Data backup method and device, and electronic device

Abstract

This disclosure relates to the field of intelligent execution technology for backup tasks in the cloud computing and big data domain, and particularly to a data backup method, apparatus, and electronic device. The data backup method, applied in a management and scheduling center, includes: sending an executability prediction request for a target backup task to a backup task checker; receiving the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; determining the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, and the response time of the executability prediction request; and determining whether the final executability result meets the backup conditions. This disclosure achieves accurate prediction and strategy optimization before backup task execution, effectively improving the reliability and success rate of backup task execution.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent execution technology for backup tasks in the field of cloud computing and big data, and in particular to a data backup method, apparatus, and electronic device. Background Technology

[0002] Backup technology is a core disaster prevention and mitigation measure in the computer system field to ensure data security and support continuous business operations. It can effectively resist data loss and corruption caused by factors such as hardware and software failures, human error, malicious tampering, hacker attacks, virus intrusions, and force majeure. Its core purpose is to maintain the continuity of business systems after a disaster and minimize unplanned downtime. Currently, when backup tasks fail, the approach is often to first locate the problem and then troubleshoot the fault, or to directly resume the backup task after the network is restored to complete the task.

[0003] However, the existing handling method has three major drawbacks due to its post-problem response logic: First, it is easy to waste system resources, such as causing redundant reading and writing of large files; second, the adaptability to abnormal scenarios is fixed and cannot cover all types of abnormalities (such as time consistency anomalies, database transaction conflicts, etc.); third, the simple retry strategy after failure will bring exponential time delay and storage resource consumption. Summary of the Invention

[0004] In view of the above-mentioned technical problems, this disclosure provides a data backup method, apparatus, and electronic device.

[0005] Firstly, this disclosure provides a data backup method applied to a management and scheduling center, the method comprising: Send an executability prediction request for the target backup task to the backup task inspector; Receive the executability prediction results of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; Based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request, determine the final executability result of the target backup task and its corresponding final execution strategy. Determine whether the final executable result meets the backup conditions; In response to the fulfillment of backup conditions, a backup command is issued to the backup task executor so that the backup task executor executes the target backup task in accordance with the final execution strategy.

[0006] Furthermore, according to the method of the first aspect of this disclosure, the executability prediction results and their corresponding execution strategies are generated based on the following: Obtain environmental metrics of the source device, operational metrics of the target device, and historical backup task execution logs; Environmental metrics, operational metrics, and historical backup task execution logs are input into the status assessment model to obtain the probability distribution of the feasibility of executing the target backup task. Based on the probability distribution of execution feasibility, the executability prediction results of the target backup task and its corresponding execution strategy are determined.

[0007] In addition, according to the method of the first aspect of this disclosure, the environmental indicators include: network connectivity, permission verification status, resource utilization, backup path anomaly marking, and time consistency; the operational indicators include: available storage capacity, network latency, and IO performance; and the historical backup task execution logs include: backup task success rate, anomaly type statistics, retry count and results.

[0008] Furthermore, according to the method of the first aspect of this disclosure, execution feasibility includes executable, non-executable, and partially executable; Based on the probability distribution of execution feasibility, the executability prediction results of the target backup task and its corresponding execution strategy are determined, including: Based on the probability distribution of execution feasibility, the execution feasibility with the highest probability in the probability distribution or the execution feasibility with a probability exceeding a preset probability threshold is determined as the executability prediction result of the target backup task. Based on the executability prediction results, the corresponding execution strategy is determined.

[0009] Furthermore, according to the method of the first aspect of this disclosure, the method also includes: When the executability prediction result is executable, the corresponding execution strategy is to immediately execute the backup jobs for all backup paths and add the target backup task to a high-priority queue for priority transmission. When the executability prediction result is partially executable, the corresponding execution strategy is to skip the abnormal path according to the preset rules and trigger the asynchronous repair task; When the feasibility prediction result is that the task is not executable, the corresponding execution strategy is to terminate the target backup task and start the fault self-healing process.

[0010] Furthermore, according to the method of the first aspect of this disclosure, the preset rule is as follows: Temporary files are automatically skipped without alerts. For non-core logs, skip them and record them, then summarize and report them daily. For the user data directory, skip it and immediately send an email alert; For core database files, terminate backups and trigger manual intervention.

[0011] Furthermore, according to the method of the first aspect of this disclosure, based on at least one executability prediction result and its corresponding execution strategy, and the response time of the executability prediction request, the final executability result of the target backup task and its corresponding final execution strategy are determined, including: If the response time for an executability prediction request does not exceed a preset response time, and if multiple consecutive executability predictions result in executability, and the number of predictions reaches the executability determination threshold, then the executability prediction result and its corresponding execution strategy will be used as the final executability result and its corresponding final execution strategy; and / or, If the response time to the executability prediction request does not exceed the preset response time, the executability prediction result is partially executable. The executability prediction result and its corresponding execution strategy are then used as the final executability result and its corresponding final execution strategy for the target backup task; and / or, If the response time of the executability prediction request exceeds the preset response time, or if the executability prediction result is unexecutable, it is recorded as an unexecutable determination. When the cumulative number of unexecutable determinations reaches the preset unexecutable determination threshold, the unexecutability and the corresponding execution strategy are determined as the final executability result and final execution strategy of the target backup task.

[0012] Furthermore, according to the method of the first aspect of this disclosure, the method also includes: If the executability prediction result is partially executable, the backup task checker returns a list of skipped abnormal paths based on the executability prediction request.

[0013] Secondly, this disclosure provides a data backup device for use in a management and dispatch center, comprising: The sending unit is used to send an executability prediction request for the target backup task to the backup task checker; The receiving unit is used to receive the executability prediction results of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request. The determining unit is used to determine the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request. The judgment unit is used to determine whether the final executability result meets the backup conditions; The issuing unit is used to issue backup instructions to the backup task executor in response to the fulfillment of backup conditions. The backup instructions include the final execution strategy corresponding to the final executability result of the backup conditions being met.

[0014] Thirdly, this disclosure provides a data backup method applied to a backup task executor, comprising: Receive backup instructions from the management and scheduling center. The backup instructions include the final executability result of the target backup task and its corresponding final execution strategy. In response to backup commands, execute the target backup task according to the final execution strategy; The backup instruction generation process includes: the management and scheduling center sending an executability prediction request for the target backup task to the backup task checker; receiving the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; determining the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request; determining whether the final executability result meets the backup conditions; and if the backup conditions are met, issuing a backup instruction containing the final executability result and the final execution strategy to the backup task executor.

[0015] Furthermore, according to the method of the third aspect of this disclosure, the method also includes: If a process blocking issue occurs during the execution of the target backup task, the target backup task will be executed based on a retry and wait strategy. The retry and wait policies include setting the maximum number of retries, the retry interval, and the maximum number of waits. Based on these policies, the target backup task is executed as follows: Wait according to the retry interval, and start retrying the target backup task. Each retry is counted as one wait. If the target backup task still fails to resume normal execution within the maximum number of retries, or when the maximum number of waits is reached, terminate the execution of the target backup task and report the error.

[0016] Furthermore, according to the method of the third aspect of this disclosure, the backup task executor includes a source device and a target device; Execute the target backup task according to the final execution strategy, including: The source device divides the data to be backed up in the target backup task into high-priority data and low-priority data according to the importance of the data. The high-priority data is transmitted synchronously in real time, and the low-priority data is transmitted when the bandwidth is idle.

[0017] Furthermore, according to the method of the third aspect of this disclosure, the method also includes: After the target backup task is completed, the execution results and exception information are collected and fed back to the management and scheduling center.

[0018] Fourthly, this disclosure provides a data backup device applied to a backup task executor, comprising: The receiving unit is used to receive backup instructions issued by the management and scheduling center. The backup instructions include the final executability result of the target backup task and its corresponding final execution strategy. The execution unit is used to execute the target backup task in response to the backup command and according to the final execution strategy; The backup instruction generation process includes: the management and scheduling center sending an executability prediction request for the target backup task to the backup task checker; receiving the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; determining the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request; determining whether the final executability result meets the backup conditions; and if the backup conditions are met, issuing a backup instruction containing the final executability result and the final execution strategy to the backup task executor.

