A method, equipment, and storage medium for determining inspection strategies before disaster recovery drills.

Abstract

This invention relates to the field of data processing technology, and in particular to a method, device, and storage medium for determining an inspection strategy before a disaster recovery exercise. The method includes: periodically calculating the resource load of the primary and backup databases for each preset indicator; calculating the target performance impact of each candidate inspection item on the primary and backup databases based on the preset resource consumption and resource load of each preset indicator; determining whether a candidate inspection item is a target inspection item based on the target performance impact and the current business importance of the database; and generating a target inspection strategy based on the obtained target inspection items. This invention, through comprehensive analysis of the database's operating status and the impact of the candidate inspection items on the database, can generate a reasonable and reliable inspection strategy, effectively avoiding abnormal situations during disaster recovery exercises, thereby reducing the probability of disaster recovery exercise failure.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, device, and storage medium for determining inspection strategies before disaster recovery drills. Background Technology

[0002] Currently, to ensure that enterprise IT systems and business processes can recover quickly and minimize losses in the event of catastrophic events such as server downtime, data loss, cyberattacks, and natural disasters, users typically conduct disaster recovery drills regularly. Among existing disaster recovery drill methods, patent CN112463455A discloses a disaster recovery drill method, apparatus, equipment, and readable storage medium, which solves the problems of existing disaster recovery drills affecting the normal operation of host services and having low drill efficiency. Patent CN118012676A discloses a disaster recovery drill method, equipment, storage medium, and apparatus, which solves the problem of high time costs associated with resynchronizing business data after the drill.

[0003] The two solutions mentioned above have improved the disaster recovery drill process, but they cannot guarantee the success rate of disaster recovery drills. Due to the long interval between current disaster recovery drills, it is not possible to conduct drills frequently. Therefore, to ensure the success rate of disaster recovery drills, and because current disaster recovery drills are inconsistent with actual scenarios, they are prone to switching failures or excessively long RTO times that lead to business unavailability, it is crucial to eliminate faults in advance and ensure the success of disaster recovery drills. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method, device, and storage medium for determining inspection strategies before disaster recovery drills. By comprehensively analyzing the database's operational status and the impact of the items to be inspected on the database, a reasonable and reliable inspection strategy can be generated, effectively avoiding abnormal situations during disaster recovery drills and thus reducing the probability of disaster recovery drill failure.

[0005] According to a first aspect of the present invention, a method for determining an inspection strategy before a disaster recovery drill is provided, comprising the following steps: S100: Periodically acquire target operating data of the primary database and backup database, and calculate the resource load level of the primary database and backup database for each preset indicator based on the target operating data; the preset indicators include CPU indicators, IO indicators and network indicators.

[0006] S200: For any preset candidate inspection item, calculate the impact of the candidate inspection item on the target performance of the primary database and the backup database based on the preset resource consumption level of the candidate inspection item on each preset indicator and the resource load level of each preset indicator.

[0007] S300 determines whether a candidate inspection item is a target inspection item based on the received current business importance of the primary database, the current business importance of the backup database, and the degree of impact of the candidate inspection item on the target performance of the primary and backup databases, respectively.

[0008] S400: Generate a target inspection strategy based on the obtained target inspection items.

[0009] According to a second aspect of the present invention, a non-transitory computer-readable storage medium is provided, wherein at least one instruction or at least one program is stored therein, the at least one instruction or the at least one program being loaded and executed by a processor to implement the above-described method for determining the inspection strategy before disaster recovery drills.

[0010] According to a third aspect of the present invention, an electronic device is provided, including a processor and the aforementioned non-transitory computer-readable storage medium.

