of cluster tasks of the i-th computing node, TED i It is represented as the cluster task execution deviation of the i-th computing node, TED 预 It is represented by the preset maximum cluster task execution deviation, TTR i It is represented as the cluster task timeout rate of the i-th computing node, and e is represented as a natural constant.

[0034] Preferably, the specific content of judging the cluster task execution efficiency index in step S6 is:

[0035] The cluster task execution performance index of each computing node is extracted and compared with the preset performance threshold. If the cluster task execution performance index of the ith computing node is greater than or equal to the preset performance threshold, the cluster task execution performance index of the ith computing node is judged to be normal. If the cluster task execution performance index of the ith computing node is less than the preset performance threshold, the cluster task execution performance index of the ith computing node is judged to be abnormal. An early warning signal is issued for the abnormal task performance result and the abnormal node number is sent to the corresponding operation and maintenance management personnel for review. The task performance abnormality data is optimized and managed through the cluster task scheduling optimization model.

[0036] Preferably, the cluster task scheduling optimization model is as follows:

[0037] maxZ=K1×(1-TED i )-K2×TTR i +K3×Sri , where maxZ represents

[0038] Output value of the optimization model for cluster task scheduling, TED i It is represented as the cluster task execution deviation of the i-th computing node, TTR i It is represented as the cluster task timeout rate of the i-th computing node, Sr i It is represented as the success rate of cluster task completion of the i-th computing node, K1, K2, and K3 are weight factors of cluster task execution deviation, cluster task timeout rate, and cluster task completion success rate, respectively, and K1+K2+K3=1.

[0039] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a computer cluster maintenance task management system, which implements the above-mentioned computer cluster maintenance task management method, including: a cluster internal node number acquisition module, a cluster node data acquisition module, a cluster operation status analysis module, a cluster operation health index judgment management module, a cluster task execution analysis module, a cluster task execution efficiency index judgment management module and a user interface establishment module.

[0040] The cluster internal node number acquisition module determines the computer cluster as the target area, acquires the number of computing nodes in the cluster, and marks each computing node as 1, 2, ..., i, ..., n in sequence, where i is the number of each computing node and n represents the total number of computing nodes in the cluster;

[0041] The cluster node data collection module is used to monitor and collect the cluster task comprehensive data of each computing node in the cluster, and transmit it to the cluster operation status analysis module and the cluster task execution analysis module, wherein the cluster task comprehensive data includes the cluster operation status data and the cluster task execution data;

[0042] The cluster operation status analysis module is used to analyze the cluster operation status data collected in the cluster node data collection module, calculate the cluster operation health index of each computing node through the operation status mathematical model, and transmit it to the cluster operation health index judgment management module;

[0043] The cluster operation health index judgment management module is used to judge the cluster operation health index, optimize and manage the cluster operation health index through the cluster operation optimization model, and obtain the optimized cluster operation health index;

[0044] The cluster task execution analysis module is used to analyze the cluster task execution data collected in the cluster node data collection module, calculate the cluster task execution efficiency index of each computing node through the task execution mathematical model, and transmit it to the cluster task execution efficiency index judgment management module;

[0045] The cluster task execution efficiency index judgment management module judges the cluster task execution efficiency index, optimizes and manages the cluster task execution efficiency index through the cluster task scheduling optimization model, and obtains the optimized cluster task execution efficiency index;

[0046] The user interface module is established to display the computer cluster operation status data and the data after the cluster maintenance task optimization management through the user interface module for operation and maintenance personnel to analyze and make decisions.

[0047] Technical effects and advantages of the present invention:

[0048] 1. The present invention collects computer cluster node data by computing node division, and can obtain the operating status and performance indicators of each node in real time. This fine-grained data collection method makes the data more accurate and comprehensive, providing a solid foundation for subsequent analysis and optimization; by analyzing the collected data, the cluster operation health indicators and cluster task execution efficiency indicators of each computing node are calculated, which can fully reflect the cluster's operating status and task execution efficiency, helping administrators to discover and solve problems in a timely manner;

[0049] 2. The present invention automatically detects and identifies abnormal data by establishing an optimized management model. This automated abnormality detection mechanism can quickly discover problems such as node failures and resource bottlenecks, reduce delays and errors caused by human intervention, and reduce management costs;

