Server operation and maintenance method, electronic equipment, storage medium and product

Abstract

The present application discloses a server operation and maintenance method, electronic device, storage medium and product, which relate to the field of computer technology, including parsing a structured instruction file corresponding to an operation and maintenance task of a target server submitted by a user; verifying and executing the atomic action based on the triggering conditions, input sequence and expected feedback of the atomic action in the structured instruction file, combined with the screen image of the target server; and determining that the operation and maintenance task of the target server has been successfully executed when the atomic action is completed. The method realizes the operation and maintenance task by parsing the structured instruction file generated based on the user's operation behavior, decomposing the operation and maintenance task into atomic actions containing triggering conditions, input sequence and expected feedback, and realizing closed-loop automated operation from trigger verification to input execution to result verification in combination with the screen image captured by the hardware interface. The method can accurately execute the operation and maintenance task in an out-of-band management scenario without the operating system, thereby improving the operation and maintenance efficiency.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to server operation and maintenance methods, electronic equipment, storage media, and products. Background Art

[0002] As the demand for intelligent operations and maintenance (O&M) in modern data centers continues to grow, remote server management faces numerous challenges. The O&M of large-scale server clusters is becoming increasingly common, creating an urgent need for efficient, consistent, and reliable O&M methods.

[0003] Currently, server O&M solutions primarily include the keyboard, video, and mouse (KVM) functionality of the baseboard management controller (BMC) and automated tools that rely on the operating system. However, KVM only supports manual operation on a single node and lacks automated scripting capabilities, resulting in low O&M efficiency and difficulty in mass deployment. Automated tools also rely on the system environment and cannot deploy tasks if the system crashes or is not booted, hindering the handling of underlying hardware failures. Summary of the Invention

[0004] This application provides a server operation and maintenance method, electronic equipment, storage medium and product to at least solve the problem in related technologies that KVM functions lack automation and automation tools rely on the system environment.

[0005] This application provides a server operation and maintenance method, including:

[0006] Based on the operation and maintenance tasks of the target server submitted by the user, the structured instruction file corresponding to the operation and maintenance tasks is parsed. The structured instruction file is pre-generated based on the user's operation behavior and includes atomic actions. The atomic actions include trigger conditions, input sequences, and expected feedback.

[0007] Based on the trigger conditions, input sequence and expected feedback of the atomic action in the structured instruction file, the atomic action is verified and executed in combination with the screen image of the target server. The screen image is captured from the target server through the hardware interface;

[0008] When the atomic action is completed, it is determined that the operation and maintenance task of the target server is successfully executed.

[0009] This application also provides a server operation and maintenance device, including:

[0010] A parsing unit is used to parse the structured instruction file corresponding to the operation and maintenance task of the target server submitted by the user. The structured instruction file is pre-generated based on the user's operation behavior and includes atomic actions. The atomic actions include trigger conditions, input sequences, and expected feedback;

[0011] An operation and maintenance unit, configured to verify and execute the atomic action based on the triggering conditions, input sequence, and expected feedback of the atomic action in the structured instruction file, combined with a screen image of the target server, where the screen image is captured from the target server via a hardware interface;

[0012] The determination unit is used to determine whether the operation and maintenance task of the target server is successfully executed when the atomic action is completed.

[0013] The present application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for implementing the steps of any of the above-mentioned server operation and maintenance methods when executing the computer program.

[0014] The present application also provides a computer-readable storage medium, in which a computer program is stored, wherein when the computer program is executed by a processor, the steps of any of the above-mentioned server operation and maintenance methods are implemented.

[0015] The present application also provides a computer program product, including a computer program, which implements the steps of any of the above-mentioned server operation and maintenance methods when executed by a processor.

[0016] Through this application, based on the operation and maintenance tasks of the target server submitted by the user, the structured instruction file corresponding to the operation and maintenance tasks is parsed. The structured instruction file is pre-generated based on the user's operation behavior. The structured instruction file includes atomic actions, and the atomic actions include trigger conditions, input sequences, and expected feedback. Based on the trigger conditions, input sequences, and expected feedback of the atomic actions in the structured instruction file, combined with the screen image of the target server, the atomic actions are verified and executed. The screen image is captured from the target server through the hardware interface. When the atomic action is executed, it is determined that the operation and maintenance tasks of the target server have been successfully executed. By parsing the structured instruction file generated based on the user's operation behavior, the operation and maintenance tasks are decomposed into atomic actions containing trigger conditions, input sequences, and expected feedback. The closed-loop automated operation from trigger verification to input execution to result verification is realized by combining the screen image captured by the hardware interface. The operation and maintenance tasks can be accurately executed in out-of-band management scenarios that are separated from the operating system (such as when the system crashes), thereby improving the operation and maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0018] Figure 1 A flow chart of a server operation and maintenance method provided in an embodiment of the present application;

[0019] Figure 2 A flowchart of another server operation and maintenance method provided in an embodiment of the present application;

[0020] Figure 3 A schematic diagram of a process for generating a structured instruction file during a specific server operation and maintenance process provided in an embodiment of the present application;

[0021] Figure 4 This is an architectural diagram of a BMC-side instruction generation system provided in an embodiment of the present application;

[0022] Figure 5 A schematic diagram of batch server cluster operation and maintenance provided in an embodiment of the present application;

[0023] Figure 6 A schematic diagram of the structure of a server operation and maintenance device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0025] It should be noted that, in the description of this application, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. The terms "first," "second," etc., in this application are used to distinguish similar objects, and are not used to describe a particular order or sequence.

[0026] As the demand for intelligent operations and maintenance (O&M) in modern data centers continues to grow, remote server management faces numerous challenges. The increasing prevalence of large-scale server cluster O&M creates an urgent need for efficient, consistent, and reliable O&M methods. However, traditional O&M models and related technologies are significantly inadequate, making them difficult to adapt to this trend.

[0027] At present, server operation and maintenance solutions based on related technologies mainly include keyboard, video and mouse (KVM) functions based on baseboard management controller (BMC) and automation tools that rely on operating system operation.

[0028] Among them, the BMC-based KVM function means that the baseboard management controller (BMC) has the KVM function, which can support real-time visual manual operation of a single node, allowing administrators to intuitively operate and manage a single server node.