[0011] The present invention has at least the following beneficial effects: This invention provides a method for determining inspection strategies before disaster recovery drills. First, the resource load levels of the primary and backup databases for each preset indicator are periodically calculated. Based on the preset resource consumption and resource load levels of each preset indicator for each candidate inspection item, the impact of the candidate inspection item on the target performance of the primary and backup databases is calculated. Then, based on the impact of the candidate inspection item on the target performance of the primary and backup databases and the current business importance of the primary and backup databases, it is determined whether the candidate inspection item is a target inspection item. Based on the obtained target inspection items, a target inspection strategy is generated. This invention, through comprehensive analysis of the database's operating status and the impact of the candidate inspection items on the database, can screen out suitable target inspection items for execution, generate reasonable and reliable inspection strategies, effectively avoid abnormal situations during disaster recovery drills, and thus reduce the probability of disaster recovery drill failure. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A flowchart illustrating the method for determining the inspection strategy before disaster recovery drills provided in this embodiment of the invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] This invention provides a method for determining the inspection strategy before a disaster recovery drill, such as... Figure 1 As shown, the method includes the following steps: S100: Periodically acquire target operating data for the primary and backup databases, and calculate the resource load level for each preset indicator for the primary and backup databases based on the target operating data. In a specific embodiment, the resource load level can be calculated once per hour.

[0016] Specifically, the preset metrics include CPU metrics, I / O metrics, and network metrics.

[0017] Furthermore, the resource load level of the CPU metric meets the following conditions: S = 100% - t%, where S represents the resource load level of the CPU metric, and t% represents the corresponding percentage of CPU idle time. That is, when calculating the resource load level of the primary database's CPU metric, t% represents the percentage of CPU idle time for the primary database. The percentage of CPU idle time is obtained through the dstat monitoring tool.

[0018] Specifically, the resource load level of the IO metric is the percentage of disk busy time obtained through the iostat monitoring tool.

[0019] Furthermore, the calculation steps for the resource load level of the network indicator are as follows: S101 collects network metrics in real time, including the receive throughput and transmit throughput of the network interface card.

[0020] S102: Based on the maximum value among several receive throughput and several transmit throughput rates within the current period, and the obtained network card speed, calculate the network utilization rate, and use the network utilization rate as a resource load indicator of the network. The network utilization rate is the percentage of the maximum value among several receive throughput and several transmit throughput rates to the network speed.

[0021] The above method calculates network utilization by using the maximum value among several receiving throughput and several sending throughput to obtain the maximum resource load level. Using the maximum resource load level as a benchmark can avoid situations where the resource load level exceeds the carrying capacity during inspection, thus ensuring the safety of the inspection. Furthermore, by calculating the resource load level of multiple indicators for each primary and backup database, it is beneficial to understand the current usage of the primary and backup databases, providing a basis for subsequent inspection item selection.

[0022] S200: For any preset candidate inspection item, calculate the impact of the candidate inspection item on the target performance of the primary database and the backup database based on the preset resource consumption level of the candidate inspection item on each preset indicator and the resource load level of each preset indicator.

[0023] Specifically, the candidate checks include any one of the following: primary / standby database log synchronization thread status anomaly check, standby database application thread status anomaly check, primary / standby database latency check, primary / standby database version consistency check, standby database access to primary database connectivity check, primary database access to standby database reverse connectivity check, primary / standby database role check, primary / standby database configuration file consistency check, and primary / standby database performance consistency check. The preset resource consumption levels for each preset metric in the candidate checks are set by those skilled in the art based on practical experience. For example, the preset resource consumption levels for CPU, IO, and network metrics in the primary / standby database latency check are 20%, 20%, and 40%, respectively; the preset resource consumption levels for CPU, IO, and network metrics in the primary / standby database version consistency check are 20%, 0, and 0%, respectively; the preset resource consumption levels for CPU, IO, and network metrics in the standby database access to primary database connectivity check are 0%, 0, and 20%, respectively; and the preset resource consumption levels for CPU, IO, and network metrics in the primary / standby database performance consistency check are 40%, 40%, and 20%, respectively.

[0024] By checking for several possible anomalies during disaster recovery drills, abnormal situations can be avoided during disaster recovery drills or failover, which helps reduce the probability of disaster recovery drill failure.

[0025] Furthermore, step S200 includes the following steps: S201, for any candidate inspection item and any preset indicator corresponding to the primary database, the sum of the preset resource consumption level of the candidate inspection item on the preset indicator and the resource load level of the preset indicator is taken as the first performance impact level of the candidate inspection item on the preset indicator. For example, the first performance impact level of the primary / standby database latency check on the CPU indicator corresponding to the primary database is the sum of the preset resource consumption level of the standby database latency check on the CPU indicator corresponding to the primary database and the resource load level of the CPU indicator corresponding to the primary database.