[0050] 3. The present invention collects computer cluster node data by computing node division, analyzes and obtains cluster operation health indicators and cluster task execution efficiency indicators of each computing node, and establishes an optimization management model through judgment and analysis of the two indicators to optimize the management of abnormal data, thereby improving the management efficiency of the computer cluster. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0052] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0053] Figure 2 It is a schematic diagram of the overall structure of the system of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0055] Example 1

[0056] See also Figure 1 As shown, the present invention provides a computer cluster maintenance task management method, comprising the following steps: step S1: obtaining the number of computing nodes inside the cluster, step S2: collecting data of each computing node in the cluster, step S3: analyzing the cluster operation status data, step S4: judging and managing the cluster operation health indicators, step S5: analyzing the cluster task execution data, step S6: judging and managing the cluster task execution efficiency indicators, and step S7: establishing a user interface module.

[0057] Step S1: Obtain the number of computing nodes in the cluster: determine the computer cluster as the target area, obtain the number of computing nodes in the cluster, and mark each computing node as 1, 2, ..., i, ..., n in sequence, where i is the number of each computing node and n represents the total number of computing nodes in the cluster;

[0058] Step S2: Collecting data from each computing node in the cluster: monitoring and collecting cluster task comprehensive data of each computing node in the cluster, wherein the cluster task comprehensive data includes cluster operation status data and cluster task execution data;

[0059] In this embodiment, it should be specifically explained that the cluster operation status data includes CPU resource utilization, memory resource utilization, disk resource utilization, network bandwidth resource utilization, the number of tasks of computing nodes, the normal operation time of the computer cluster, the total operation time of the computer cluster and the total number of cluster failures; the cluster task execution data includes the actual execution time of the cluster task, the number of successfully completed cluster tasks, the number of cluster task timeouts, the total number of cluster tasks and the cluster task response time.

[0060] Step S3: Analyze the cluster operation status data: Analyze the cluster operation status data collected in step S2, and calculate the cluster operation health index of each computing node through the operation status mathematical model;

[0061] In this embodiment, it should be specifically explained that the steps for calculating the cluster operation health index of each computing node are as follows:

[0062] Step S31: Calculate the cluster resource utilization RU of each nodei , the calculation model is as follows:

[0063] Among them, CPU i It is expressed as the CPU resource utilization of the i-th computing node, Memu i It is expressed as the memory resource utilization of the i-th computing node, Disu i Expressed as the disk resource utilization of the i-th computing node, Bwu i It is expressed as the network bandwidth resource utilization rate of the i-th computing node, where i is the number of each computing node;

[0064] Step S32: Calculate the cluster node load balancing degree NLB of each node i , the calculation model is as follows:

[0065] Among them, TN i It is represented by the number of tasks of the i-th computing node, n is represented by the total number of computing nodes in the cluster, TN max It is expressed as the maximum number of tasks, that is, the number of tasks of the computing node with the largest number of tasks among all computing nodes, TN min It is expressed as the minimum number of tasks, that is, the number of tasks of the computing node with the least number of tasks among all computing nodes;

[0066] In this embodiment, it should be specifically noted that by calculating the cluster node load balance of each computing node, operation and maintenance personnel can evaluate the fairness and efficiency of task scheduling, optimize task allocation strategies, avoid resource bottlenecks, and improve the overall performance and stability of the cluster; the cluster node load balance value ranges from 0 to 1, and the closer the cluster node load balance value is to 1, the more evenly the tasks are distributed among the nodes in the cluster.

[0067] Step S33: Calculate the cluster operation health index OHI of each node i , the calculation model is as follows:

[0068] Among them, OHI i Represented as the cluster operation health indicator of the i-th computing node, RU i It is expressed as the cluster resource utilization of the i-th computing node, NLB i It is represented as the cluster node load balancing degree of the i-th computing node, Ut i It is represented as the normal operation time of the computer cluster of the i-th computing node, Trt i Ff is the total running time of the computer cluster of the i-th computing node. i It is represented as the total number of cluster failures of the i-th computing node, n is represented as the total number of computing nodes in the cluster, and e is represented as a natural constant.