[0019] Fifthly, this disclosure provides an electronic device, including: a memory for storing computer-readable instructions; and a processor for executing the computer-readable instructions, causing the electronic device to perform the method as described in any of the embodiments of the first or fourth aspect.

[0020] This disclosure provides a data backup method, apparatus, and electronic device. The method involves a management and scheduling center initiating an executability prediction request for a target backup task to a backup task checker. The final executability result and execution strategy are determined by combining the prediction results, corresponding execution strategies, and request response time. After determining that the backup conditions are met, a backup instruction is issued to the backup task executor, enabling the executor to execute the target backup task according to the final execution strategy. This achieves accurate prediction and strategy optimization before backup task execution, effectively improving the reliability and success rate of backup task execution, avoiding system resource waste and redundant large file read / write operations caused by backup execution failures, significantly reducing time delays and additional storage resource consumption, thereby ensuring stable and reliable data backup, enhancing the disaster prevention and mitigation capabilities of business systems, and minimizing the risk of unplanned downtime.

[0021] It should be understood that both the foregoing general description and the following detailed description are exemplary and intended to provide further illustration of the claimed technology. Attached Figure Description

[0022] The above and other objects, features, and advantages of this disclosure will become more apparent from the more detailed description of the embodiments thereof in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this disclosure and form part of the specification. They are used together with the embodiments of this disclosure to explain the disclosure and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0023] Figure 1 A schematic diagram of a data backup system provided in an embodiment of this disclosure; Figure 2 A flowchart illustrating a data backup method provided in an embodiment of this disclosure; Figure 3 A flowchart illustrating yet another data backup method provided in this disclosure embodiment; Figure 4 A flowchart illustrating yet another data backup method provided in this disclosure embodiment; Figure 5 A flowchart illustrating yet another data backup method provided in this disclosure embodiment; Figure 6 A flowchart illustrating the executability prediction method provided in this embodiment of the disclosure; Figure 7 A structural block diagram of a data backup device provided in an embodiment of this disclosure; Figure 8 A structural block diagram of another data backup device provided in the embodiments of this disclosure; Figure 9 This is a hardware block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this disclosure more apparent, exemplary embodiments according to this disclosure will now be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this disclosure, and not all embodiments of this disclosure. It should be understood that this disclosure is not limited to the exemplary embodiments described herein.

[0025] The success rate of existing backup jobs is affected by various factors, such as the source machine being offline or off-network, the backup job path not existing or being occupied, file corruption, backup software program malfunctions, and insufficient permissions on the source machine. These issues can all lead to backup job failures during operation. In the industry, backup job failures are typically addressed by operations and maintenance personnel through log analysis and a step-by-step elimination process to locate the problem, while simultaneously optimizing backup strategies (such as incremental backups and data compression) to improve reliability. This approach is based on troubleshooting and debugging after a backup job failure and is considered an error recovery measure.

[0026] Professional backup vendors such as Aishu Software use technologies such as breakpoint reconnection and breakpoint resume for offline and off-network scenarios, allowing backup operations to continue when the source device is back online within a specified time.

[0027] Mainstream cloud vendors such as AWS Backup provide cross-region replication that automatically handles cross-region network issues, monitors backup task status in real time, and triggers alarms; Alibaba Cloud Hybrid Cloud Backup (HBR) supports backup task priority scheduling to ensure that critical business data is transmitted first; Dell EMC PowerProtect accelerates data transmission, reduces the risk of network interruption, and integrates with Kubernetes to achieve consistent backup of containerized applications.

[0028] Based on the implementation solutions of mainstream cloud vendors and professional backup vendors, the following two methods can be summarized to improve the success rate of backup jobs: 1. For scenarios involving network jitter or disconnection, technical means are used to support the continued execution of backup operations after the network environment is restored. Real-time monitoring is also used to detect problems and issue alarms, and other measures are taken to restore the network environment.

[0029] 2. After a backup job fails, troubleshoot, locate, and resolve the issue, and establish a product knowledge base to continuously optimize product capabilities for known problems, thereby improving the success rate of product job execution. Additionally, by detecting known problem scenarios, such as communication connection anomalies at the backup source due to network outages or client offline status, the task will be directly terminated upon initiation.

[0030] Existing solutions for backup job failures, while improving the fault tolerance of backup tasks, still involve performing layer-by-layer verification of file paths and attributes when the task is triggered, i.e., real-time verification and repeated verification after interruption during the operation. This has significant drawbacks, namely, post-processing of problems leads to resource waste: when the file size is large, it will consume a long time of redundant read and write operations before the system reports an error, causing the transmitted data to become invalid; if the application attempts to overwrite a running locked document, it may cause target file conflicts or abnormal source file data.

[0031] While some solutions attempt to detect problems in advance, their problem scenarios are fixed and lack dynamic coverage: common error scenarios typically include individual, routine triggering conditions such as whether the disk is full or whether the file path can be recognized, without a systematic verification process to ensure comprehensive problem investigation; if a full-scale logic is initiated while the original database transaction is writing, it may disrupt the temporal continuity of the dataset; differences in system time zones may lead to errors in the last modification comparison, making it impossible to detect the true change status, etc., which are often difficult to detect; the resulting failure retries often bring exponential time delays and higher total storage costs.

[0032] To address the aforementioned technical issues, this disclosure provides a data backup method.

[0033] Figure 1 A schematic diagram of a data backup system provided in an embodiment of this disclosure.

[0034] like Figure 1 As shown, the data backup system 100 may include a management and scheduling center 101, a backup task checker 102, and a backup task executor 103.

[0035] Understandably, before providing services, the management and dispatch center 101 binds to and authenticates the target storage media in the target device. If the target storage media is used as the primary storage medium, it pre-binds to the AS / SK of multiple target storage media across multiple clusters. After AK authentication is successful, the management and dispatch center gains the authority to store and retrieve data from the target storage media. Through platform-side AK authentication, actual tenants in the backup source environment do not need to re-bind the target storage media and related authentication information during use, thus achieving centralized management of tenant data in the dispatch center. The backup task checker 102 is integrated into the management and scheduling center 101. The managed backup data is encapsulated into an API and added to the plugin of the backup task checker 102, allowing it to obtain real-time information from the scheduling center, including: number of clients (online / offline), application and host job tasks (total / success / failure / in progress / partial success), backup source data volume (connected, not connected, scanning, unknown), data capacity of newly added / released / net-added application datasets / host backup datasets / storage pools, number of recovery attempts (total / success / failure) and the amount of data recovered, number of task executions (backup / recovery, total / success / failure), average speed, peak / valley speed, traffic, etc. The backup task checker 102 automatically checks the environment of the source device to be backed up, possessing scanning, automatic waiting, and retry capabilities, and is an intelligent application capable of predicting whether a job is executable. Before job execution, the backup task checker 102 captures source environment problems that could cause abnormal interruptions to the backup task. It then returns the predicted results of the check information to the management and scheduling center 101, including whether an anomaly occurred, a summary of the anomaly, and suggested solutions. This avoids direct interruption when encountering anomalies, optimizes job performance, improves job success rate, reduces unnecessary storage and network resource consumption, and lowers I / O overhead. The backup task executor 103 is an intelligent application based on the backup task checker 102, capable of automatic waiting, retrying, and automatically skipping blocked paths during job execution. Based on the identified situation and solutions, during backup task execution, it can automatically skip blocked paths such as occupied files, automatically wait and retry jobs on that path, and if the backup is still unavailable due to being occupied after the retry count, it skips the task directly.

[0036] Specifically, the management and scheduling center 101 acts as the core control unit, responsible for coordinating and scheduling the overall process, connecting various components, and ensuring the orderly execution of tasks. Before initiating a backup job, the backup task checker 102 checks the source device environment and target device to be backed up, predicts whether the backup task is executable, and then returns the prediction result to the management and scheduling center 101. The management and scheduling center 101 determines whether the backup conditions are met based on the returned prediction result: if the detection status is executable or partially executable, it issues an execution command to the backup task executor 103; if there is an abnormality, it terminates the process and issues an abnormality alarm.

[0037] Figure 2 A data backup method provided for embodiments of this disclosure. For example... Figure 2 As shown, this method is executed by the management and scheduling center and specifically includes the following steps: Step 201: Send an executability prediction request for the target backup task to the backup task checker.