[0026] S202, based on n preset indicators corresponding to the main database, determine the maximum value among the n first performance impact degrees corresponding to the candidate inspection items as the target performance impact degree of the candidate inspection items on the main database.

[0027] S203, for any candidate inspection item and any preset index corresponding to the backup database, the sum of the preset resource consumption degree of the candidate inspection item on the preset index and the resource load degree of the preset index is taken as the second performance impact degree of the candidate inspection item on the preset index; it can be understood that the calculation method of the second performance impact degree is the same as that of the first performance impact degree.

[0028] S204, based on n preset indicators corresponding to the backup database, the maximum value among the n second performance impact degrees corresponding to the candidate inspection item is determined as the target performance impact degree of the candidate inspection item on the backup database.

[0029] As mentioned above, since CPU, I / O, and network metrics can all affect the operation of the primary and backup databases, the maximum value should be used as the benchmark when calculating the degree of performance impact. That is, if the estimated performance impact of the candidate check item on any preset metric exceeds the carrying capacity threshold, it is more likely to be disallowed. Furthermore, when calculating the target performance impact, the impact of the check item on the database is added on the basis of the current database operation status. The performance impact is more reasonable and accurate based on both internal and external factors.

[0030] S300: Based on the received current business importance of the primary database, the current business importance of the backup database, and the impact of the candidate inspection items on the target performance of the primary and backup databases, determine whether a candidate inspection item is a target inspection item. In specific implementation, the current business importance is the level of importance obtained by those skilled in the art after rating and scoring the current business importance of the primary and backup databases.

[0031] In one specific embodiment, step S300 includes the following steps: S301. Based on the current business importance of the main database, the impact of the candidate inspection items on the target performance of the main database, and the preset weights corresponding to the business importance and the impact on the target performance, calculate the first execution interference score corresponding to the candidate inspection items.

[0032] Specifically, the first execution interference score F1 meets the following conditions: F1 = W1 × K1 + W2 × K2, where K1 is the importance of the current business of the main database, K2 is the impact of the candidate inspection items on the target performance of the main database, W1 is the preset weight corresponding to the importance of the business, and W2 is the preset weight corresponding to the impact on the target performance.

[0033] S302. Based on the current business importance of the backup database, the impact of the candidate inspection items on the target performance of the backup database, and the preset weights corresponding to the business importance and the impact on the target performance, calculate the second execution interference score corresponding to the candidate inspection items.

[0034] Specifically, the second execution interference score F2 meets the following conditions: F2 = W1 × K3 + W2 × K4, where K3 is the current business importance of the backup database, and K4 is the impact of the candidate inspection items on the target performance of the backup database.

[0035] S303, when the maximum value of the first execution interference score and the second execution interference score corresponding to the candidate inspection item is less than a preset scoring threshold, the candidate inspection item is determined to be the target inspection item. Those skilled in the art can set the preset scoring threshold according to actual needs, which will not be elaborated here.

[0036] As mentioned above, when determining whether a candidate inspection item is a target inspection item, the importance of the current business is also considered in addition to the impact of the inspection item on the database performance. During peak business hours or when core business is running, and when the impact on database performance is high, the inspection priority of the inspection item should be reduced to avoid cumulative performance risks. Furthermore, the candidate inspection item is only determined to be a target inspection item when the maximum score value is greater than the preset score threshold, so as to avoid excessive impact on database performance. This effectively ensures the smooth inspection of the inspection items and reduces the risk of system business failures.

[0037] S400: Generate a target inspection strategy based on the obtained target inspection items.

[0038] Specifically, the S400 procedure includes the following steps: S401, obtain the attribute information corresponding to each candidate check item; the attribute information includes the performance impact parameters of executing the candidate check item, the estimated time consumption, and the dependency relationship with other candidate check items; it can be understood that the dependency relationship between a candidate check item and other candidate check items refers to the dependency relationship between a candidate check item and any other candidate check item besides itself. For example, the primary / secondary database log synchronization thread status anomaly check must be performed before the primary / secondary database delay check can be performed, because if the thread does not exist, the delay value is meaningless.