[0029] Automation tools (such as Keyshot) are tools that use these tools to implement certain automated operations at the operating system level, aiming to improve the efficiency of some operation and maintenance tasks.

[0030] As can be seen, the KVM function only supports manual operation of a single node and seriously lacks the ability to execute automated scripts. When faced with complex O&M tasks for multiple servers, such as system configuration modifications and BIOS parameter adjustments, manual operations are required on each server, resulting in extremely low O&M efficiency and difficulty in achieving mass deployment. Furthermore, automated tools are highly dependent on the operating environment at the operating system level. If a server system crashes or has not yet started, automated O&M tasks cannot be deployed, and the underlying hardware faults cannot be detected and repaired in a timely manner, greatly affecting the timeliness of troubleshooting.

[0031] At the same time, related technologies also face challenges with large-scale cluster operations and maintenance. This involves using purely manual operations to maintain large server clusters, resulting in low maintenance efficiency and poor operational consistency due to varying operating habits among administrators. This also results in low fault tolerance, making it fundamentally unable to meet the requirements of intelligent operations and maintenance in modern data centers. Furthermore, related technologies lack the ability to integrate out-of-band management channels with automated operational processes. When the operating system becomes unavailable, full-stack automated control, from the hardware layer to the system layer, cannot be achieved, significantly limiting the flexibility and reliability of operations and maintenance.

[0032] In order to solve the problems existing in the relevant solutions, the embodiment of the present application proposes to deeply integrate the KVM function of BMC with the automation script engine to realize the construction of an automated management system at the BMC level. As the core technology of remote server management, the KVM function of BMC can break through the limitations of the operating system and support remote control when the server is not started; and the automation script engine can convert complex operation and maintenance processes into executable instruction sequences. After the two are combined, even if the operating system crashes or is not running, the system can automatically perform operation and maintenance tasks such as diagnosis and repair according to the preset script. For example, in a large-scale data center scenario, the server operation and maintenance method of the present application can replace manual operation of each server, automatically complete batch maintenance and troubleshooting of servers, significantly improve operation and maintenance efficiency and operation consistency, enhance system reliability, and achieve more intelligent and efficient remote management of servers.

[0033] The server operation and maintenance method of the present application relates to the field of server management technology. By deeply integrating the KVM function of the baseboard management controller (BMC) with the automation script engine, out-of-band automated operation and maintenance control is realized, which is suitable for remote host (HOST) configuration scenarios in the baseboard management controller (BMC) environment.

[0034] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0035] Figure 1 A flow chart of a server operation and maintenance method provided in an embodiment of the present disclosure.

[0036] like Figure 1 As shown, the method comprises the following steps:

[0037] Step 101: Based on the operation and maintenance task of the target server submitted by the user, a structured instruction file corresponding to the operation and maintenance task is parsed.

[0038] In the present application, the structured instruction file is pre-generated based on the user's operation behavior. The structured instruction file includes atomic actions, and the atomic actions include trigger conditions, input sequences, and expected feedback.

[0039] In some embodiments, the present application parses a pre-generated structured instruction file based on an operation and maintenance task submitted by a user (such as "configuring BIOS startup items").

[0040] A structured instruction file is a standardized configuration file (such as JSON or XML format) generated based on user operation behaviors (such as mouse clicks and keyboard input), including atomic actions, which is used to describe the complete operation and maintenance process under the current operation and maintenance task.

[0041] An atomic action is the smallest operation unit, consisting of three elements: trigger conditions, input sequence, and expected feedback.

[0042] The trigger condition indicates the environmental state that needs to be verified before execution; the input sequence indicates the specific operation instructions; and the expected feedback indicates the result that should be presented after the operation.

[0043] Step 102 : Based on the triggering conditions, input sequence, and expected feedback of the atomic action in the structured instruction file and in combination with the screen image of the target server, verify and execute the atomic action.

[0044] In this application, the screen image is captured from the target server through the hardware interface.

[0045] In some embodiments, the real-time screen image of the target server is obtained through a hardware interface, the trigger condition is verified in combination with an image matching algorithm, the input sequence is executed, and the operation result is verified through expected feedback, forming a closed loop from condition verification to operation execution to result confirmation.

[0046] The hardware interface is an underlying communication channel that does not rely on the operating system (such as the BMC's VGA capture interface), ensuring that screen information can still be obtained when the system crashes.

[0047] The screen image is a screenshot of the server interface at a fixed frame rate (such as 30fps) and is used to compare the similarity between the current image and the preset trigger image, and the difference between the image after the operation and the expected feedback image.

[0048] Step 103: When the atomic action is completed, it is determined that the operation and maintenance task of the target server is successfully executed.

[0049] In some embodiments, when the input events in the input sequence of the atomic action of the current structured instruction file are executed in sequence and pass verification, it is determined that the operation and maintenance task of the current target server is successful.

[0050] Among them, it should be noted that the target server operation and maintenance tasks of this application are applicable to the maintenance operations of a single target server, and can also be extended to clustered operation and maintenance scenarios of batch target servers. For the maintenance operations of a single target server, its structured instruction file can be uploaded by the user or pre-generated and stored by the substrate management system directly according to the user's operation behavior. For the clustered operation and maintenance scenarios of batch target servers, users can batch upload the structured instruction files corresponding to the current operation and maintenance tasks to multiple target servers, so that multiple target servers can execute the operation and maintenance tasks synchronously, thereby improving operation and maintenance efficiency.

[0051] Through this application, based on the operation and maintenance tasks of the target server submitted by the user, the structured instruction file corresponding to the operation and maintenance tasks is parsed. The structured instruction file is pre-generated based on the user's operation behavior. The structured instruction file includes atomic actions, and the atomic actions include trigger conditions, input sequences, and expected feedback. Based on the trigger conditions, input sequences, and expected feedback of the atomic actions in the structured instruction file, combined with the screen image of the target server, the atomic actions are verified and executed. The screen image is captured from the target server through the hardware interface. When the atomic action is executed, it is determined that the operation and maintenance tasks of the target server have been successfully executed. By parsing the structured instruction file generated based on the user's operation behavior, the operation and maintenance tasks are decomposed into atomic actions containing trigger conditions, input sequences, and expected feedback. The closed-loop automated operation from trigger verification to input execution to result verification is realized by combining the screen image captured by the hardware interface. The operation and maintenance tasks can be accurately executed in out-of-band management scenarios that are separated from the operating system (such as when the system crashes), thereby improving the operation and maintenance efficiency.