[0038] In one embodiment of this disclosure, the target backup task refers to the task that currently requires a backup operation. Before issuing the actual backup instruction, the management and scheduling center sends an executability prediction request to the backup task checker for the target backup task to be executed, so that the backup task checker can determine in advance whether the target backup task has the conditions for execution.

[0039] For example, as a further step, the executability prediction request carries backup-related information such as the backup type of the target backup task, the target device path, and the source device path.

[0040] Specifically, a target backup task refers to a task that currently requires backup. Before issuing the actual backup instruction, the management and scheduling center sends an executability prediction request to the backup task checker for the target backup task to be executed. The backup task checker parses the executability prediction request and pre-determines whether the target backup task meets the conditions for execution.

[0041] Step 202: Receive the executability prediction results of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request.

[0042] In one embodiment of this disclosure, the execution strategy refers to the execution rules determined for the target backup task based on the executability prediction results. After receiving the executability prediction request, the backup task checker makes a preliminary judgment on the executability of the target backup task, and generates a corresponding executability prediction result and execution strategy based on the prediction result. Then, it returns both to the management and scheduling center for reception.

[0043] Step 203: Based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request, determine the final executability result of the target backup task and its corresponding final execution strategy.

[0044] In one embodiment of this disclosure, the executability results obtained from one or more predictions, the corresponding execution strategies, and the response time of each prediction request are comprehensively weighted and judged to make a final prediction on the executability of the target backup task. Based on the final prediction result, a final executability result and a corresponding final execution strategy are generated. The response time can be used to assess the load of the backup task checker and the network status, avoiding the use of unreliable prediction results when the checker is busy or abnormal.

[0045] Step 204: Determine whether the final executable result meets the backup conditions.

[0046] In one embodiment of this disclosure, the backup conditions include whether the final executability result is executable or partially executable. By comparing the final executability result with preset backup execution conditions, it is determined whether the target backup task meets the criteria for allowed execution, thereby deciding whether to issue an actual backup execution instruction. Partially executable means that only a portion of the data is backed up, rather than a complete backup.

[0047] Step 205: In response to the fulfillment of the backup conditions, a backup instruction is issued to the backup task executor so that the backup task executor executes the target backup task in accordance with the final execution strategy.

[0048] In one embodiment of this disclosure, when the final executability result meets the backup conditions, the management and scheduling center issues a backup instruction to the backup task executor. The backup instruction carries a final execution strategy corresponding to the current final executability result, so that the backup task executor executes the target backup task in accordance with the final execution strategy.

[0049] In summary, according to the technical solution provided in the embodiments of this disclosure, this disclosure initiates an executability prediction request for the target backup task to the backup task checker through the management and scheduling center. The final executability result and the final execution strategy are determined by combining the prediction result, the corresponding execution strategy, and the request response time. After determining that the backup conditions are met, a backup instruction is issued to the backup task executor so that the backup task executor executes the target backup task according to the final execution strategy. This achieves accurate prediction and strategy optimization before the backup task is executed, effectively improving the reliability and success rate of the backup task execution, avoiding the waste of system resources and redundant reading and writing of large files caused by backup execution failure, significantly reducing time delay and additional storage resource consumption, thereby ensuring stable and reliable data backup, enhancing the disaster prevention and mitigation capabilities of the business system, and minimizing the risk of unplanned downtime.

[0050] Figure 3 This is a flowchart illustrating another data backup method provided in an embodiment of the present disclosure.

[0051] like Figure 3 As shown, the specific steps for generating the executability prediction results and their corresponding execution strategies are as follows: Step 301: Obtain the environmental indicators of the source device, the operating indicators of the target device, and the historical backup task execution logs.

[0052] In one embodiment of this disclosure, environmental metrics include: network connectivity, permission verification status, resource utilization, backup path anomaly markers, and time consistency; operational metrics include: available storage capacity, network latency, and IO performance; historical backup task execution logs include: backup task success rate, anomaly type statistics, retry counts and results. Network connectivity can be obtained through ICMP Ping + TCP port probing (e.g., database port 3306). Permission verification status can be obtained by simulating read / write test files (file backup) / executing SELECT 1 (database backup). Resource utilization can be obtained by monitoring CPU / memory / disk IO and calculating the resource pressure index. Path anomalies can be obtained by traversing the target backup task's path and marking locked files and paths with insufficient permissions. Time consistency can be determined by checking whether the system time of the source device and the target device are consistent. Available storage capacity can be determined by checking the available storage space size of the target device. Network latency can be determined by executing a Ping command on the backup server to send network probe packets to the gateway where the target device is located. IO performance can be determined by testing the random write performance of the target device. The source device refers to the device where the data to be backed up resides, and the target device refers to the device on which the data to be backed up will be written. From three dimensions—the data source side (i.e., the source device), the storage target side (i.e., the target device), and the historical backup task execution performance—we obtain the environmental metrics of the source device, the operational metrics of the target device, and the historical backup task execution logs.

[0053] For example, taking the environmental metrics of the source device as an example, detecting the following metrics can obtain network connectivity, permission verification status, resource utilization, backup path anomaly markers, and time consistency:

[0054] For example, available storage, network latency, and IO performance in the operational metrics can be obtained by acquiring the following metrics:

[0055] For example, the success rate, anomaly type statistics, retry count, and results of backup tasks in the historical backup task execution log can be obtained through the metrics in the following table:

[0056] Step 302: Input environmental indicators, operational indicators, and historical backup task execution logs into the status assessment model to obtain the probability distribution of the execution feasibility of the target backup task.

[0057] In one embodiment of this disclosure, execution feasibility includes executable, non-executable, and partially executable. Acquired environmental metrics, operational metrics, and historical backup task execution logs are input into a pre-trained state evaluation model, which outputs a probability distribution of the execution feasibility of the target backup task. The state evaluation model is obtained through the following method: Historical environmental metrics, historical operational metrics, and historical backup task execution logs are used as model training features. Final status labels (executable, non-executable, partially executable, etc.) marked by operations experts or system administrators based on the actual execution results of backup tasks are used as supervision labels. A supervised learning classification model (such as a random forest) is trained using these features and labels to obtain the status evaluation model. Historical environmental metrics, historical operational metrics, and historical backup task execution logs together constitute historical executable check records, which are complete check data automatically recorded by the system during each health check. The final status labels are posterior evaluation markers formed based on the actual execution results of backup tasks.

[0058] For example, before inputting environmental metrics, operational metrics, and historical backup task execution logs into the state assessment model, the model further includes processing the environmental metrics, operational metrics, and historical backup task execution logs, such as: For numerical features, resource occupancy is normalized to the 0-1 range, network latency is discretized by binning, and storage capacity percentage is calculated. For categorical features, path permission status is encoded in binary, and backup types are categorized. Simultaneously, continuous executable check results are extracted to form time-series features to determine whether relevant indicators continuously meet preset thresholds. Furthermore, a rule engine pre-labels known anomalies such as path occupancy and insufficient permissions, and an isolated forest algorithm is used to detect unknown anomaly patterns: a binary tree is constructed based on various input indicators, environmental indicators are identified during the executability check, and for unmarked abnormal parameters, the algorithm calculates the corresponding anomaly score and determines its status, thus completing the anomaly feature labeling. Labeling anomaly features aims to identify abnormal states and potential risks in indicators and use them as features in the inference of the state assessment model, enabling the model to fully learn the correlation between anomaly patterns and backup execution results, improving the accuracy and robustness of feasibility prediction.

[0059] Step 303: Based on the probability distribution of execution feasibility, determine the executability prediction results of the target backup task and its corresponding execution strategy.

[0060] In one embodiment of this disclosure, the executability prediction result of the target backup task and its corresponding execution strategy are determined based on the probability distribution of executability, including: Based on the probability distribution of execution feasibility, the execution feasibility with the highest probability in the probability distribution or the execution feasibility with a probability exceeding a preset probability threshold is determined as the executability prediction result of the target backup task. Based on the executability prediction results, the corresponding execution strategy is determined.

[0061] Specifically, execution feasibility includes executable, non-executable, and partially executable. Based on the probability distribution of execution feasibility output by the model, the execution feasibility with the highest probability or the probability above a preset threshold is taken as the execution feasibility prediction result of the target backup task, and the corresponding execution strategy is matched according to the execution feasibility prediction result.