[0039] Specifically, the estimated time consumption is the average of several times consumed based on several trials or historical data of performing the candidate inspection items.

[0040] S402, Based on historical operation and maintenance data, determine the correlation probability between each candidate inspection item and the failure of the disaster recovery exercise; the historical operation and maintenance data includes historical logs and historical fault report texts. For example, historical operation and maintenance data shows that the probability of the disaster recovery exercise failing due to an anomaly in the primary and standby database latency check is 80%, while the probability of the disaster recovery exercise failing due to an anomaly in the primary and standby database version consistency check is 5%.

[0041] Preferably, the grey relational analysis method is used to calculate the association probability based on historical operation and maintenance data. Since grey relational analysis is a conventional method for calculating association degree, those skilled in the art are familiar with the specific implementation methods for calculating association probability using grey relational analysis, and will not be described in detail here.

[0042] S403, based on the performance impact parameters, estimated time consumption, and correlation probability between each candidate inspection item and the failure of the disaster recovery exercise, the target weight value of each candidate inspection item is calculated.

[0043] Specifically, the calculation steps for the target weight value of any of the candidate inspection items are as follows: S4031, for any candidate inspection item, the average value of the preset resource consumption degree of the candidate inspection item on several preset indicators is determined as the performance impact parameter corresponding to the candidate inspection item.

[0044] S4032, the target weight value of the candidate inspection item is obtained by weighted summing based on three dimensions: the performance impact parameter, the estimated time consumption, and the correlation probability between the candidate inspection item and the failure of the disaster recovery exercise, as well as the preset weight coefficient corresponding to each dimension; among them, the performance impact parameter and the estimated time consumption are inversely proportional to the target weight value, and the correlation probability is directly proportional to the target weight value.

[0045] Specifically, the target weight value W of any of the candidate inspection items meets the following conditions: W = α × (1 - C) + β × (1 - T0) + γ × P, where α is the preset weight coefficient corresponding to the performance impact parameter, β is the preset weight coefficient corresponding to the estimated time consumption, γ is the preset weight coefficient corresponding to the correlation probability, C is the performance impact parameter corresponding to the candidate inspection item, T0 is the normalized value corresponding to the estimated time consumption, and P is the correlation probability.

[0046] As mentioned above, since the inspection items have a significant impact on database performance and take a long time, the probability of failure is relatively high, which will affect subsequent inspections. Therefore, the inspection order should be lower, while those with a high probability of correlation with the failure of disaster recovery drills should be inspected first to promptly detect and eliminate problems and avoid affecting subsequent inspections. Based on this, the above target weight values ​​are obtained, making the target weight values ​​of the candidate inspection items more reasonable and reliable.

[0047] S404 constructs a directed graph of check items, treating each candidate check item as a node and its dependencies as edges. In the specific implementation, for two candidate check items without dependencies, the edges between them are undirected, meaning the order of execution is not restricted.

[0048] S405. A graph optimization algorithm is used to sort the nodes in the directed graph of the inspection items, generate an inspection item sequence that satisfies the dependency constraints and has the optimal target weight value, and use the generated inspection item sequence as the target inspection strategy.

[0049] Specifically, optimal target weight ranking means arranging the candidate inspection items in descending order of target weight value, while satisfying dependency constraints. It can be understood as follows: based on the above method of obtaining target weight values, the target inspection item with the larger the target weight value should be inspected earlier.

[0050] As described above, after obtaining several target inspection items, each inspection item is ranked by comprehensively considering its performance impact on the database, estimated time consumption, and correlation probability with the disaster recovery exercise. Inspection items with high failure probability and low time consumption are executed first, allowing for the discovery of potential problems via the shortest path. This significantly improves inspection efficiency. Furthermore, by considering the performance impact of inspection items on the database, the execution order is optimized, minimizing the performance interference of the inspection process on the database itself. Through dependency constraints, the reliability and accuracy of the inspection results are improved, effectively avoiding abnormal situations during the disaster recovery exercise and greatly reducing the probability of failure.