[0052] Figure 2 The following further illustrates a flow chart of another server operation and maintenance method proposed in the present disclosure. Figure 1 In the embodiment shown, step 101 and step 102 are further explained. Figure 2 The following steps may be included.

[0053] Step 201: Acquire an input event sent by a browser and a screen image corresponding to the input event.

[0054] In this application, an input event refers to the user operation behavior performed by the browser in the remote control interface of the baseboard management controller for the operation and maintenance tasks of the target server, which is recorded by the browser. The screen image includes the trigger screen image and feedback screen image of the target server captured at a fixed frame rate through the hardware interface when the user triggers the input event and after the input event is executed.

[0055] In some embodiments, during the server operation and maintenance automation process, the present application dynamically captures user operation behaviors and converts them into reusable instruction scripts.

[0056] Specifically, when a user logs in to the remote KVM control page through the BMC's web interface, activates the "Command Recording" function, and then performs a target operation, such as powering on or off the computer, or performing BIOS settings, the browser and the BMC establish a two-way communication channel via the WebSocket protocol, transmitting the user's input events for the target operation and the corresponding server video stream data in real time.

[0057] To ensure the accuracy and reliability of data transmission, all messages transmitted by both parties follow a specific format, including a fixed magic number, message type (HID mouse input, HID keyboard input, etc.), data packet status (reply request, encryption identifier, CRC checksum, etc.), message body CRC checksum value, message body length, and variable-length message body content.

[0058] The KVM function in the BMC uses a USB interface (the hardware interface in this application) connected to the host (the target server being managed) and a HID (the human interface device in this application) driver to virtualize input devices such as the host's mouse and keyboard. It also captures video image information output by the VGA graphics card (the screen image of the target server in this application). This enables the BMC to simulate real user operations and monitor changes in server interface status.

[0059] In recording mode, the BMC encodes captured input events and screen images. Input events are subdivided into mouse events and keyboard events, and the corresponding timestamps are recorded.

[0060] For mouse events, you can capture mouse movements, clicks, drags, and other operations through WebSocket, recording coordinates (using relative coordinates to adapt to different resolutions), timestamps, and operation types. For keyboard events, you can capture key press and release events, record unified USB HID key codes and timestamps, and eliminate the impact of operating system differences.

[0061] For screen image processing, this application uses a VGA capture module to capture host screen images at a fixed frame rate of 15fps and save them as RGB bitmaps or compressed formats such as JPEG / PNG. Based on the BMC's hardware clock, each input event and screen image is timestamped with microsecond-level precision, ensuring strict timing consistency between operation events and image frames.

[0062] Step 202 : In response to a user completion operation triggered by the user on the remote control interface, the input event is time-sequentially aligned with the triggering screen image and the feedback evaluation image to obtain a structured instruction file of the operation and maintenance task.

[0063] In some embodiments, based on a timestamp with preset precision, for each input event, the input event is time-aligned with the trigger screen image and feedback screen image corresponding to the input event; multiple trigger screen images after time alignment are combined into trigger conditions, multiple input events after time alignment are combined into multiple input sequences, and multiple feedback screen images after time alignment are combined into multiple expected feedbacks to obtain atomic actions corresponding to the operation and maintenance tasks; the atomic actions are encapsulated according to a preset format to obtain a structured instruction file.

[0064] Among them, the present application can adopt Delta Encoding technology to only store the difference between consecutive frames, greatly reducing the data storage volume. In other words, based on the timestamp of preset precision, for each input event, before the input event and the trigger screen image and feedback screen image corresponding to the input event are time-aligned, the present application includes: differential encoding the trigger screen image and the feedback screen image respectively to obtain trigger image difference data of each trigger screen image and the adjacent image of each trigger screen image, and feedback image difference data of each feedback screen image and the adjacent image of each feedback screen image; based on the trigger image difference data and the feedback image difference data, obtaining the trigger screen image and feedback screen image to be processed.

[0065] Specifically, after completing the user operation recording, the present application can perform differential encoding on the trigger screen image and the feedback screen image before timing alignment. By calculating the difference between each trigger screen image and the adjacent image, trigger image difference data is obtained; similarly, feedback image difference data between the feedback screen image and the adjacent image is obtained. Based on this difference data, the trigger screen image and feedback screen image to be processed are generated to reduce the amount of data and improve the efficiency of subsequent processing.

[0066] Using timestamps with a preset precision, each input event can be accurately associated with its corresponding trigger screen image (before the operation) and feedback screen image (after the operation). Subsequently, multiple trigger screen images, aligned in time, are combined into trigger conditions, multiple input events are sequentially integrated into input sequences, and multiple feedback screen images are aggregated into expected feedback, thus forming the smallest operational unit of the operation and maintenance task—the atomic action.

[0067] For example, the atomic action "click BIOS menu item A" is triggered by the matching of the BIOS main interface image. The input sequence includes mouse movement and click events, and the expected feedback is the appearance of the "Save&Exit" dialog box.

[0068] The atomic actions are encapsulated in a preset format (such as YAML) to form a structured instruction file that includes the operation type, coordinate positioning, input sequence, and image matching conditions.

[0069] After the user terminates recording, the BMC completes the timing alignment and logical association between the input events and the image frames, generating a structured instruction file. The user can download this file and upload it to the target server where the operation and maintenance task needs to be performed. The BMC's script engine automatically replays the instructions in subsequent configuration tasks, enabling batch replication of server configurations, such as performing BIOS parameter modifications on multiple servers.

[0070] Step 203: Based on the operation and maintenance task of the target server submitted by the user, a structured instruction file corresponding to the operation and maintenance task is parsed.

[0071] In some embodiments, after the BMC receives the target server operation and maintenance task submitted by the user, the present application may parse and execute the structured instruction file corresponding to the operation and maintenance task.

