Server power on / off control system, server and method

Abstract

This application discloses a server power-on / off control system, server, and method, relating to the field of server power-on / off control technology. It includes a front-end assembly and a rear-end assembly, each equipped with a first baseboard management controller and a second baseboard management controller that communicate with each other. The system achieves automated control of the server power-on / off sequence through the coordinated operation of the first and second baseboard management controllers: during power-on, the first baseboard management controller in the front-end first triggers the power-on of the rear-end assembly and polls its status, confirming readiness before powering on the front-end; during power-off, the second baseboard management controller in the rear-end monitors the power-off status of the front-end and confirms safe shutdown before powering off the rear-end. This solves the problem in related technologies where reliance on independent script control and lack of interconnected monitoring leads to mutual interference between the power-on / off sequences of the front and rear-end assemblies, avoiding problems caused by illegal operations. It achieves timing safety and system stability under the coordinated control of the front and rear-end assemblies.

Description

Technical Field

[0001] This application relates to the field of server power-on / off control technology, and in particular to a server power-on / off control system, server, and method. Background Technology

[0002] A GPUBOX server typically consists of two parts: a GPUBOX (GPU resource module) and a host unit (or server). These two parts are connected via a high-speed PCIe interconnect interface and use two completely independent power supply systems. Powering on and off are independent of each other. Therefore, improper operation during powering on and off can easily lead to data loss or hardware damage. Summary of the Invention

[0003] This application provides a server power-on / off control system, server, and method to at least solve the problems in related technologies where the start and end of server power-on are independently controlled by scripts, which can easily lead to improper operation during power-on / off, resulting in data loss or hardware damage.

[0004] This application provides a server power-on / off control system, including: a front-end assembly, which includes a first baseboard management controller; and a rear-end assembly, which includes a second baseboard management controller. The first baseboard management controller and the second baseboard management controller communicate with each other. When the server is powered on, the first baseboard management controller acquires the server's power-on sequence and controls the front-end assembly and the rear-end assembly to perform power-on actions according to the power-on sequence. During the control process, the power-on status of the rear-end assembly is polled, and the power-on actions of the front-end assembly and the rear-end assembly are controlled according to a first polling result, where the first polling result is the polling result of the power-on status of the rear-end assembly. When the server is powered off, the second baseboard management controller acquires the server's power-off sequence and controls the front-end assembly and the rear-end assembly to perform power-off actions according to the power-off sequence. During the control process, the power-off status of the front-end assembly is polled, and the power-off actions of the rear-end assembly are controlled according to a second polling result, where the second polling result is the polling result of the power-off status of the front-end assembly.

[0005] This application also provides a server, including the structure of the above-described server power on / off control system.

[0006] This application also provides a server power-on / off control method, applied to the first baseboard management controller of the aforementioned server power-on / off control system. The method includes the following steps: acquiring the server power-on sequence when the server is powered on; controlling the head unit and tail unit to perform power-on actions according to the power-on sequence, wherein, during the control process, the power-on status of the tail unit is polled, and the power-on actions of the head unit and tail unit are controlled according to the first polling result, wherein the first polling result is the polling result of the power-on status of the tail unit.

[0007] This application also provides a server power-on / off control method, applied to the second baseboard management controller of the aforementioned server power-on / off control system. The method includes the following steps: acquiring the server's power-off sequence when the server is powered off; controlling the head assembly and tail assembly to perform power-off actions according to the power-off sequence, wherein, during the control process, the power-off state of the head assembly is polled, and the power-off action of the tail assembly is controlled according to the second polling result, the second polling result being the polling result of the power-off state of the head assembly.

[0008] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for implementing the above-described server power-on / off control method when executing the computer program.

[0009] This application also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the above-described server power-on / off control method.

[0010] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described server power-on / off control method.

[0011] This application establishes a linkage control mechanism by equipping the head unit and tail unit with first and second baseboard management controllers that can communicate with each other. During power-on, the first baseboard management controller controls the power-on process according to the power-on sequence, polling the power-on status of the tail unit to ensure its readiness before controlling the head unit to power on. During power-off, the second baseboard management controller controls the power-off process according to the power-off sequence, polling the power-off status of the head unit to ensure its safe shutdown before controlling the tail unit to power off. This mechanism solves the technical problems in related technologies caused by a lack of effective linkage and difficulty in ensuring correct power-on / off sequences, leading to system instability, data loss, hardware damage, and susceptibility to illegal operations. It achieves the technical effects of ensuring correct server power-on / off sequences, improving system stability and security, and reducing the impact of operational errors. Attached Figure Description

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

[0013] Figure 1 A structural diagram of a server power-on / off control system provided in this application embodiment;

[0014] Figure 2 A schematic diagram illustrating the implementation principle of the server power-on / off control system provided in this application embodiment;

[0015] Figure 3 This application provides a schematic diagram of the structure of a server according to an embodiment of the present application.