[0051] Furthermore, the method also includes the following steps: S10, based on historical data, obtain several historical disaster recovery exercise results and the most recent database check results before each historical disaster recovery exercise.

[0052] S20: When both the historical disaster recovery exercise result and the most recent database check result before the historical disaster recovery exercise are successful or both are unsuccessful, accumulate 1 success based on the original success count; otherwise, accumulate 0 successes to obtain the current success count corresponding to each historical disaster recovery exercise result; the initial value of the original success count is 0.

[0053] S30 updates the inspection success rate based on the current number of successes corresponding to each historical disaster recovery exercise result, and generates an inspection report text based on the changing trend of the inspection success rate.

[0054] The above-mentioned method of generating an inspection success rate by statistically analyzing the number of times the inspection results match the disaster recovery exercise results is beneficial for users to review their inspection strategies and add new inspection items, so that the inspection success rate can be continuously iterated and gradually increased. On this basis, it can also improve the probability of success of disaster recovery exercises.

[0055] Embodiments of the present invention also provide a non-transitory computer-readable storage medium, which can be disposed in an electronic device to store at least one instruction or at least one program related to implementing a method in the method embodiment. The at least one instruction or the at least one program is loaded and executed by the processor to implement the disaster recovery drill pre-examination inspection strategy determination method provided in the above embodiments.

[0056] Embodiments of the present invention also provide an electronic device, including a processor and the aforementioned non-transitory computer-readable storage medium.

[0057] While specific embodiments of the invention have been described in detail by way of example, those skilled in the art should understand that the examples are for illustrative purposes only and not intended to limit the scope of the invention. It should also be understood that various modifications can be made to the embodiments without departing from the scope and spirit of the invention. The scope of the invention is defined by the appended claims.

Claims

1. A method for determining inspection strategies before disaster recovery drills, characterized in that, The method includes the following steps: S100: Periodically acquire target operating data of the primary database and backup database, and calculate the resource load level of the primary database and backup database for each preset indicator based on the target operating data; the preset indicators include CPU indicators, IO indicators and network indicators. S200: For any preset candidate inspection item, calculate the impact of the candidate inspection item on the target performance of the primary database and the backup database based on the preset resource consumption level of the candidate inspection item on each preset indicator and the resource load level of each preset indicator. S300: Based on the received current business importance of the primary database, the current business importance of the backup database, and the impact of the candidate inspection items on the target performance of the primary and backup databases, determine whether the candidate inspection items are target inspection items. S400 generates a target inspection strategy based on the obtained target inspection items.

2. The method for determining the inspection strategy before disaster recovery drills according to claim 1, characterized in that, The optional checks are any one of the following: check for abnormal status of the primary / secondary database log synchronization thread, check for abnormal status of the secondary database application thread, check for primary / secondary database latency, check for primary / secondary database version consistency, check for connectivity between the secondary database and the primary database, check for reverse connectivity between the primary database and the secondary database, check for primary / secondary database roles, check for consistency of primary / secondary database configuration files, and check for consistency of primary / secondary database performance.

3. The method for determining the inspection strategy before disaster recovery drills according to claim 1, characterized in that, The calculation steps for the resource load level of the network indicator are as follows: S101 collects network metrics in real time, including the receive throughput and transmit throughput of the network interface card. S102, calculate network utilization based on the maximum value among several receiving throughput and several transmitting throughput in the current period, as well as the obtained network card speed, and use network utilization as a resource load level of network indicators.

4. The method for determining the inspection strategy before disaster recovery drills according to claim 1, characterized in that, The S200 procedure includes the following steps: S201, for any candidate inspection item and any preset indicator corresponding to the main database, the sum of the preset resource consumption degree of the candidate inspection item on the preset indicator and the resource load degree of the preset indicator is taken as the first performance impact degree of the candidate inspection item on the preset indicator. S202, based on n preset indicators corresponding to the main database, the maximum value among the n first performance impact degrees corresponding to the candidate inspection items is determined as the target performance impact degree of the candidate inspection items on the main database; S203, for any candidate inspection item and any preset index corresponding to the backup database, the sum of the preset resource consumption degree of the candidate inspection item on the preset index and the resource load degree of the preset index is taken as the second performance impact degree of the candidate inspection item on the preset index. S204, based on the n preset indicators corresponding to the backup database, determine the maximum value among the n second performance impact degrees corresponding to the candidate inspection item as the target performance impact degree of the candidate inspection item on the backup database.