[0062] For example, the probability value ranges from 0 to 100%, and the sum of the probability of being executable, the probability of being non-executable, and the probability of being partially executable is 100%. Based on the executable probability, non-executable probability, and partially executable probability output by the model, the executableness with the highest probability is selected as the executableness prediction result of the target backup task. Different thresholds (i.e., preset thresholds) can also be set for the executable probability, non-executable probability, and partially executable probability. The execution feasibility that first meets the corresponding threshold is determined as the executableness prediction result of the target backup task, and the corresponding execution strategy is matched according to the result.

[0063] Furthermore, when the executability prediction result is executable, the corresponding execution strategy is to immediately execute the backup jobs for all backup paths and add the target backup task to a high-priority queue for priority transmission. When the executability prediction result is partially executable, the corresponding execution strategy is to skip the abnormal path according to the preset rules and trigger the asynchronous repair task; When the feasibility prediction result is that the task is not executable, the corresponding execution strategy is to terminate the target backup task and start the fault self-healing process.

[0064] Specifically, based on the executability prediction results, corresponding backup strategies are automatically matched and executed. When the executability prediction result is executable, backup jobs for all backup paths of the target task are immediately executed, and the backups are added to a high-priority queue for priority transmission. When the executability prediction result is partially executable, paths with anomalies are skipped according to preset rules, and an asynchronous repair task is triggered in the background. When the executability prediction result is not executable, the current target backup task is terminated directly, and a fault self-healing process is initiated. The preset rules are as follows: Temporary files are automatically skipped without alerts. For non-core logs, skip them and record them, then summarize and report them daily. For the user data directory, skip it and immediately send an email alert; For core database files, terminate backups and trigger manual intervention.

[0065] Furthermore, if the executability prediction result is partially executable, the backup task checker returns a list of skipped abnormal paths based on the executability prediction request.

[0066] Specifically, if the executability prediction result is partially executable, the backup task checker will also return the list of skipped abnormal paths to the management and scheduling center. When the management and scheduling center determines that the final executability result of the target backup task is partially executable, the backup instruction sent to the backup task executor also carries the list of skipped abnormal paths. The backup task executor, based on the final execution strategy corresponding to partial executability, automatically skips the corresponding abnormal paths according to preset rules and the skipped path list during the execution of the target backup task.

[0067] In summary, according to the technical solution provided in the embodiments of this disclosure, this disclosure comprehensively collects multi-dimensional detection indicators from three dimensions: the environment of the source device, the operation of the target device, and historical backup execution logs. These indicators are then input into a state evaluation model trained based on supervised learning, which outputs a probability distribution of the feasibility of backup task execution. Based on the highest probability or a preset threshold, the executability prediction result is determined and a corresponding execution strategy is matched. This enables the accurate identification of potential risks and intelligent determination of the executability status of the task before the backup task is executed, effectively improving the stability and success rate of backup task execution, reducing the backup failure rate caused by environmental anomalies, insufficient resources, or historical risks, and simultaneously achieving adaptive matching of execution strategies, thereby enhancing the reliability and robustness of the backup system.

[0068] Furthermore, during the feature engineering and anomaly labeling of environmental indicators, operational indicators, and historical backup task execution logs, unsupervised clustering models (such as K-means) are used to perform cluster analysis on various indicators and anomaly features. This allows for the autonomous discovery of frequently occurring similar anomaly combinations, such as anomaly scenarios involving network jitter and high CPU usage concurrency. Based on the clustering results, corresponding handling rules are automatically generated to enrich anomaly identification and handling strategies. Simultaneously, when determining the executability prediction results based on the probability distribution of execution feasibility, a dynamic threshold adjustment mechanism is adopted. Based on historical operational data (such as the success rate of backup tasks in the past 24 hours) statistically analyzed using a sliding window, the judgment thresholds for executability, non-executability, and partial executability are updated in real time. This enables the thresholds to adapt to changes in the system environment, thereby further improving the prediction accuracy of the state assessment model and the rationality of the execution strategy matching.

[0069] The aforementioned threshold can be determined using the following formula:

[0070] Furthermore, based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request, the final executability result of the target backup task and its corresponding final execution strategy are determined, including: If the response time for the executability prediction request does not exceed the preset response time, and if multiple consecutive executability prediction results are executable, and the number of predictions reaches the executability determination threshold, the executability prediction result and its corresponding execution strategy will be used as the final executability result and its corresponding final execution strategy.

[0071] Specifically, a preset response time is set, such as 30 seconds for file-based backup tasks and 60 seconds for database-based tasks. If the response time for the executability prediction request does not exceed the preset response time, and if multiple consecutive executability predictions result in executability, and the number of consecutive successes reaches the preset executability judgment threshold, then the environment of the current target device and source device is considered stable and reliable. The executability prediction result and its corresponding execution strategy are then taken as the final executability result and its corresponding final execution strategy.

[0072] Furthermore, based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request, the final executability result of the target backup task and its corresponding final execution strategy are determined, including: If the response time of the executability prediction request does not exceed the preset response time, the executability prediction result is partially executable. The executability prediction result and its corresponding execution strategy are taken as the final executability result and its corresponding final execution strategy of the target backup task.

[0073] Specifically, a preset response time is set, such as 30 seconds for file-based backup tasks and 60 seconds for database-based tasks. If the response time for the executability prediction request does not exceed the preset response time, and the executability prediction result is partially executable, then the partially executable result and its corresponding execution strategy are taken as the final executability result and its corresponding final execution strategy.

[0074] In practical applications, a check interval is set, which is the time interval between two consecutive executability prediction requests. For example, a regular backup task might have a check interval of 5 minutes, while a high-priority backup task might have a check interval of 1 minute. The check interval balances resource consumption and real-time requirements. Furthermore, the check interval can be dynamically adjusted using an exponential smoothing method.

[0075] Furthermore, based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request, the final executability result of the target backup task and its corresponding final execution strategy are determined, including: If the response time of the executability prediction request exceeds the preset response time, or if the executability prediction result is unexecutable, it is recorded as an unexecutable determination. When the cumulative number of unexecutable determinations reaches the preset unexecutable determination threshold, the unexecutability and the corresponding execution strategy are determined as the final executability result and final execution strategy of the target backup task.

[0076] Specifically, a preset response time is set, such as 30 seconds for file-based backup tasks and 60 seconds for database-based tasks. When the management and scheduling center initiates an executability prediction request, each occurrence of either a timeout or a prediction result indicating non-executability is recorded as an unexecutability determination. When the cumulative number of such unexecutability determinations reaches the preset unexecutability determination threshold, the management and scheduling center will determine the final executability result as unexecutable and adopt the corresponding execution strategy.

[0077] Figure 4 This is a flowchart illustrating another data backup method provided in this disclosure. The method is executed by a backup task executor and specifically includes the following steps: Step 401: Receive the backup instruction issued by the management and scheduling center. The backup instruction includes the final executability result of the target backup task and its corresponding final execution strategy. Step 402: In response to the backup command, execute the target backup task according to the final execution strategy; The backup instruction generation process includes: the management and scheduling center sending an executability prediction request for the target backup task to the backup task checker; receiving the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; determining the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request; determining whether the final executability result meets the backup conditions; and if the backup conditions are met, issuing a backup instruction containing the final executability result and the final execution strategy to the backup task executor.

[0078] Specifically, the target backup task refers to the task that currently requires a backup operation. Before issuing the actual backup instruction, the management and scheduling center sends an executability prediction request to the backup task checker for the target backup task to be executed. This allows the backup task checker to pre-determine whether the target backup task meets the conditions for execution. Further, the executability prediction request carries backup-related information such as the backup type, target device path, and source device path of the target backup task. The backup task checker parses the executability prediction request and pre-determines whether the target backup task meets the conditions for execution. The execution strategy refers to the execution rules determined for the target backup task based on the executability prediction results. After receiving the executability prediction request, the backup task checker pre-judges the executability of the target backup task and generates corresponding executability prediction results and execution strategies based on the pre-judge results. These are then returned to the management and scheduling center for reception. By comprehensively weighting and analyzing the executability results obtained from one or more predictions, the corresponding execution strategies, and the response time of each prediction request, a final executability prediction for the target backup task is made. Based on this final prediction, a final executability result and corresponding final execution strategy are generated. The response time can be used to assess the load of the backup task checker and network status, avoiding the use of unreliable prediction results when the checker is busy or abnormal. Backup conditions include whether the final executability result is executable or partially executable. By comparing the final executability result with preset backup execution conditions, it is determined whether the target backup task meets the criteria for allowed execution, thus deciding whether to issue an actual backup execution command. Partially executable means that only a portion of the data is backed up, rather than a complete backup. When the final executability result meets the backup conditions, the management scheduling center issues a backup command to the backup task executor. This command carries the final execution strategy corresponding to the current final executability result, enabling the backup task executor to execute the target backup task according to this strategy. The backup task executor parses the backup instructions to obtain the final executability result of the target backup task and its corresponding final execution strategy, and then executes the target backup task according to the final execution strategy.