[0016] Figure 4 A flowchart of a server power-on / off control method provided in an embodiment of this application;

[0017] Figure 5 The control logic flowchart of the server power-on / off control system provided in the embodiments of this application;

[0018] Figure 6 A flowchart illustrating a server power-on / off control method provided in another embodiment of this application;

[0019] Figure 7 A control logic flowchart for a server power-on / off control system provided in another embodiment of this application;

[0020] Figure 8 A block diagram of a server power-on / off control device provided in an embodiment of this application;

[0021] Figure 9 A block diagram of a server power-on / off control device provided in another embodiment of this application;

[0022] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0024] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0025] Currently, with the rapid development of AI, big data processing, and other technologies, the demand for AI servers with high-density computing and easy deployment and expansion capabilities has surged. GPUBOX models, due to their modularity, low power consumption, and low cost, have become an important choice for server application vendors deploying AI tasks. However, the GPU BOX and GPU head unit are powered independently and their power on / off cycles do not affect each other. The BOX, on the other hand, must follow a specific sequence: "power on the GPU head for recognition, power off the GPU head, then power off the BOX." Incorrect timing can easily lead to data loss, hardware damage, and affect system stability.

[0026] To ensure reliable operation across all production and maintenance scenarios and avoid failures caused by human error or timing issues, a solution is needed that can automatically manage power-on / off timing and possess status monitoring and fault tolerance capabilities. The power-on / off handling methods for GPUBOX models in related technologies are as follows:

[0027] In related technology (1), PowerBox is used in conjunction with automation scripts to achieve test automation, save resources and improve test coverage, and solve the problems of server AC power-on and power-off testing relying on manual operation, low efficiency and poor accuracy; however, it is only applicable to laboratory testing, does not realize the interconnection between the machine head and BOX and status monitoring, and cannot solve the timing stability and illegal operation problems in production and online applications.

[0028] In related technology (2), automated testing is achieved by remote control and fatigue testing using Python scripts, which solves the problems of HOST-BOX products relying on specific equipment (such as PDUs and cabinets) for ACreboot testing and high testing costs; however, it relies on scripts to independently control power on and off, and there is no interconnection or status management between the machine head and the BOX, which cannot avoid the risks of timing and illegal operation in production and online applications.

[0029] In related technology (3), efficiency is improved by testing in the order of power-on and disconnection; however, it only addresses the testing process and solves the problem of low power-on testing efficiency of composite servers (including HOST and BOX), but does not realize the linkage control between the head unit and BOX, and cannot solve the stability problems caused by timing and illegal operations in production and online applications.

[0030] To address the aforementioned problems, this application proposes a server power-on / off control system, a server, and a method. To enable those skilled in the art to better understand the present application, further detailed descriptions are provided below in conjunction with the accompanying drawings and specific embodiments.

[0031] This section describes the specific application environment architecture or specific hardware architecture that the server power-on / off control system relies on.

[0032] Figure 1This is a structural diagram of a server power-on / off control system provided in an embodiment of this application.

[0033] like Figure 1 As shown, the server power-on / off control system 10 includes: a front assembly 110, a rear assembly 120, a first baseboard management controller 130, and a second baseboard management controller 140.

[0034] The head assembly 110 includes a first baseboard management controller 130, and the tail assembly 120 includes a second baseboard management controller 140. The first baseboard management controller 130 and the second baseboard management controller 140 communicate with each other. When the server is powered on, the first baseboard management controller 130 obtains the server's power-on sequence and controls the head assembly 110 and the tail assembly 120 to perform power-on actions according to the power-on sequence. During the control process, the power-on status of the tail assembly 120 is polled, and the head assembly is controlled according to the first polling result. The power-on actions of component 110 and tail component 120 are initiated by polling the power-on state of tail component 120 in the first polling result. When the server is powered off, the second baseboard management controller 140 obtains the server's shutdown sequence and controls the head component 110 and tail component 120 to perform shutdown actions according to the shutdown sequence. During the control process, the power-off state of head component 110 is polled, and the shutdown action of tail component 120 is controlled according to the second polling result, which is the polling result of the power-off state of head component 110.

[0035] It is understood that, in this embodiment of the application, the power-on status of the tail assembly 120 can be polled to ensure its readiness before controlling the head assembly to power on; during shutdown, the second baseboard management controller 140, according to the shutdown sequence, polls the power-off status of the head assembly 110 to ensure its safe shutdown before controlling the tail assembly to power off, thus forming a linkage control mechanism. Therefore, it can solve the technical problems in related technologies caused by the lack of effective linkage and difficulty in ensuring the correct power-on and power-off sequence, resulting in system instability, data loss or hardware damage, and susceptibility to illegal operations. This achieves the technical effects of ensuring the correct power-on and power-off sequence of the server, improving system stability and security, and reducing the impact of operational errors.