[0072] Specifically, this application can obtain and verify the legitimacy of structured instruction files uploaded by users through the Redfish interface. The structured instruction files (including operation sequences, image matching conditions, etc.) are encapsulated in JSON format that complies with the Redfish Schema specification, and the operation type is defined as POST.

[0073] Users can upload files to the BMC's Redfish endpoint (for example, ` / redfish / v1 / Managers / bmc / Actions / Oem / UploadScript`) through the HTTPS protocol so that the BMC can obtain the structured instruction file.

[0074] Before parsing the structured instruction file, this application must first obtain the structured instruction file corresponding to the operation and maintenance task and perform a security check on the structured instruction file. When the security check result meets the preset security conditions, the structured instruction file is parsed. Specifically, the BMC performs a double security check on the structured instruction file: first, it verifies the integrity of the file through digital signatures to prevent malicious tampering; second, it verifies user permissions (administrator role required) to ensure the legality of the operation. If the security check passes, the BMC will parse the structured instruction file.

[0075] After verification, this application can trigger instruction execution through Redfish's Action interface (such as ` / redfish / v1 / Managers / bmc / Action / Oem / RunScript`) and use Redfish's TaskService interface (such as ` / redfish / v1 / TaskService / Tasks / {TaskID}`) to track task progress in real time.

[0076] Step 204: When determining to trigger an atomic action in the structured instruction file, determine the first input event in the input sequence that currently needs to be triggered.

[0077] In the present application, the trigger conditions and expected feedback are related to the input sequence, the input sequence includes multiple input events that need to be executed, the trigger conditions include multiple preset trigger screen images before the execution of the multiple input events, and the expected feedback includes multiple preset feedback screen images after the execution of the multiple input events.

[0078] In some embodiments, after the BMC parses the structured instruction file, the present application can call the virtual input device driver and image matching module according to a preset timing: by simulating input events of input devices such as a mouse and keyboard, combining the real-time captured screen image of the target server with the preset trigger screen image and the preset feedback screen image for comparison, dynamically adjust the instruction execution conditions, and ensure that the operation is triggered only under the target interface state (such as the completion of the BIOS main menu loading).

[0079] Specifically, after BMC parses the structured instruction file, when BMC determines that it needs to trigger an atomic action in the structured instruction file, it will first locate the first input event in the input sequence that needs to be executed. In this mechanism, the trigger condition and expected feedback are deeply associated with the input sequence: the input sequence contains multiple input events that are executed in sequence (such as mouse clicks, keyboard input, etc.), and the trigger condition is composed of preset trigger screen images before the execution of multiple input events (such as the interface status that needs to be verified before the operation), and the expected feedback is composed of preset feedback screen images after the execution of multiple input events (such as the result interface that needs to be verified after the operation). Through this association mechanism, it is ensured that the execution of each input event is based on the correct interface state, and the validity of the operation is verified through image matching after execution, forming a closed-loop control logic of "trigger condition verification to input event execution to expected feedback verification".

[0080] Step 205 : Execute the first input event and verify whether the first input event is successfully executed according to the first preset trigger screen image in the trigger condition corresponding to the first input event and the first preset feedback screen image in the expected feedback.

[0081] In some embodiments, when determining the first input event in a triggered input sequence, the first screen image and the first preset trigger screen image are compared to determine whether the first screen image and the first preset trigger screen image match, and the first screen image is the screen image of the target server when the first input event is triggered; when the first screen image and the first preset trigger screen image match, the virtual human-computer interface device driver is called to execute the first input event; the second screen image and the first preset feedback screen image are compared to determine whether the second screen image and the first preset feedback screen image match, and the second screen image is the screen image of the target server after the first input event is executed; when the second screen image matches the first preset feedback screen image, it is determined that the first input event is executed successfully.

[0082] In other words, when the BMC triggers the first input event in the input sequence, the following verification execution process must be followed: Trigger condition verification: compare the current screen image (first screen image) with the preset trigger image, and execute the input event only when a match is successful; Input event execution: call the virtual HID driver to simulate mouse / keyboard operations (such as clicking coordinates (300, 200)); Expected feedback verification: after execution, compare the new screen image (second screen image) with the preset feedback image to confirm that the operation is effective.

[0083] Among them, comparing the first screen image and the first preset trigger screen image to determine whether the first screen image and the first preset trigger screen image match each other includes: preprocessing the first screen image and the first preset trigger screen image to obtain the preprocessed first screen image and the first preset trigger screen image; determining the normalized mutual correlation value between the preprocessed first screen image and the first preset trigger screen image through a matching algorithm; when the normalized mutual correlation value is greater than or equal to a preset mutual correlation threshold, extracting feature points of the preprocessed first screen image and the first preset trigger screen image respectively, and performing feature point matching based on the feature points to obtain a Hamming distance; when the Hamming distance is less than or equal to the preset distance threshold, determining that the first screen image and the first preset trigger screen image match each other.

[0084] Among them, comparing the second screen image with the first preset feedback screen image to determine whether the second screen image and the first preset feedback screen image match each other includes: preprocessing the second screen image and the first preset feedback screen image to obtain the preprocessed second screen image and the first preset feedback screen image; determining the normalized mutual correlation value between the preprocessed second screen image and the first preset feedback screen image through a matching algorithm; when the normalized mutual correlation value is greater than or equal to a preset mutual correlation threshold, extracting feature points of the preprocessed second screen image and the first preset feedback screen image respectively, and performing feature point matching based on the feature points to obtain a Hamming distance; when the Hamming distance is less than or equal to the preset distance threshold, determining that the second screen image and the first preset feedback screen image match each other.

[0085] Specifically, when the first screen image is compared with the first preset trigger screen image and the second screen image is compared with the second preset trigger image, a double-layer verification of “template matching + feature point matching” is adopted.

[0086] Before matching, the image needs to be preprocessed, that is, the image resolution needs to be unified (such as scaling to 720p), the brightness and contrast need to be adjusted, and key areas (such as the BIOS menu button, OS login box, etc.) need to be focused to reduce the amount of calculation.