[0079] In summary, according to the technical solution provided in the embodiments of this disclosure, this disclosure initiates an executability prediction request for the target backup task to the backup task checker through the management and scheduling center. The final executability result and the final execution strategy are determined by combining the prediction result, the corresponding execution strategy, and the request response time. After determining that the backup conditions are met, a backup instruction is issued to the backup task executor so that the backup task executor executes the target backup task according to the final execution strategy. This achieves accurate prediction and strategy optimization before the backup task is executed, effectively improving the reliability and success rate of the backup task execution, avoiding the waste of system resources and redundant reading and writing of large files caused by backup execution failure, significantly reducing time delay and additional storage resource consumption, thereby ensuring stable and reliable data backup, enhancing the disaster prevention and mitigation capabilities of the business system, and minimizing the risk of unplanned downtime.

[0080] Furthermore, data backup methods also include: If a process blocking problem occurs during the execution of the target backup task, the target backup task will be executed based on a retry and wait strategy. The retry and waiting strategy includes setting a maximum number of retries, a retry interval, and a maximum number of waits. Executing the target backup task based on the retry and waiting strategy includes: Waiting according to the retry interval, initiating retry to execute the target backup task, each retry is counted as one wait; if the target backup task still fails to resume normal execution within the maximum number of retry attempts, or when the maximum number of wait attempts is reached, the execution of the target backup task is terminated and an anomaly is reported.

[0081] Specifically, during the execution of the target backup task, a retry and waiting strategy is designed to ensure the robustness of the job in response to potential process blocking issues (such as insufficient target resources, busy I / O, network congestion, or lock contention). This strategy achieves controllable handling of blocking scenarios by configuring three core parameters: maximum number of retries, retry interval, and maximum number of waits. When a process is blocked, the executor will enter a waiting state according to the preset retry interval. After the wait is completed, it will retry the execution of the target backup task, and each retry is counted as one wait. If the task still has not resumed normal execution after reaching the maximum number of retries, or if the cumulative number of waits has reached the maximum wait threshold, the execution of the current target backup task will be terminated immediately, and a detailed exception log will be reported to the management and scheduling center. This prevents the task from blocking indefinitely and consuming system resources, while ensuring that maintenance personnel can intervene in a timely manner.

[0082] Furthermore, the backup task executor includes a source device and a target device; Execute the target backup task according to the final execution strategy, including: The source device divides the data to be backed up in the target backup task into high-priority data and low-priority data according to the importance of the data. The high-priority data is transmitted synchronously in real time, and the low-priority data is transmitted when the bandwidth is idle.

[0083] Specifically, the backup task executor consists of a source device and a target device. The source device is responsible for data reading and scheduling transmission, while the target device is responsible for receiving and storing backup data. When executing the target backup task according to the final execution strategy, the source device first divides the data to be backed up into high-priority and low-priority data based on the business importance of the data: for high-priority core business data, a real-time synchronous transmission strategy is adopted to ensure that critical data can be backed up to the target device in a timely manner; for low-priority non-critical data (such as logs, cache files, etc.), a bandwidth idle time transmission strategy is adopted to complete data transmission during periods of abundant network resources or low business hours. This ensures the timeliness of core data backup while avoiding backup traffic from consuming the main business bandwidth, achieving coordinated stability between the backup task and business operations.

[0084] Furthermore, the method also includes: After the target backup task is completed, the execution results and exception information are collected and fed back to the management and scheduling center.

[0085] Specifically, execution results include success, failure, in progress, and partial success. Anomaly information includes: permission verification anomalies, backup path anomalies, file usage anomalies, insufficient resource anomalies, IO performance anomalies, data verification anomalies, process blocking anomalies, task timeout anomalies, time synchronization anomalies, and related information such as system load, resource utilization, task execution status, and historical anomaly records at the time of the anomaly. After the target sub-task is completed, the backup task executor collects the execution results and anomaly information and feeds them back to the management and scheduling center. Simultaneously, feature extraction is performed on the execution results and anomaly information, and anomaly patterns are identified through cluster analysis, classifying anomalies into known categories and new categories. Existing handling rules are optimized for known categories of anomalies, and new handling rules are generated for new categories of anomalies. The updated rules are stored in the policy knowledge base for continuous optimization of subsequent executability prediction and anomaly handling capabilities through closed-loop learning.

[0086] To explain in detail the data backup method provided in this disclosure, Figure 4 This is a flowchart illustrating another data backup method provided in an embodiment of the present disclosure.

[0087] like Figure 5 As shown, the method specifically includes the following steps: Step 501: The management and scheduling center sends an executability prediction request for the backup task to the backup task checker.

[0088] Specifically, before initiating a backup job, the management scheduling center sends a backup task executability prediction request to the backup task checker.

[0089] Step 502: The backup task checker predicts the executability of the backup task based on the received executability prediction request, and generates executability prediction results and their corresponding execution strategies.

[0090] Specifically, after receiving an executability prediction request, the backup task checker performs an executability prediction on the backup task, obtains the executability prediction results and their corresponding execution strategies, including: Acquire environmental metrics from the source device, operational metrics from the target device, and historical backup task execution logs. Preprocess this data, including feature extraction and anomaly labeling. Input the preprocessed data into the Zhangtai evaluation model, which outputs a probability distribution of the backup task's execution feasibility (including executable, inexecutable, and partially executable). Select the execution feasibility with the highest probability in the probability distribution, or the execution feasibility with a probability exceeding a preset probability threshold, as the predicted executability result for the backup task.

[0091] The specific methods for feature extraction and anomaly feature labeling are as follows: Numerical characteristics: resource utilization (normalized to 0-1), network latency (binning), and storage capacity percentage.

[0092] Category-based characteristics: path permission status (binary encoding), backup type (file / database / virtualization).

[0093] Temporal characteristics: Time series of continuously executable check results (used to determine whether the threshold is continuously met).

[0094] Anomaly feature markers: Anomalies (such as paths being occupied or insufficient permissions) are pre-marked based on the rule engine.

[0095] The Isolation Forest algorithm is used to detect unforeseen anomalous patterns: a binary tree is constructed from the input data metrics, and during runtime inspection, the source environment is identified. For unlabeled anomalous parameters, the algorithm calculates anomaly scores and defines the state of the parameter.

[0096] Step 503: The management and scheduling center receives the executability prediction results and corresponding execution strategies returned by the backup task checker based on the executability prediction request, and finally judges the final executability result of the backup task.

[0097] Specifically, the backup task checker returns the executability prediction results from the executability prediction request to the management and scheduling center. If the response time to the executability prediction request does not exceed a preset response time, and multiple consecutive executability predictions are all executable, and the number of predictions reaches the executability determination threshold, the management and scheduling center will use the executability prediction results and their corresponding execution strategies as the final executability result and its corresponding final execution strategy. If the response time to the executability prediction request does not exceed the preset response time, and the executability prediction result is partially executable, the executability prediction result and its corresponding execution strategy will be used as the final executability result and its corresponding final execution strategy for the target backup task. If the response time to the executability prediction request exceeds the preset response time, or if the executability prediction result is unexecutable, it will be recorded as an unexecutable determination. When the cumulative number of unexecutable determinations reaches the preset unexecutable determination threshold, the unexecutability and its corresponding execution strategy will be determined as the final executability result and final execution strategy for the target backup task.