[0069] Step S4: judging and managing the cluster operation health index: judging the cluster operation health index, optimizing and managing the cluster operation health index through the cluster operation optimization model, and obtaining the optimized cluster operation health index;

[0070] In this embodiment, it should be specifically explained that the specific content of judging the cluster operation health index in step S4 is:

[0071] The cluster operation health index of each computing node is extracted and compared with the preset health threshold. If the cluster operation health index of the ith computing node is greater than or equal to the preset health threshold, the cluster operation health index of the ith computing node is judged to be normal. If the cluster operation health index of the ith computing node is less than the preset health threshold, the cluster operation health index of the ith computing node is judged to be abnormal. An early warning signal is issued for the abnormal cluster operation result and the abnormal node number is sent to the corresponding operation and maintenance management personnel for review. The cluster operation abnormal data is optimized and managed through the cluster operation optimization model.

[0072] In this embodiment, it should be specifically explained that the cluster operation optimization model is as follows:

[0073] maxY=W1×RU i +W2×NLB i +W3×(1-Fr i ), where maxY represents

[0074] The output value of the cluster running optimization model, RU i It is expressed as the cluster resource utilization of the i-th computing node, NLB i It is represented as the cluster node load balancing degree of the i-th computing node, Fr i It is represented as the cluster operation failure rate of the i-th computing node, W1, W2, and W3 are respectively represented as the weight factors of cluster resource utilization, cluster node load balance, and cluster operation failure rate, and W1+W2+W3=1;

[0075] In this embodiment, it should be specifically explained that the specific optimization content in step S4 is: improving cluster resource utilization and cluster node load balancing, reducing cluster operation failure rate, and optimizing cluster operation; importing the cluster operation optimization model output value into the cluster operation health index in step S3, and obtaining the optimized cluster operation health index = maxY×OHI i .

[0076] Step S5: Analyze the cluster task execution data: Analyze the cluster task execution data collected in step S2, and calculate the cluster task execution efficiency index of each computing node through the task execution mathematical model;

[0077] In this embodiment, it should be specifically explained that the calculation steps of the cluster task execution efficiency index of each computing node are as follows:

[0078] Step S51: Calculate the cluster task execution deviation TED of each node i , the calculation model is as follows:

[0079] Among them, Aet i It is represented by the actual execution time of the cluster task of the i-th computing node, Ed 预 represents the ideal execution time preset for the cluster task of the i-th computing node, Ctn i It is represented as the number of cluster tasks successfully completed by the i-th computing node;

[0080] Step S52: Calculate the cluster task timeout rate TTR of each node i , the calculation model is as follows:

[0081] Among them, Ton i Represented as the number of cluster task timeouts of the i-th computing node, ttn i It is represented as the total number of cluster tasks of the i-th computing node;

[0082] Step S53: Calculate the cluster task execution efficiency index EEI of each computing node i , the calculation model is as follows:

[0083] Among them, EEI i It is represented as the cluster task execution efficiency index of the i-th computing node, Ctn i It is expressed as the number of cluster tasks successfully completed by the i-th computing node, rt i Represented as the cluster task response time of the i-th computing node, ttn i It is represented as the total number of cluster tasks of the i-th computing node, TED i It is represented as the cluster task execution deviation of the i-th computing node, TED 预 It is represented by the preset maximum cluster task execution deviation, TTR i It is represented as the cluster task timeout rate of the i-th computing node, and e is represented as a natural constant.

[0084] Step S6: judging and managing the cluster task execution efficiency index: judging the cluster task execution efficiency index, optimizing and managing the cluster task execution efficiency index through the cluster task scheduling optimization model, and obtaining the optimized cluster task execution efficiency index;

[0085] In this embodiment, it should be specifically explained that the specific content of judging the cluster task execution efficiency index in step S6 is:

[0086] The cluster task execution performance index of each computing node is extracted and compared with the preset performance threshold. If the cluster task execution performance index of the ith computing node is greater than or equal to the preset performance threshold, the cluster task execution performance index of the ith computing node is judged to be normal. If the cluster task execution performance index of the ith computing node is less than the preset performance threshold, the cluster task execution performance index of the ith computing node is judged to be abnormal. An early warning signal is issued for the abnormal task performance result and the abnormal node number is sent to the corresponding operation and maintenance management personnel for review. The task performance abnormality data is optimized and managed through the cluster task scheduling optimization model.