[0087] In an optional embodiment of the present application, the template matching of the present application may adopt an NCC algorithm to calculate the normalized cross-correlation value between the first screen image and the first preset trigger screen image. The NCC algorithm is as follows:

[0088]

[0089] Where, I is the first input screen image, (x, y) is the pixel coordinate, I(x, y) is the pixel value of the pixel point (x, y), T is the first preset trigger screen image, is the mean value of the pixels of the first screen image, The NCC is the average value of the pixels in the first preset trigger screen image. The NCC value is the normalized cross-correlation value, which ranges from -1 to 1. A value closer to 1 indicates a higher match. A preset cross-correlation threshold (e.g., NCC ≥ 0.85) is set based on the scenario. A match is considered successful if the threshold is exceeded.

[0090] When the normalized cross-correlation value is greater than or equal to the preset cross-correlation threshold, feature point matching is performed. The feature point matching uses the ORB (Oriented FAST and Rotated BRIEF, improved FAST corner detection + BRIEF descriptor) algorithm, which is faster and suitable for real-time systems. First, it is necessary to extract the 256-bit binary descriptors of the second screen image and the first preset feedback screen image based on the BRIEF algorithm, and calculate the Hamming distance (exclusive OR operation) for matching. Among them, this application can dynamically adjust the number of feature points according to the complexity of the scene (for example, the BIOS interface only requires 50 feature points, and the complex OS interface requires 500 feature points).

[0091] When the Hamming distance is less than or equal to the preset distance threshold, it is determined that the second screen image matches the first preset feedback screen image.

[0092] It is understandable that the matching of the second screen image of the present application with the first preset feedback screen image can also be achieved by referring to the above steps, which will not be repeated here.

[0093] In addition, if the trigger condition or feedback verification fails, it will retry according to the preset strategy (such as waiting for 5 seconds and then retrying 3 times). If it still fails, it will trigger a hardware-level repair (such as a power reset).

[0094] In other words, when the first screen image does not match the first preset trigger screen image, or when the second screen image does not match the first preset feedback screen image, the first input event execution is determined to be abnormal, and the exception handling mechanism is triggered. Triggering the exception handling mechanism includes: repeatedly triggering the first input event according to a preset wait time (e.g., 5 seconds for the first time, doubling the wait time thereafter) and a preset number of retries (e.g., a maximum of 3 retries); and, when the first screen image after the first input event is repeatedly triggered does not match the first preset trigger screen image, or when the second screen image after the first input event is repeatedly triggered does not match the first preset feedback screen image, performing a repair operation (e.g., restarting the fan, resetting the power supply) via the hardware interface.

[0095] Specifically, during the execution of input events, the BMC continuously monitors the operating status. If anomalies such as VGA image matching failure or host unresponsiveness occur, an alarm is immediately triggered. The BMC also collects real-time server hardware sensor data (such as temperature and voltage) and initiates automatic repair actions (such as fan speed adjustment and power reset) if a hardware anomaly is detected.

[0096] For recoverable errors (such as communication interruption caused by network jitter), an exponential backoff strategy is used to automatically retry the operation. That is, the interval between each retry increases exponentially (such as 5 seconds for the first time and 10 seconds for the second time) until the operation is successful or the maximum number of retries is reached, ensuring the reliability and stability of operation and maintenance tasks.

[0097] Step 206 : When the first input event is successfully executed, triggering a second input event is determined according to the order of the multiple input events in the input sequence, until the multiple input events are successfully executed.

[0098] In some embodiments, once the first input event successfully executes (i.e., passes trigger condition verification and matches expected feedback), the BMC automatically triggers subsequent events according to the pre-set order of the input sequence. For example, if the input sequence is "Click the BIOS menu → Select Advanced Options → Change Boot Order," the next action will be executed immediately after the previous one completes, until all input events have been executed. This chained execution mechanism ensures the complete execution of complex O&M tasks (such as multi-step BIOS configuration).

[0099] During the execution of the input event, the BMC may also obtain the key task status after the input event is triggered in the input sequence and the server parameters of the target server corresponding to the key task status at a preset period (eg, every 30 seconds).

[0100] Key mission status data includes recorded operation execution progress, image matching success rate, number of exception retries, etc.; server parameters include hardware indicators such as CPU temperature, memory usage, power status, etc. obtained through the BMC interface (i.e., the preset interface of this application), as well as configuration parameters such as BIOS version and boot order.

[0101] Mission-critical status data and server parameter data can be used to assess operational impact in real time and provide a basis for report generation.

[0102] When generating an operation and maintenance report, the total number of tasks, number of successes / failures, and average execution time can also be counted; the trigger time, execution result, image matching score of each input event, temperature change curve before and after the operation, and power supply stability analysis can be determined; and a comparison table of target parameters and actual effective values ​​can be determined (such as expected BIOS version vs. actual version).

[0103] After obtaining the operation and maintenance report, it can be pushed to the administrator's mailbox (i.e. the target mailbox of this application) or the enterprise storage server (i.e. the storage server of this application) through the Redfish EventDestination interface; and stored locally in the BMC log partition (i.e. the target storage area of ​​this application), supporting the retrieval of historical operation and maintenance reports by timestamp, task ID or error type.

[0104] Step 207: When the execution of multiple input events is completed, it is determined that the operation and maintenance task of the target server is executed successfully.

[0105] In some embodiments, when all input events in the input sequence are successfully executed in sequence, the BMC will confirm that the operation and maintenance task of the target server is successfully executed, and at the same time, it can also feedback the successful execution of the task to the user.

[0106] For automated server cluster operations, users can submit batch operations for multiple servers through the Redfish API or web management interface. Because structured command files are encapsulated in a batch execution format compliant with the Redfish protocol, each server can utilize these files for batch operations, enabling efficient, automated management of large-scale server operations.

[0107] In summary, this application generates reusable structured instruction files by recording user operation behaviors, completely innovating the repetitive manual configuration mode in related technologies and greatly improving server deployment efficiency; using image matching technology to accurately identify screen images such as BIOS menus and OS login pages to ensure that instructions are executed in the correct environment; supporting cross-platform operation recording and widely adapting to various server types and operating system environments; structured instruction files support parameterized modification and can be flexibly adjusted according to different server configuration requirements; through anomaly detection and automatic recovery mechanisms, effectively ensuring the success rate of tasks in large-scale deployment scenarios.