[0036] In this embodiment, the power-on sequence is that the tail assembly 120 is powered on earlier than the head assembly 110, and the power-off sequence is that the tail assembly 120 is powered off later than the head assembly 110.

[0037] Understandably, the power-on sequence is clearly defined as follows: tail assembly 120 powers on first, followed by head assembly 110; the power-off sequence is as follows: tail assembly 120 powers off last, followed by head assembly 110. This timing design precisely matches the core requirement of GPU BOX models: "the BOX must be powered on first for head assembly recognition, and then powered off to prevent data loss." It fundamentally avoids issues such as head assembly failing to recognize the tail assembly, data loss, or hardware damage caused by incorrect timing. Simultaneously, it lays the foundation for subsequent automated power-on and power-off control, reducing the risk of errors due to manual operation.

[0038] It should be noted that the BOX is a functional hardware module within the tail assembly, integrated into the tail assembly. Its core states are "power-on" and "power-off," which are the key criteria for the head unit to determine whether to perform power-on / power-off actions. As a slave / extension device of the head unit, it must be powered on and recognized by the head unit before the head unit can start. When shutting down, it must also be powered off only after the head unit is shut down, thereby ensuring the stability of the server's power-on / off sequence and data security.

[0039] In the embodiments of this application, such as Figure 2 As shown, at least one command transmission link is provided between the head assembly 110 and the tail assembly 120. The first baseboard management controller 130 and the second baseboard management controller 140 send target commands through the command transmission link. The target commands include at least one of a first polling command, a power-on command, a second polling command, and a power-off command. The first polling command is used to poll the power-on status of the tail assembly 120 when the server is powered on, and the second polling command is used to poll the power-off status of the head assembly 110 when the server is powered off.

[0040] Understandably, a command transmission link is established between the head assembly 110 and the tail assembly 120. The two baseboard management controllers send target commands through the command transmission link to accurately control the key aspects of the server's power-on / power-off sequence. By constructing a command transmission link and interacting with target commands, active monitoring and control of the head-to-tail and tail-to-head status is achieved, avoiding system failures and data loss caused by timing disorder.

[0041] Specifically, the front-end assembly 110 is the core control unit of the server, integrating the first baseboard management controller 130 and timing sub-components, responsible for the system's main control logic and timing scheduling. The rear-end assembly 120 is... Figure 2The right-hand "tail" area houses the server's expansion unit, integrating a second baseboard management controller 140 and a switching sub-component, working in conjunction with the front-end to achieve computing expansion functionality. The first baseboard management controller 130 is the front-end management module, sending and receiving commands via a command transmission link to control the power-on / off sequence of the front and tail. The second baseboard management controller 140 is the tail-end management module, receiving commands from the front-end, providing feedback on the tail-end status, and coordinating the execution of timing control. The command transmission link is a dedicated management link, serving as the channel for command interaction between the first and second baseboard management controllers. It transmits control commands such as power-on / power-off / polling, ensuring reliable command interaction between the baseboard management controllers. Target commands include: First polling command: Check the tail-end power-on status upon power-on; Power-on command: The control component executes the power-on action; Second polling command: Check the front-end power-off status upon power-off; Power-off command: The control component executes the power-off action.

[0042] In the embodiments of this application, such as Figure 2 As shown, the head assembly 110 includes at least one timing sub-component, the first baseboard management controller 130 is connected to the timing sub-component, the tail assembly 120 includes at least one switching sub-component, the second baseboard management controller 140 is connected to the switching sub-component, and the two ends of the command transmission link are connected to the timing sub-component and the switching sub-component.

[0043] It is understandable that by configuring at least one timing sub-component in the head assembly 110 and at least one switching sub-component in the tail assembly 120, the first baseboard management controller 130 is connected to the timing sub-component and the second baseboard management controller 140 is connected to the switching sub-component, and the two ends of the command transmission link are connected to the timing sub-component and the switching sub-component, a control architecture from the controller to the sub-component and then connected by the link is constructed.

[0044] Specifically, the timing sub-component belongs to the signal conditioning / timing control module within the front of the machine, ensuring stable command / data transmission timing. The switching sub-component belongs to the PCIe switching module within the rear of the machine, responsible for data / command forwarding and switching on the rear side.

[0045] In the embodiments of this application, such as Figure 2 As shown, an electrical interconnection link is provided between the timing sub-component and the switching sub-component, and the first baseboard management controller 130 and the second baseboard management controller 140 communicate with each other through the electrical interconnection link.