5. The method for determining the inspection strategy before disaster recovery drills according to claim 1, characterized in that, The S300 procedure includes the following steps: S301. Based on the current business importance of the main database, the impact of the candidate inspection items on the target performance of the main database, and the preset weights corresponding to the business importance and the impact on the target performance, calculate the first execution interference score corresponding to the candidate inspection items. S302, based on the current business importance of the backup database, the impact of the candidate inspection items on the target performance of the backup database, and the preset weights corresponding to the business importance and the impact on the target performance, calculate the second execution interference score corresponding to the candidate inspection items; S303, when the maximum value of the first execution interference score and the second execution interference score corresponding to the candidate inspection item is less than the preset score threshold, the candidate inspection item is determined to be the target inspection item.

6. The method for determining the inspection strategy before disaster recovery drills according to claim 1, characterized in that, The S400 procedure includes the following steps: S401, Obtain the attribute information corresponding to each candidate inspection item; the attribute information includes the performance impact parameters of executing the candidate inspection item, the estimated time consumption, and the dependency relationship with other candidate inspection items; S402, Based on historical operation and maintenance data, determine the correlation probability between each candidate inspection item and the failure of the disaster recovery exercise; the historical operation and maintenance data includes historical logs and historical fault report texts; S403, based on the performance impact parameters, time consumption estimates, and the correlation probability between each candidate inspection item and the failure of the disaster recovery exercise, the target weight value of each candidate inspection item is calculated. S404: Construct a directed graph of inspection items by treating each candidate inspection item as a node and the dependency relationship as an edge. S405. A graph optimization algorithm is used to sort the nodes in the directed graph of the inspection items, generate an inspection item sequence that satisfies the dependency constraints and has the optimal target weight value, and use the generated inspection item sequence as the target inspection strategy.

7. The method for determining the inspection strategy before disaster recovery drills according to claim 6, characterized in that, The steps for calculating the target weight value of any of the candidate inspection items are as follows: S4031, for any candidate inspection item, the average value of the preset resource consumption degree of the candidate inspection item on several preset indicators is determined as the performance impact parameter corresponding to the candidate inspection item. S4032, the target weight value of the candidate inspection item is obtained by weighted summing based on three dimensions: the performance impact parameter, the estimated time consumption, and the correlation probability between the candidate inspection item and the failure of the disaster recovery exercise, as well as the preset weight coefficient corresponding to each dimension; among them, the performance impact parameter and the estimated time consumption are inversely proportional to the target weight value, and the correlation probability is directly proportional to the target weight value.

8. The method for determining the inspection strategy before disaster recovery drills according to claim 1, characterized in that, The method further includes the following steps: S10, based on historical data, obtain several historical disaster recovery exercise results and the most recent database check results before each historical disaster recovery exercise; S20: When both the historical disaster recovery exercise result and the most recent database check result before the historical disaster recovery exercise are successful or both are unsuccessful, accumulate 1 success on the basis of the original success count; otherwise, accumulate 0 successes to obtain the current success count corresponding to each historical disaster recovery exercise result. The initial value for the original number of successes is 0; S30 updates the inspection success rate based on the current number of successes corresponding to each historical disaster recovery exercise result, and generates an inspection report text based on the changing trend of the inspection success rate.

9. A non-transitory computer-readable storage medium, wherein the storage medium stores at least one instruction or at least one program segment, characterized in that, The at least one instruction or the at least one program segment is loaded and executed by the processor to implement the method for determining the inspection strategy before disaster recovery drills as described in any one of claims 1-8.

10. An electronic device, characterized in that, Includes a processor and the non-transitory computer-readable storage medium as described in claim 9.

Images