[0098] When the executability prediction result is executable, the corresponding execution strategy is to immediately execute the backup jobs of all backup paths and add the target backup task to a high-priority queue for priority transmission. When the executability prediction result is partially executable, the corresponding execution strategy is to skip the abnormal path according to the preset rules and trigger the asynchronous repair task; When the feasibility prediction result is that the task is not executable, the corresponding execution strategy is to terminate the target backup task and start the fault self-healing process.

[0099] The default rules are: Temporary files are automatically skipped without alerts. For non-core logs, skip them and record them, then summarize and report them daily. For the user data directory, skip it and immediately send an email alert; For core database files, terminate backups and trigger manual intervention.

[0100] Furthermore, the determination of the executability prediction results and their corresponding execution strategies obtained by the above model, as well as the final executability results and their corresponding final execution strategies for the backup tasks, can be found in [reference needed]. Figure 6 , Figure 6 This is a flowchart illustrating the executability prediction method provided in embodiments of this disclosure. Figure 6 As shown, the specific steps include the following: 6.1 Trigger check.

[0101] Specifically, when the backup task checker receives an executability prediction request, which carries information such as the backup type, target device path, and source device path of the backup task, it triggers the executability prediction method for the backup task.

[0102] 6.2 Parameter loading.

[0103] Specifically, after receiving an executability prediction request, the backup task checker parses the request to obtain the backup type and loads the core parameters of the prediction mechanism based on the backup type. The parameter loading process includes global default parameters, backup type templates, and task coverage parameters. Among these, the core parameters of the prediction mechanism include: Backup types, such as file backup, database backup, virtualization environment backup, etc.; Check the response timeout period; if the check fails after this time in the source environment, the check will be reported as failed. The inspection interval is the time between a successful inspection and a subsequent inspection. The executability threshold, i.e., how many consecutive successful backups are required before the device to be backed up to the target device can be defined as executable, can be adjusted according to business needs to determine whether to set it and the parameter value set. The threshold for determining an unexecutable state is the number of consecutive failures required before the device to be backed up to the target device can be defined as unexecutable. Whether to set this threshold and the parameter value can be adjusted according to business needs. Intelligent fast execution means that for scenarios that affect the backup of a small amount of data on the source side, such as when the directory or path is occupied or some read permissions are restricted, the backup can be skipped intelligently without affecting the overall execution progress. The retry and waiting strategy involves setting parameters such as the number of retries, waiting interval, and waiting count for the source to check the blocked path. Based on this strategy, the location and range of the blocked path can be located, and the recommended execution strategy for the path can be returned to the scheduling center.

[0104]

[0105] 6.3 State detection.

[0106] Specifically, obtain the environmental indicators of the source device and the operational indicators of the target device.

[0107] The environmental metrics include: network connectivity, permission verification status, resource utilization, backup path anomaly flags, and time consistency; the operational metrics include: available storage, network latency, and IO performance; historical backup task execution logs include: backup task success rate, anomaly type statistics, and retries and results. These environmental and operational metrics can be obtained through the following detection items:

[0108] 6.4 Result determination.

[0109] Specifically, if the executable threshold is reached consecutively (e.g., 3 successful attempts), the backup task is marked as "executable".

[0110] If the inexecutable threshold is reached consecutively (e.g., 2 failures), the backup task is marked as "inexecutable".

[0111] If a partial path is detected as abnormal but meets the "fast execution" condition (e.g., the directory is occupied but can be skipped), return "conditionally executable" (i.e., the aforementioned part is executable).

[0112] 6.5 Strategy Recommendations.

[0113] Specifically, if the flag is executable, return a full execution suggestion. If the flag is not executable, return a termination suggestion. If the flag is conditionally executable, return the skipped path and the number of retries.

[0114] 6.6 Task Strategy.

[0115] Specifically, the management and dispatch center issues backup instructions based on policy recommendations.

[0116]

[0117] Step 504: After determining that the final executable prediction result meets the backup conditions, the management and scheduling center issues a backup instruction to the backup task executor so that the backup task executor executes the backup task according to the final execution strategy.

[0118] The aforementioned state assessment model can be optimized in the following ways: Specifically, the state assessment model is used to output the probability distribution of backup tasks as executable, partially executable, and inexecutable based on source-end environmental indicators, target-end operational indicators, and historical backup logs, thereby determining the executability prediction results of backup tasks. The objective function, based on the output of the state assessment model, is used to dynamically adjust the weight coefficients of success rate, backup rate, and resource consumption under the executable state determined by the state assessment model, according to real-time business load, network bandwidth, and resource consumption. Under the premise of ensuring the normal operation of the main business, through continuous iterative optimization, the backup strategy achieves the optimal balance of success rate, efficiency, and resource consumption in different business environments. Therefore, the objective function needs to be continuously and dynamically adjusted based on the results of the state assessment model and the business operation status. Through continuous learning and feedback loops, the status assessment model and backup handling strategy are further iterated and optimized. After each backup task is executed, the execution results and status data are used as new samples to update the model. The model parameters and handling rules are continuously optimized through anomaly rule mining, rule validity evaluation, and dynamic updates of the knowledge base. The overall model adopts a composite model architecture that combines supervised and unsupervised learning. It has dynamic parameter adaptation, strong interpretability, and closed-loop self-optimization capabilities, which can continuously expand the coverage of anomaly scenarios and continuously improve the execution reliability and system robustness of backup tasks.

[0119] The formula for the objective function is shown below: Optimization target = α × success rate + β × transmission rate - γ × (CPU usage + memory usage).

[0120] Dynamic parameter adjustment: The weights of α, β, and γ are adjusted based on the real-time resource utilization rate.

[0121] In the continuous learning and feedback loop, for new abnormal events, features are first extracted to transform the abnormal events into computable feature vectors. Then, through cluster analysis, the clustered anomalies are divided into new categories or known categories. If it is a new category, new handling rules are generated; if it is a known category, existing handling rules are optimized. Finally, the newly generated or optimized rules are stored in the strategy knowledge base to realize the continuous iteration and expansion of the anomaly handling strategy.

[0122] Step 505: The backup task executor executes the backup task.

[0123] Specifically, after the management and scheduling center issues a backup command to the backup task executor, the backup task executor starts the backup process. During the backup process, there are two parallel sub-processes: one for the source device and one for the target device.

[0124] in: The source device preprocesses the data to be backed up, including compression, encryption, or chunking.

[0125] The processed data is synchronized to the transmission queue in real time. The data is then transmitted over the network to the storage medium (such as cloud storage, local disk, etc.) of the target device.

[0126] The transmission progress and status of the source and target devices are fed back to the management and scheduling center in real time.

[0127] After the backup task executor completes the backup, it returns the final execution result (success / partial success details / failure details) to the management and scheduling center. Backup data is persistently stored on the target device's storage medium, supporting object storage, block storage, or local storage.

[0128] Specifically, the backup task executor executes intelligent backup parameters according to backup instructions and, through the continuous learning capabilities of the intelligent agent, quickly makes decisions and executes in cases of abnormal source device environments: for scenarios that have been learned and trained, it can automatically locate the abnormal path and scope at the source end and execute skip or wait commands according to plan; in the face of sudden service connection anomalies, including network traffic anomalies, it executes according to existing training plans. The detailed workflow between the management and scheduling center and the backup task executor is as follows: 1. Job scheduling and executability check: Input: Backup task metadata (source path, target storage type, SLA requirements).

[0129] Executable determination rules:

[0130]

[0131] 2. Start the backup task executor: Parameter loading order: ① Backup type adapter (matches backup tools and strategies based on data type) parameter configuration reference:

[0132] ② Intelligent and fast execution: For scenarios where partial execution is possible, such as when directories or paths are occupied or partial read permissions are restricted, which may affect the backup of a small amount of data on the source end, the system can intelligently wait, retry, or skip during backup job execution without affecting the overall job progress. The execution strategy is dynamically adjusted based on the executable check results to achieve "partially successful" backup. The process is as follows: Upon receiving a backup instruction, if the target backup task is partially executable, a preset skip rule is loaded and a dynamic exclusion list is generated. The backup is executed based on the exclusion list, and the skipped items are recorded. Finally, a partial success log is reported. If the task is fully executable, the backup is executed normally and completely, and a success log is reported. This ensures that the core data backup is completed while enabling flexible handling of abnormal scenarios and traceability of execution results.