[0108] based on Figure 1 、 Figure 2 The embodiment shown, as Figure 3 As shown, the present application provides a flow chart of generating a structured instruction file during a specific server operation and maintenance process.

[0109] Reference Figure 3 The user logs in to the BMC webpage on a PC browser. This is the starting point for operations, building a bridge for interaction with the server management module. Once logged in, the KVM page opens. KVM enables remote control of the server, providing an environment foundation for subsequent operations.

[0110] Click the "Generate Commands - Start" button to start recording. Then, perform actual operations on the host through KVM. These operations will be recorded and used to generate commands. After completing the operations, click the "Generate Commands - End" button to terminate the recording. Finally, click the "Download Commands" button to obtain a structured command file containing the operation logic. This can be reused in subsequent server configurations to improve deployment efficiency.

[0111] Logging in to the BMC webpage and opening the KVM page on a PC browser establishes a remote control channel. This step requires the browser to continuously capture user operations on the KVM interface (i.e., input events, user actions on target server operation and maintenance tasks) and simultaneously capture screen images (triggering / feedback screen images) at a fixed frame rate through the hardware interface. This is a preparatory step for the user.

[0112] Clicking the "Generate Command - Start" button starts the data collection mark. At this time, the browser starts recording, and the BMC starts to obtain the browser-recorded user's behavior of operating the HOST through KVM (input events). At the same time, the BMC captures the screen images before and after the operation (trigger / feedback images) according to the rules.

[0113] During the KVM operation host phase, the BMC obtains input events recorded by the browser while collecting screen images during and after the operation is triggered. It also performs differential encoding on the images (compressing redundant data while retaining differential information) to make the data more lightweight.

[0114] Clicking "Generate Command - End" terminates acquisition. The BMC now obtains the complete sequence of input events and corresponding screen images. It then aligns the input events and trigger / feedback screen images based on timestamps, breaking this data down into "trigger conditions (trigger image combinations), input sequences (operation event combinations), and expected feedback (feedback image combinations)." These data are then packaged into atomic actions, ultimately generating a structured command file for download and reuse. Click the "Download Command" button to download.

[0115] For ease of understanding, Figure 4 As shown, the present application also provides an architecture diagram of a BMC-side instruction generation system.

[0116] Reference Figure 4 ,The PC browser is the user operation entrance, running the ,“command generation” and “command download” functions, communicating with the BMC through ,network protocols to achieve remote operation and maintenance control.

[0117] BMC is the core module of server management, which undertakes the functions of "KVM (remote control)" and "recording operation". It simulates user operations to the HOST (host, that is, the target server of this application) and records operation data to generate structured instruction files.

[0118] The HOST is the server that is actually operated and maintained, and receives user operations (such as configuration and debugging) simulated by the BMC.

[0119] During the structured command file generation phase, after the user clicks "Generate Command" in the PC browser, the user's operation behavior (i.e., input events in this application) is sent to the BMC's KVM module via the Websocket protocol (real-time, two-way communication, suitable for operation stream transmission). The BMC's KVM module then executes the input events on the host side through the USB channel (simulating local USB operations, allowing remote operations to be performed as if they were local), enabling remote control of the server (such as BIOS configuration and system debugging).

[0120] At the same time, the BMC's operation recording module will synchronously capture and record the user's operation behaviors performed through KVM, accumulating data for generating structured instructions.

[0121] During the structured instruction file download phase, when the user completes the operation and needs to reuse the structured instruction file, the user can click "Download Instructions" in the PC browser. The BMC uses the Redfish protocol (a standardized RESTful protocol for server hardware management, suitable for structured data transmission) to return the structured instruction file (including input sequence, trigger conditions, and expected feedback) that was previously recorded in the operation module to the PC browser for the user to download and save for subsequent server batch deployment or repeated operation and maintenance.

[0122] Further, such as Figure 5 As shown, the present application provides a schematic diagram of batch server cluster operation and maintenance.

[0123] Reference Figure 5 After obtaining the structured instruction file, batch operation and maintenance of the server cluster can be performed based on the structured instruction file.

[0124] In other words, after the first server obtains the structured instruction file corresponding to the current operation and maintenance task, when subsequent servers are started, the BMC can directly load the structured instruction file uploaded by the user or stored by itself, and automatically trigger the operation and maintenance task (such as the BIOS configuration process) without manual intervention.

[0125] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0126] The embodiment of the present application further provides a server operation and maintenance device 600, Figure 6 A schematic diagram of the structure of a server operation and maintenance device provided in an embodiment of the present disclosure is shown in FIG. Figure 6 Shown, including:

[0127] Parsing unit 610 is used to parse the structured instruction file corresponding to the operation and maintenance task of the target server submitted by the user. The structured instruction file is pre-generated based on the user's operation behavior and includes atomic actions. The atomic actions include trigger conditions, input sequences, and expected feedback.

[0128] An operation and maintenance unit 620 is configured to verify and execute the atomic action based on the triggering conditions, input sequence, and expected feedback of the atomic action in the structured instruction file and in combination with a screen image of the target server, where the screen image is captured from the target server via a hardware interface;

[0129] The determining unit 630 is configured to determine whether the operation and maintenance task of the target server is successfully executed when the atomic action is completed.

[0130] Furthermore, in a possible implementation of the embodiment of the present disclosure, the trigger condition and the expected feedback are related to an input sequence, the input sequence includes multiple input events that need to be executed, the trigger condition includes multiple preset trigger screen images before the execution of the multiple input events, and the expected feedback includes multiple preset feedback screen images after the execution of the multiple input events. The operation and maintenance unit 620 is used to: when determining to trigger an atomic action in a structured instruction file, determine the first input event in the input sequence that currently needs to be triggered; execute the first input event and verify whether the first input event is executed successfully according to the first preset trigger screen image in the trigger condition corresponding to the first input event and the first preset feedback screen image in the expected feedback; when the first input event is executed successfully, determine to trigger the second input event according to the arrangement order of the multiple input events in the input sequence, until the multiple input events are executed successfully.