[0087] In this embodiment, it should be specifically explained that the cluster task scheduling optimization model is as follows:

[0088] maxZ=K1×(1-TED i )-K2×TTR i +K3×Sr i , where maxZ represents

[0089] Output value of the optimization model for cluster task scheduling, TED i It is represented as the cluster task execution deviation of the i-th computing node, TTR i It is represented as the cluster task timeout rate of the i-th computing node, Sr i It is represented as the success rate of cluster task completion of the i-th computing node, K1, K2, K3 are weight factors of cluster task execution deviation, cluster task timeout rate and cluster task completion success rate, respectively, and K1+K2+K3=1;

[0090] In this embodiment, it should be specifically explained that the specific optimization content in step S6 is: reducing the cluster task execution deviation and the cluster task timeout rate, improving the success rate of cluster task completion, and optimizing the scheduling of cluster tasks; importing the cluster task scheduling optimization model output value into the cluster task execution efficiency index in step S5, and obtaining the optimized cluster task execution efficiency index = maxZ×EEI i .

[0091] In this embodiment, it should be specifically explained that the specific optimization content of the cluster task scheduling optimization model is: for nodes with large cluster task execution deviation values ​​and cluster task timeouts, the task allocation strategy is dynamically adjusted according to the real-time load and resource status of the nodes: tasks are assigned to nodes with sufficient resources and low load to reduce task execution deviation and timeout rate; for nodes with a higher success rate in completing cluster tasks, more computing resources are preferentially allocated to increase the probability of their on-time completion.

[0092] Step S7: Establishing a user interface module: Displaying the computer cluster operation status data and the data after the cluster maintenance task optimization management through the user interface module for operation and maintenance personnel to analyze and make decisions.

[0093] Example 2

[0094] See also Figure 2 As shown, the present invention provides a computer cluster maintenance task management system, including a cluster internal node number acquisition module, a cluster node data acquisition module, a cluster operation status analysis module, a cluster operation health index judgment management module, a cluster task execution analysis module, a cluster task execution efficiency index judgment management module and a user interface establishment module.

[0095] The cluster internal node number acquisition module determines the computer cluster as the target area, acquires the number of computing nodes in the cluster, and marks each computing node as 1, 2, ..., i, ..., n in sequence, where i is the number of each computing node and n represents the total number of computing nodes in the cluster;

[0096] The cluster node data collection module is used to monitor and collect the cluster task comprehensive data of each computing node in the cluster, and transmit it to the cluster operation status analysis module and the cluster task execution analysis module, wherein the cluster task comprehensive data includes the cluster operation status data and the cluster task execution data;

[0097] The cluster operation status analysis module is used to analyze the cluster operation status data collected in the cluster node data collection module, calculate the cluster operation health index of each computing node through the operation status mathematical model, and transmit it to the cluster operation health index judgment management module;

[0098] The cluster operation health index judgment management module is used to judge the cluster operation health index, optimize and manage the cluster operation health index through the cluster operation optimization model, and obtain the optimized cluster operation health index;

[0099] The cluster task execution analysis module is used to analyze the cluster task execution data collected in the cluster node data collection module, calculate the cluster task execution efficiency index of each computing node through the task execution mathematical model, and transmit it to the cluster task execution efficiency index judgment management module;

[0100] The cluster task execution efficiency index judgment management module judges the cluster task execution efficiency index, optimizes and manages the cluster task execution efficiency index through the cluster task scheduling optimization model, and obtains the optimized cluster task execution efficiency index;

[0101] The user interface module is established to display the computer cluster operation status data and the data after the cluster maintenance task optimization management through the user interface module for operation and maintenance personnel to analyze and make decisions.

[0102] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0103] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A computer cluster maintenance task management method, characterized in that: The following steps are involved: Step S1: Obtain the number of computing nodes in the cluster: determine the computer cluster as the target area, obtain the number of computing nodes in the cluster, and mark each computing node as 1, 2, ..., i, ..., n in sequence, where i is the number of each computing node and n represents the total number of computing nodes in the cluster; Step S2: Collecting data from each computing node in the cluster: monitoring and collecting cluster task comprehensive data of each computing node in the cluster, wherein the cluster task comprehensive data includes cluster operation status data and cluster task execution data; Step S3: Analyze the cluster operation status data: Analyze the cluster operation status data collected in step S2, and calculate the cluster operation health index of each computing node through the operation status mathematical model; Step S4: judging and managing the cluster operation health index: judging the cluster operation health index, optimizing and managing the cluster operation health index through the cluster operation optimization model, and obtaining the optimized cluster operation health index; Step S5: Analyze the cluster task execution data: Analyze the cluster task execution data collected in step S2, and calculate the cluster task execution efficiency index of each computing node through the task execution mathematical model; Step S6: judging and managing the cluster task execution efficiency index: judging the cluster task execution efficiency index, optimizing and managing the cluster task execution efficiency index through the cluster task scheduling optimization model, and obtaining the optimized cluster task execution efficiency index; Step S7: Establishing a user interface module: Displaying the computer cluster operation status data and the data after the cluster maintenance task optimization management through the user interface module for operation and maintenance personnel to analyze and make decisions.