[0133] ③ Retry and wait strategy, which sets parameters such as the number of retries, waiting interval, and waiting count when the backup task encounters reading problems or other issues that cause process blocking during execution, and executes automatically within the identified path range; Retry and wait strategy engine (automatically retry or wait when an exception occurs to avoid task interruption).

[0134] The parameter configuration table is as follows:

[0135] ④ Skip strategy, which means setting the path or file to skip when the executor encounters process blocking during execution, and automatically collecting the skipped range, logs, etc., and returning them to the management and scheduling center.

[0136] Multi-level skip rule (safely skipping non-critical paths when retry is not possible, ensuring task progress):

[0137] Parameter configuration table:

[0138] The exception handling of the intelligent backup task executor includes path-level exception handling and network-level exception handling.

[0139] Path-level exception handling is as follows:

[0140] Network-level anomaly handling includes adaptive degradation transmission and breakpoint resume protection mechanisms as follows: Adaptive degradation transmission:

[0141] Resumable downloads while preserving the transmission state machine:

[0142] 3. Source-end device processing sub-process: Data preprocessing:

[0143] The real-time synchronization mechanism employs a dual-queue design: Specifically, based on the importance of the data service, the data blocks to be transmitted are diverted to different priority queues: urgent data enters the high-priority queue and is transmitted immediately to ensure the timeliness of core data backup; data that can be delayed enters the low-priority queue and is transmitted when bandwidth is idle to complete data synchronization during network off-peak hours, avoiding the competition for main service bandwidth resources, thereby achieving hierarchical scheduling and resource optimization of backup data transmission.

[0144] Priority criteria:

[0145] 4. Intelligent transmission strategy at the target end:

[0146] 5. Learning feedback system operation mechanism: After the backup task is completed, feature extraction and anomaly pattern recognition are performed based on the execution results: if a new type of anomaly is identified, a new handling rule is generated; if a known type of anomaly is identified, the existing handling rule is optimized; then the newly generated or optimized rule is updated to the policy knowledge base to guide the executability prediction and policy selection of subsequent backup tasks, thereby realizing a complete feedback loop from execution results to policy optimization, and continuously improving the system's anomaly detection capability and handling efficiency.

[0147] The task executor receives backup execution instructions from the management and scheduling center, obtains the final executability result of the target backup task and the corresponding final execution strategy; during the execution of the backup job, if an anomaly is identified, it executes the corresponding rules according to the handling plan in the executability prediction report; at the same time, the task executor has automatic learning capabilities, and can continuously optimize existing rules or generate new rules for new anomaly types based on the execution results, constantly enriching its own strategy knowledge base; after the job is completed, it reports logs and reports containing anomaly handling details to the management and scheduling center, thereby stably handling various anomaly scenarios during job execution, improving job execution efficiency, and saving time, network traffic, and storage costs caused by job failures.

[0148] The above-mentioned feasibility prediction results of backup tasks and their corresponding execution strategies obtained through the state assessment model can achieve the following technical effects: Precise Status Awareness: Based on multi-dimensional indicators such as source environment, target operation and historical backup logs, the feasibility of backup tasks is quantitatively evaluated, enabling early detection of risks in backup task execution.

[0149] Elastic strategy generation: Based on real-time business load, network bandwidth and resource usage, the backup execution strategy is dynamically adapted to ensure the stable operation of the main business while optimizing backup efficiency.

[0150] Closed-loop learning optimization: Relying on incremental learning and anomaly rule mining mechanisms, the state assessment model and handling rules are continuously iterated and optimized to improve the accuracy of anomaly detection and the efficiency of fault handling.

[0151] Before a backup task is executed, the management and scheduling center sends an executability prediction request to the backup task checker. Combining the prediction results, corresponding execution strategies, and request response time, the final executability result and final execution strategy are comprehensively determined. After determining that the backup conditions are met, a backup command is issued to the backup task executor, enabling the executor to execute the target backup task according to the final execution strategy. This achieves accurate prediction and pre-optimization of strategies before backup task execution, significantly improving the reliability of large-scale model backup tasks. Preventative interception: Identify and intercept abnormal environments at the source / target end in advance to avoid backup data corruption or task failure due to environmental issues.

[0152] Partial disaster recovery capability: By employing tiered skipping and asynchronous repair strategies in "partially executable" scenarios, the impact of non-critical path anomalies on the overall backup task is reduced, ensuring the completion of core data backup.

[0153] Dynamic adaptive optimization: Based on historical backup execution data, continuously iterate and improve the inspection rules and model parameters to reduce the cost of manual intervention and improve the system's adaptability and robustness.

[0154] Existing solutions are all aimed at continuously optimizing products to address known and new problems, reducing or avoiding the possibility of the same problems recurring. These are all post-incident remedial measures and cannot cope with all unexpected problems. When the problem optimization capabilities are not available or when new problems are encountered, backup operations may still fail, resulting in a waste of resources such as time, network traffic, storage resources, and service process consumption.

[0155] The data backup method disclosed herein has the following advantages compared to existing solutions: 1. This disclosure provides two core intelligent methods: an intelligent prediction method for executability and an intelligent backup task execution method. These methods can pre-check the source environment and target storage media to determine the executability of the backup task. For unusable, non-executable check results, the task can be terminated early to avoid blocking during execution. For partially non-executable and partially executable check results, strategies such as waiting, skipping, and retrying can be automatically executed based on intelligent prediction results and suggested solutions, ensuring the task continues even when partial blocking occurs during execution. Furthermore, both core intelligent methods possess continuous learning capabilities, continuously enriching the knowledge base, optimizing component algorithm logic, continuously optimizing existing solutions, and identifying and generating new solutions based on diverse application scenarios, thereby continuously optimizing the handling of new and old problems encountered during production operations.

[0156] 2. This disclosure uses an executable inspection mechanism to scan and inspect the environmental indicators of the source device to be backed up and the operating indicators of the target device. Compared with solutions where the known environment detection results are only pass or fail, the inspection solution of this disclosure incorporates executable inspection-related parameter configuration and dynamic algorithms. In addition to detecting environmental unavailability in advance, it also has the ability to automatically identify and judge environmental availability without human intervention, which belongs to artificial intelligence-driven operation and maintenance automation.

[0157] 3. This disclosure forms a data moat through a closed-loop learning system, making it difficult for competitors to replicate the dynamic strategy optimization capabilities. It prioritizes tasks with high priority in checking and feedback, improves backup job efficiency, and reduces resource overload and process occupation issues during service operation.

[0158] This disclosure also provides a data backup device. Figure 7 A structural block diagram of a data backup device provided in an embodiment of this disclosure, such as... Figure 7 As shown, the data backup device 700 includes: a sending unit 701, a receiving unit 702, a determining unit 703, a judging unit 704, and a sending unit 705.

[0159] In one exemplary embodiment, the sending unit 701 is configured to send an executability prediction request for the target backup task to the backup task checker. In one exemplary embodiment, the receiving unit 702 is configured to receive the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request. In one exemplary embodiment, the determining unit 703 is configured to determine the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request. In one exemplary embodiment, the determination unit 704 is used to determine whether the final executability result meets the backup conditions; In one exemplary embodiment, the issuing unit 705 is configured to issue a backup instruction to the backup task executor in response to the fulfillment of backup conditions. The backup instruction includes the final execution strategy corresponding to the final executability result of the backup conditions being met.

[0160] This disclosure also provides a data backup device. Figure 8 A structural block diagram of another data backup device provided in this disclosure embodiment is shown below. Figure 8 As shown, the data backup device 800 includes a receiving unit 801 and an execution unit 802.

[0161] In one exemplary embodiment, the receiving unit 801 is configured to receive a backup instruction issued by the management and scheduling center. The backup instruction includes the final executability result of the target backup task and its corresponding final execution strategy. In one exemplary embodiment, the execution unit 802 is configured to execute a target backup task in response to a backup instruction, according to a final execution strategy. The backup instruction generation process includes: the management and scheduling center sending an executability prediction request for the target backup task to the backup task checker; receiving the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; determining the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request; determining whether the final executability result meets the backup conditions; and if the backup conditions are met, issuing a backup instruction containing the final executability result and the final execution strategy to the backup task executor.