[0131] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: when determining the first input event in the trigger input sequence, compare the first screen image with the first preset trigger screen image to determine whether the first screen image and the first preset trigger screen image match, and the first screen image is the screen image of the target server when the first input event is triggered; when the first screen image and the first preset trigger screen image match, call the virtual human-machine interface device driver to execute the first input event; compare the second screen image with the first preset feedback screen image to determine whether the second screen image and the first preset feedback screen image match, and the second screen image is the screen image of the target server after the first input event is executed; when the second screen image matches the first preset feedback screen image, determine that the first input event is executed successfully.

[0132] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: preprocess the first screen image and the first preset trigger screen image to obtain the preprocessed first screen image and the first preset trigger screen image; determine the normalized mutual correlation value between the preprocessed first screen image and the first preset trigger screen image through a matching algorithm; when the normalized mutual correlation value is greater than or equal to a preset mutual correlation threshold, extract feature points of the preprocessed first screen image and the first preset trigger screen image respectively, and perform feature point matching based on the feature points to obtain a Hamming distance; when the Hamming distance is less than or equal to the preset distance threshold, determine that the first screen image and the first preset trigger screen image are matched.

[0133] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: preprocess the second screen image and the first preset feedback screen image to obtain the preprocessed second screen image and the first preset feedback screen image; determine the normalized mutual correlation value between the preprocessed second screen image and the first preset feedback screen image through a matching algorithm; when the normalized mutual correlation value is greater than or equal to a preset mutual correlation threshold, extract feature points of the preprocessed second screen image and the first preset feedback screen image respectively, and perform feature point matching based on the feature points to obtain a Hamming distance; when the Hamming distance is less than or equal to the preset distance threshold, determine that the second screen image and the first preset feedback screen image match.

[0134] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: after comparing the second screen image with the first preset feedback screen image to determine whether the second screen image matches the first preset feedback screen image, when the first screen image does not match the first preset trigger screen image or the second screen image does not match the first preset feedback screen image, determine that the first input event execution is abnormal and trigger the exception handling mechanism.

[0135] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: repeatedly trigger the first input event according to a preset waiting time and a preset number of retries; when the first screen image after the first input event is repeatedly triggered does not match the first preset trigger screen image or the second screen image after the first input event is repeatedly triggered does not match the first preset feedback screen image, perform a repair operation through the hardware interface.

[0136] Furthermore, in a possible implementation of the embodiment of the present disclosure, the parsing unit 610 is used to: obtain a structured instruction file corresponding to the operation and maintenance task; perform a security check on the structured instruction file; and parse the structured instruction file when the security check result meets the preset security conditions.

[0137] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: obtain the critical task status after the input event in the input sequence is triggered and the server parameters of the target server corresponding to the critical task status according to a preset period; generate an operation and maintenance report based on the critical task status and server parameters; store the operation and maintenance report in the target storage area, and push it to the target mailbox or storage server through a preset interface.

[0138] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: obtain the input event sent by the browser and the screen image corresponding to the input event before parsing the structured instruction file corresponding to the operation and maintenance task based on the target server submitted by the user, the input event refers to the user operation behavior recorded by the browser through the browser side in the remote control interface of the baseboard management controller for the operation and maintenance task of the target server, and the screen image includes the trigger screen image and feedback screen image of the target server captured at a fixed frame rate through the hardware interface when the user triggers the input event and after the input event is executed; in response to the user completion operation triggered by the user in the remote control interface, the input event is time-aligned with the trigger screen image and the feedback evaluation image to obtain the structured instruction file of the operation and maintenance task.

[0139] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: based on a timestamp with preset precision, for each input event, time-align the input event with the trigger screen image and feedback screen image corresponding to the input event; combine the multiple trigger screen images after time alignment into trigger conditions, combine the multiple input events after time alignment into multiple input sequences, and combine the multiple feedback screen images after time alignment into multiple expected feedbacks to obtain the atomic action corresponding to the operation and maintenance task; encapsulate the atomic action according to a preset format to obtain a structured instruction file.

[0140] Furthermore, in a possible implementation of the embodiment of the present disclosure, the operation and maintenance unit 620 is used to: before timing-aligning the input event with the trigger screen image and feedback screen image corresponding to the input event based on a timestamp with preset precision, for each input event, perform difference encoding on the trigger screen image and the feedback screen image, respectively, to obtain trigger image difference data between each trigger screen image and each adjacent image of the trigger screen image, and feedback image difference data between each feedback screen image and each adjacent image of the feedback screen image; and obtain the trigger screen image and feedback screen image to be processed based on the trigger image difference data and the feedback image difference data.

[0141] For the description of the features in the embodiment corresponding to the server operation and maintenance device, please refer to the relevant description of the embodiment corresponding to the server operation and maintenance method, and will not be repeated here.

[0142] An embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any of the above-mentioned server operation and maintenance method embodiments.

[0143] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above-mentioned server operation and maintenance method embodiments when running.

[0144] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0145] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any of the above-mentioned server operation and maintenance method embodiments are implemented.

[0146] An embodiment of the present application also provides another computer program product, including a non-volatile computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps in any of the above-mentioned server operation and maintenance method embodiments.