2. A computer cluster maintenance task management method according to claim 1, characterized in that: The cluster operation status data includes CPU resource utilization, memory resource utilization, disk resource utilization, network bandwidth resource utilization, the number of tasks of computing nodes, the normal operation time of the computer cluster, the total operation time of the computer cluster and the total number of cluster failures; the cluster task execution data includes the actual execution time of the cluster task, the number of successfully completed cluster tasks, the number of cluster task timeouts, the total number of cluster tasks and the cluster task response time.

3. A computer cluster maintenance task management method according to claim 1, characterized in that: The calculation steps of the cluster operation health index of each computing node are as follows: Step S31: Calculate the cluster resource utilization RU of each node i , the calculation model is as follows: Among them, CPU i It is expressed as the CPU resource utilization of the i-th computing node, Memu i It is expressed as the memory resource utilization of the i-th computing node, Disu i Expressed as the disk resource utilization of the i-th computing node, Bwu i It is expressed as the network bandwidth resource utilization rate of the i-th computing node, where i is the number of each computing node; Step S32: Calculate the cluster node load balancing degree NLB of each node i , the calculation model is as follows: Among them, TN i It is represented by the number of tasks of the i-th computing node, n is represented by the total number of computing nodes in the cluster, TN max It is expressed as the maximum number of tasks, that is, the number of tasks of the computing node with the largest number of tasks among all computing nodes, TN min It is expressed as the minimum number of tasks, that is, the number of tasks of the computing node with the least number of tasks among all computing nodes; Step S33: Calculate the cluster operation health index OHI of each node i , the calculation model is as follows: Among them, OHI i Represented as the cluster operation health indicator of the i-th computing node, RU i It is expressed as the cluster resource utilization of the i-th computing node, NLB i It is represented as the cluster node load balancing degree of the i-th computing node, Ut i It is represented as the normal operation time of the computer cluster of the i-th computing node, Trt i Ff is the total running time of the computer cluster of the i-th computing node. i It is represented as the total number of cluster failures of the i-th computing node, n is represented as the total number of computing nodes in the cluster, and e is represented as a natural constant.

4. A computer cluster maintenance task management method according to claim 1, characterized in that: The specific content of judging the cluster operation health index in step S4 is: The cluster operation health index of each computing node is extracted and compared with the preset health threshold. If the cluster operation health index of the ith computing node is greater than or equal to the preset health threshold, the cluster operation health index of the ith computing node is judged to be normal. If the cluster operation health index of the ith computing node is less than the preset health threshold, the cluster operation health index of the ith computing node is judged to be abnormal. An early warning signal is issued for the abnormal cluster operation result and the abnormal node number is sent to the corresponding operation and maintenance management personnel for review. The cluster operation abnormal data is optimized and managed through the cluster operation optimization model.

5. A computer cluster maintenance task management method according to claim 1, characterized in that: The cluster operation optimization model is specifically as follows: maxY=W1×RU i +W2×NLB i +W3×(1-Fr i ), where maxY represents The output value of the cluster running optimization model, RU i It is expressed as the cluster resource utilization of the i-th computing node, NLB i It is represented as the cluster node load balancing degree of the i-th computing node, Fr i It is represented as the cluster operation failure rate of the i-th computing node, W1, W2, and W3 are respectively represented as the weight factors of cluster resource utilization, cluster node load balance, and cluster operation failure rate, and W1+W2+W3=1; The specific optimization content in step S4 is: importing the output value of the cluster operation optimization model into the cluster operation health index in step S3, and obtaining the optimized cluster operation health index = maxY×OHI i .