[0162] Figure 9 This is a hardware block diagram of an electronic device provided according to an embodiment of the present disclosure. The electronic device 900 according to an embodiment of the present disclosure includes at least a processor and a memory for storing computer-readable instructions. When the computer-readable instructions are loaded and executed by the processor, the processor performs the data backup method described in any of the preceding embodiments of the present disclosure.

[0163] Figure 9 The illustrated electronic device 900 specifically includes a central processing unit (CPU) 901, a graphics processing unit (GPU) 902, and a memory 903. These units are interconnected via a bus 904. The CPU 901 and / or GPU 902 can function as the aforementioned processor, and the memory 903 can function as the aforementioned memory storing computer-readable instructions. Furthermore, the electronic device 900 may also include a communication unit 905, a storage unit 906, an output unit 907, an input unit 908, and an external device 909, all of which are also connected to the bus 904.

[0164] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0165] The data backup method disclosed herein has the ability to automatically wait, retry, and automatically skip blocked paths during task execution. It can predict and anticipate problems in advance, and can also respond to unexpected scenarios that may affect task failure by waiting and retrying based on the predictive analysis of executability during the backup task execution process, thereby ensuring the success rate of the backup task. It enables the backup task to check in advance whether the disaster recovery machine to be backed up has the ability to execute the task, intelligently generate solutions to hidden problems based on the check results, and automatically output relevant execution parameter scripts to the backup task executor to deal with unexpected situations during the execution process.

[0166] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0167] The block diagrams of devices, apparatuses, devices, and systems disclosed herein are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0168] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0169] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0170] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0171] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0172] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A data backup method, characterized in that, Applied to a management and scheduling center, the method includes: Send an executability prediction request for the target backup task to the backup task inspector; Receive the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; Based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request, the final executability result of the target backup task and its corresponding final execution strategy are determined. Determine whether the final executability result meets the backup conditions; In response to the fulfillment of the backup conditions, a backup instruction is issued to the backup task executor so that the backup task executor executes the target backup task in accordance with the final execution strategy.

2. The data backup method according to claim 1, characterized in that, The executability prediction results and their corresponding execution strategies are generated based on the following methods: Obtain environmental metrics of the source device, operational metrics of the target device, and historical backup task execution logs; The environmental indicators, the operational indicators, and the historical backup task execution logs are input into the status evaluation model to obtain the probability distribution of the execution feasibility of the target backup task; Based on the probability distribution of the execution feasibility, the executability prediction result of the target backup task and its corresponding execution strategy are determined.

3. The data backup method according to claim 2, characterized in that, The environmental metrics include: network connectivity, permission verification status, resource utilization, backup path anomaly markers, and time consistency; the operational metrics include: available storage capacity, network latency, and IO performance; the historical backup task execution logs include: backup task success rate, anomaly type statistics, retry count and results.

4. The data backup method according to claim 2, characterized in that, The feasibility of execution includes executable, non-executable, and partially executable. The step of determining the executability prediction result of the target backup task and its corresponding execution strategy based on the probability distribution of the executability includes: Based on the probability distribution of the execution feasibility, the execution feasibility with the highest probability in the probability distribution or the execution feasibility with a probability exceeding a preset probability threshold is determined as the executability prediction result of the target backup task. Based on the executability prediction results, the corresponding execution strategy is determined.

5. The data backup method according to claim 4, characterized in that, The method further includes: When the executability prediction result is executable, the corresponding execution strategy is to immediately execute the backup jobs for all backup paths and add the target backup task to a high-priority queue for priority transmission. When the executability prediction result is partially executable, the corresponding execution strategy is to skip the abnormal path according to the preset rules and trigger the asynchronous repair task; When the feasibility prediction result is that the task is not executable, the corresponding execution strategy is to terminate the target backup task and start the fault self-healing process.

6. The data backup method according to claim 5, characterized in that, The preset rule is as follows: Temporary files are automatically skipped without alerts. For non-core logs, skip them and record them, then summarize and report them daily. For the user data directory, skip it and immediately send an email alert; For core database files, terminate backups and trigger manual intervention.

7. The data backup method according to claim 4, characterized in that, The step of determining the final executability result and corresponding final execution strategy of the target backup task based on at least one executability prediction result and its corresponding execution strategy, and the response time of the executability prediction request, includes: If the response time to the executability prediction request does not exceed a preset response time, and if multiple consecutive executability prediction results are executable, and the number of predictions reaches an executability determination threshold, then the executability prediction result and its corresponding execution strategy are taken as the final executability result and its corresponding final execution strategy; and / or, If the response time to the executability prediction request does not exceed a preset response time, the executability prediction result is partially executable, and the executability prediction result and its corresponding execution strategy are taken as the final executability result and its corresponding final execution strategy for the target backup task; and / or, If the response time of the executability prediction request exceeds the preset response time, or if the executability prediction result is unexecutable, it is recorded as an unexecutable determination. When the cumulative number of unexecutable determinations reaches the preset unexecutable determination threshold, the unexecutability and the corresponding execution strategy are determined as the final executability result and final execution strategy of the target backup task.

8. The data backup method according to claim 5, characterized in that, The method further includes: If the executability prediction result is partially executable, the backup task checker returns a list of skipped abnormal paths based on the executability prediction request.

9. A data backup device, characterized in that, Applications include management and dispatch centers, including: The sending unit is used to send an executability prediction request for the target backup task to the backup task checker; The receiving unit is configured to receive the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request. The determining unit is configured to determine the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, as well as the response time of the executability prediction request. A judgment unit is used to determine whether the final executability result meets the backup conditions; The issuing unit is used to issue a backup instruction to the backup task executor in response to the fulfillment of the backup conditions. The backup instruction includes the final execution strategy corresponding to the final executability result that meets the backup conditions.

10. A data backup method, characterized in that, Applied to backup task executors, including: Receive backup instructions issued by the management and scheduling center, the backup instructions including the final executability result of the target backup task and its corresponding final execution strategy; In response to the backup instruction, the target backup task is executed according to the final execution strategy; The backup instruction generation process includes: the management scheduling center sending an executability prediction request for the target backup task to the backup task checker; receiving the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; determining the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, and the response time of the executability prediction request; determining whether the final executability result meets the backup conditions; and if the backup conditions are met, issuing a backup instruction containing the final executability result and the final execution strategy to the backup task executor.

11. The data backup method according to claim 10, characterized in that, The method further includes: If a process blocking problem occurs during the execution of the target backup task, the target backup task will be executed based on a retry and wait strategy. The retry and waiting strategy includes setting a maximum number of retries, a retry interval, and a maximum number of waits. Executing the target backup task based on the retry and waiting strategy includes: Waiting according to the retry interval, initiating retry to execute the target backup task, each retry is counted as one wait; if the target backup task still fails to resume normal execution within the maximum number of retry attempts, or when the maximum number of wait attempts is reached, the execution of the target backup task is terminated and an anomaly is reported.

12. The data backup method according to claim 10, characterized in that, The backup task executor includes a source device and a target device; Executing the target backup task according to the final execution strategy includes: The source device divides the data to be backed up in the target backup task into high-priority data and low-priority data according to the importance of the data. The high-priority data is transmitted synchronously in real time, and the low-priority data is transmitted when the bandwidth is idle.

13. The data backup method according to claim 10, characterized in that, The method further includes: After the target backup task is completed, the execution results and exception information are collected and fed back to the management and scheduling center.

14. A data backup device, characterized in that, Applied to backup task executors, including: The receiving unit is used to receive backup instructions issued by the management and scheduling center. The backup instructions include the final executability result of the target backup task and its corresponding final execution strategy. An execution unit is configured to execute the target backup task in response to the backup instruction, according to the final execution strategy; The backup instruction generation process includes: the management scheduling center sending an executability prediction request for the target backup task to the backup task checker; receiving the executability prediction result of the target backup task and its corresponding execution strategy returned by the backup task checker based on the executability prediction request; determining the final executability result of the target backup task and its corresponding final execution strategy based on at least one executability prediction result and its corresponding execution strategy, and the response time of the executability prediction request; determining whether the final executability result meets the backup conditions; and if the backup conditions are met, issuing a backup instruction containing the final executability result and the final execution strategy to the backup task executor.

15. An electronic device, characterized in that, include: Memory, used to store computer-readable instructions; as well as A processor for executing the computer-readable instructions, causing the electronic device to perform the data backup method as described in any one of claims 1 to 8 or 10-13.

Images