[0147] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0148] The above is a detailed introduction to a server operation and maintenance method, electronic device, storage medium and product provided by the present application. This article uses specific examples to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A server operation and maintenance method, characterized in that: include: Based on the operation and maintenance task of the target server submitted by the user, a structured instruction file corresponding to the operation and maintenance task is parsed. The structured instruction file is pre-generated based on the user's operation behavior. The structured instruction file includes atomic actions, and the atomic actions include trigger conditions, input sequences, and expected feedback; Based on the trigger condition, input sequence, and expected feedback of the atomic action in the structured instruction file, and in combination with a screen image of the target server, verifying and executing the atomic action, wherein the screen image is captured from the target server via a hardware interface, the trigger condition and the expected feedback are related to the input sequence, wherein the input sequence includes multiple input events to be executed, the trigger condition includes multiple preset trigger screen images before the multiple input events are executed, and the expected feedback includes multiple preset feedback screen images after the multiple input events are executed; When the atomic action is completed, it is determined that the operation and maintenance task of the target server is successfully executed; The verifying and executing operation on the atomic action based on the triggering condition, input sequence, and expected feedback of the atomic action in the structured instruction file and in combination with the screen image of the target server includes: When it is determined to trigger the atomic action in the structured instruction file, determining the first input event in the input sequence that currently needs to be triggered; When it is determined that a first input event in the input sequence is triggered, comparing a first screen image with a first preset trigger screen image to determine whether the first screen image and the first preset trigger screen image match, the first screen image being a screen image of the target server when the first input event is triggered; When the first screen image matches the first preset trigger screen image, calling a virtual human-machine interface device driver to execute the first input event; comparing a second screen image with a first preset feedback screen image to determine whether the second screen image matches the first preset feedback screen image, where the second screen image is a screen image of the target server after the first input event is executed; When the second screen image matches the first preset feedback screen image, determining that the first input event is executed successfully; When the first input event is successfully executed, triggering a second input event is determined according to the arrangement order of the multiple input events in the input sequence until the multiple input events are successfully executed.

2. The method according to claim 1, characterized in that The comparing the first screen image and the first preset trigger screen image to determine whether the first screen image and the first preset trigger screen image match includes: Preprocessing the first screen image and the first preset trigger screen image to obtain a preprocessed first screen image and a first preset trigger screen image; Determining a normalized cross-correlation value between the preprocessed first screen image and the first preset trigger screen image through a matching algorithm; When the normalized mutual correlation value is greater than or equal to a preset mutual correlation threshold, extracting feature points of the preprocessed first screen image and the first preset trigger screen image respectively, and performing feature point matching based on the feature points to obtain a Hamming distance; When the Hamming distance is less than or equal to a preset distance threshold, it is determined that the first screen image matches the first preset trigger screen image.

3. The method according to claim 1, characterized in that The comparing the second screen image with the first preset feedback screen image to determine whether the second screen image matches the first preset feedback screen image includes: preprocessing the second screen image and the first preset feedback screen image to obtain a preprocessed second screen image and a first preset feedback screen image; determining a normalized mutual correlation value between the preprocessed second screen image and the first preset feedback screen image through a matching algorithm; When the normalized mutual correlation value is greater than or equal to a preset mutual correlation threshold, extracting feature points of the preprocessed second screen image and the first preset feedback screen image respectively, and performing feature point matching based on the feature points to obtain a Hamming distance; When the Hamming distance is less than or equal to a preset distance threshold, it is determined that the second screen image matches the first preset feedback screen image.

4. The method according to claim 1, wherein After comparing the second screen image with the first preset feedback screen image to determine whether the second screen image matches the first preset feedback screen image, the method includes: When the first screen image does not match the first preset trigger screen image or the second screen image does not match the first preset feedback screen image, it is determined that the first input event is executed abnormally, and an exception handling mechanism is triggered.

5. The method according to claim 4, characterized in that The trigger exception handling mechanism includes: Repeatedly triggering the first input event according to a preset waiting time and a preset number of retries; When the first screen image after the first input event is repeatedly triggered does not match the first preset trigger screen image or the second screen image after the first input event is repeatedly triggered does not match the first preset feedback screen image, a repair operation is performed through the hardware interface.

6. The method according to claim 1, characterized in that The step of parsing the structured instruction file corresponding to the target server operation and maintenance task based on the target server operation and maintenance task submitted by the user includes: Obtaining a structured instruction file corresponding to the operation and maintenance task; Performing a security check on the structured instruction file; When the security verification result of the security verification meets the preset security condition, the structured instruction file is parsed.

7. The method according to any one of claims 1 to 6, characterized in that The method further comprises: According to a preset period, obtaining the key task status after the input event in the input sequence is triggered and the server parameters of the target server corresponding to the key task status; Generate an operation and maintenance report based on the key task status and the server parameters; The operation and maintenance report is stored in the target storage area and pushed to the target mailbox or storage server through a preset interface.

8. The method according to claim 1, characterized in that Before parsing the structured instruction file corresponding to the operation and maintenance task of the target server submitted by the user, the method includes: Obtaining an input event sent by the browser and a screen image corresponding to the input event, wherein the input event refers to a user operation behavior performed by the browser in a remote control interface of a baseboard management controller for an operation and maintenance task of a target server, and the screen image includes a trigger screen image and a feedback screen image of the target server captured at a fixed frame rate through the hardware interface when the user triggers the input event and after the input event is executed; In response to a user completion operation triggered by the user on the remote control interface, the input event is time-sequentially aligned with the trigger screen image and the feedback evaluation image to obtain a structured instruction file for the operation and maintenance task.

9. The method according to claim 8, characterized in that The step of performing time sequence alignment on the input event, the trigger screen image, and the feedback evaluation image to obtain the structured instruction file of the operation and maintenance task includes: Based on a timestamp with preset accuracy, for each input event, align the input event with the trigger screen image and the feedback screen image corresponding to the input event; Combining multiple trigger screen images after time sequence alignment into trigger conditions, combining multiple input events after time sequence alignment into multiple input sequences, and combining multiple feedback screen images after time sequence alignment into multiple expected feedbacks to obtain atomic actions corresponding to the operation and maintenance task; The atomic action is encapsulated according to a preset format to obtain the structured instruction file.

10. The method according to claim 9, characterized in that Before performing time sequence alignment on each input event based on the timestamp with preset precision, with respect to the input event, with the trigger screen image and the feedback screen image corresponding to the input event, the method includes: performing difference encoding on the trigger screen image and the feedback screen image respectively to obtain trigger image difference data between each trigger screen image and an image adjacent to each trigger screen image, and feedback image difference data between each feedback screen image and an image adjacent to each feedback screen image; Based on the trigger image difference data and the feedback image difference data, a trigger screen image and a feedback screen image to be processed are obtained.

11. An electronic device, characterized in that: include: memory for storing computer programs; A processor, configured to implement the steps of the server operation and maintenance method according to any one of claims 1 to 10 when executing the computer program.

12. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the server operation and maintenance method according to any one of claims 1 to 10 are implemented.

13. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the server operation and maintenance method according to any one of claims 1 to 10 are implemented.

Images