6. A computer cluster maintenance task management method according to claim 1, characterized in that: The specific steps for calculating the cluster task execution efficiency index of each computing node are as follows: Step S51: Calculate the cluster task execution deviation TED of each node i , the calculation model is as follows: Among them, Aet i It is represented by the actual execution time of the cluster task of the i-th computing node, Ed 预 represents the ideal execution time preset for the cluster task of the i-th computing node, Ctn i It is represented as the number of cluster tasks successfully completed by the i-th computing node; Step S52: Calculate the cluster task timeout rate TTR of each node i , the calculation model is as follows: Among them, Ton i Represented as the number of cluster task timeouts of the i-th computing node, ttn i It is represented as the total number of cluster tasks of the i-th computing node; Step S53: Calculate the cluster task execution efficiency index EEI of each computing node i , the calculation model is as follows: Among them, EEI i It is represented as the cluster task execution efficiency index of the i-th computing node, Ctn i It is expressed as the number of cluster tasks successfully completed by the i-th computing node, rt i Represented as the cluster task response time of the i-th computing node, ttn i It is represented as the total number of cluster tasks of the i-th computing node, TED i It is represented as the cluster task execution deviation of the i-th computing node, TED 预 It is represented by the preset maximum cluster task execution deviation, TTR i It is represented as the cluster task timeout rate of the i-th computing node, and e is represented as a natural constant.

7. A computer cluster maintenance task management method according to claim 1, characterized in that: The specific content of judging the cluster task execution performance index in step S6 is: The cluster task execution performance index of each computing node is extracted and compared with the preset performance threshold. If the cluster task execution performance index of the ith computing node is greater than or equal to the preset performance threshold, the cluster task execution performance index of the ith computing node is judged to be normal. If the cluster task execution performance index of the ith computing node is less than the preset performance threshold, the cluster task execution performance index of the ith computing node is judged to be abnormal. An early warning signal is issued for the abnormal task performance result and the abnormal node number is sent to the corresponding operation and maintenance management personnel for review. The task performance abnormality data is optimized and managed through the cluster task scheduling optimization model.

8. A computer cluster maintenance task management method according to claim 1, characterized in that: The cluster task scheduling optimization model is as follows: maxZ=K1×(1-TED i )-K2×TTR i +K3×Sr i , where maxZ represents Output value of the optimization model for cluster task scheduling, TED i It is represented as the cluster task execution deviation of the i-th computing node, TTR i It is represented as the cluster task timeout rate of the i-th computing node, Sr i It is represented as the success rate of cluster task completion of the i-th computing node, K1, K2, K3 are weight factors of cluster task execution deviation, cluster task timeout rate and cluster task completion success rate, respectively, and K1+K2+K3=1; The specific optimization content in step S6 is: importing the output value of the cluster task scheduling optimization model into the cluster task execution efficiency index in step S5, and obtaining the optimized cluster task execution efficiency index = maxZ × EEI i .

9. A computer cluster maintenance task management system, used to implement a computer cluster maintenance task management method as described in any one of claims 1 to 8, characterized in that: include: Cluster internal node number acquisition module, cluster node data collection module, cluster operation status analysis module, cluster operation health index judgment management module, cluster task execution analysis module, cluster task execution efficiency index judgment management module and user interface establishment module; The cluster internal node number acquisition module determines the computer cluster as the target area, acquires the number of computing nodes in the cluster, and marks each computing node as 1, 2, ..., i, ..., n in sequence, where i is the number of each computing node and n represents the total number of computing nodes in the cluster; The cluster node data collection module is used to monitor and collect the cluster task comprehensive data of each computing node in the cluster, and transmit it to the cluster operation status analysis module and the cluster task execution analysis module, wherein the cluster task comprehensive data includes the cluster operation status data and the cluster task execution data; The cluster operation status analysis module is used to analyze the cluster operation status data collected in the cluster node data collection module, calculate the cluster operation health index of each computing node through the operation status mathematical model, and transmit it to the cluster operation health index judgment management module; The cluster operation health index judgment management module is used to judge the cluster operation health index, optimize and manage the cluster operation health index through the cluster operation optimization model, and obtain the optimized cluster operation health index; The cluster task execution analysis module is used to analyze the cluster task execution data collected in the cluster node data collection module, calculate the cluster task execution efficiency index of each computing node through the task execution mathematical model, and transmit it to the cluster task execution efficiency index judgment management module; The cluster task execution efficiency index judgment management module judges the cluster task execution efficiency index, optimizes and manages the cluster task execution efficiency index through the cluster task scheduling optimization model, and obtains the optimized cluster task execution efficiency index; The user interface module is established to display the computer cluster operation status data and the data after the cluster maintenance task optimization management through the user interface module for operation and maintenance personnel to analyze and make decisions.

Images