[0046] Among them, the electrical interconnection link can be a PCIe link, which is a hardware physical link that connects the timing sub-component and the switching sub-component and carries high-speed transmission of commands / data.

[0047] It is understandable that bidirectional communication between the first baseboard management controller 130 and the second baseboard management controller 140 is achieved by setting up an electrical interconnection link between the timing sub-component at the front and the switching sub-component at the rear. This provides a stable and high-speed physical channel for interaction between the baseboard management controllers, avoiding the risk of timing control failure due to communication delays / interruptions.

[0048] In this embodiment of the application, the first baseboard management controller is further configured to, after failing to obtain the first polling result, start timing from the time the first polling instruction is sent, and count whether the first polling result is obtained within a preset time period. If the first polling result is not obtained within the preset time period, the head unit performs a power-on action. If the second baseboard management controller does not provide feedback, the first polling result is not obtained.

[0049] The preset duration can be set according to the actual situation, such as 15 minutes or 16 minutes.

[0050] Understandably, to balance server power-on / off requirements with system stability, multiple fault-tolerance mechanisms are added to the process, such as power-on status detection fault tolerance: the power-on process includes detection of the BOX power-on status; if the BOX status is not detected for more than 15 minutes, the machine head will force a power-on operation. This embodiment also adds power-on status detection fault tolerance: if the first baseboard management controller cannot complete the BOX power-on operation within 15 minutes, the machine head will force a power-on operation.

[0051] According to the server power-on / off control system proposed in this application, the power-on status of the tail component is polled to ensure its readiness before the head component is powered on. During power-off, the second baseboard management controller polls the power-off status of the head component to ensure its safe power-off before controlling the tail component to power off, forming a linkage control mechanism. Therefore, it can solve the technical problems in related technologies caused by the lack of effective linkage and difficulty in ensuring the correct power-on / off sequence, such as system instability, data loss or hardware damage, and susceptibility to illegal operations. It achieves the technical effects of ensuring the correct server power-on / off sequence, improving system stability and security, and reducing the impact of operational errors.

[0052] Through the above description of the embodiments, those skilled in the art can clearly understand that the system according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0053] The implementation principle of the server power-on / off control system will be further explained below through a specific embodiment, which is divided into hardware implementation and software implementation.

[0054] At the hardware implementation level: such as Figure 2 As shown, the first baseboard management controller 130 of the nose assembly 110 and the second baseboard management controller 140 of the tail assembly 120 are interconnected via an electrical interconnection link. In addition, a command transmission link is added. This command transmission link serves as the core command transmission channel, enabling communication between the nose and tail baseboard management controllers at the hardware level. This provides a stable physical foundation for the nose to monitor the tail's power-on status in real time and for the tail to monitor the nose's power-off status in real time.

[0055] At the software implementation level, by embedding power-on / off procedures into the firmware code of the first and second baseboard management controllers, the baseboard management controllers automatically run relevant programs after startup, achieving power-on / off timing control without manual intervention. The specific process is as follows:

[0056] (1) The boot process integrated into the firmware of the baseboard management controller includes:

[0057] After the server AC is powered on, the first baseboard management controller of the front assembly starts up and sends commands (ipmitool -b 10 -t 50 chassis status) through the command transmission link to continuously poll the power-on status of the rear BOX.

[0058] If the command returns "Power off" (BOX not powered on), then a command is sent through the second baseboard management controller of the tail assembly (ipmitool -b 10 -t 50 chassis power on) to power on the BOX;

[0059] After waiting one minute to confirm that the tail section is powered on, the head section will automatically perform a power-on operation to ensure the timing rule of "tail section powered on first, head section powered on later".

[0060] (2) The shutdown process integrated into the baseboard management controller firmware includes:

[0061] After the power-on / off command is issued, the machine head will prioritize the power-off action;

[0062] The second baseboard management controller polls the power-on status of the machine head via the command transmission link: if the machine head is still "Poweron" (not powered off), the tail power-off command will not take effect; if the machine head is "Power off" (power off), the tail power-off command will be executed immediately, following the timing rule of "machine head powered off first, tail powered off last".

[0063] It should be noted that the head unit does not have a BOX, but it has a clear power-on state (Power on / Power off), and this state is recorded and managed in real time by its built-in first baseboard management controller.

[0064] Meanwhile, to balance server power-on / off requirements with system stability, multiple fault-tolerance mechanisms have been added to the process:

[0065] Power-on status detection fault tolerance (1): Add detection of BOX power-on status to the power-on process. If the BOX status cannot be detected for more than 15 minutes, the machine head will be forced to perform a power-on operation.

[0066] Power-on status detection fault tolerance (II): If the first baseboard management controller cannot complete the power-on operation of the BOX within 15 minutes, the machine head will be forced to perform a power-on operation.

[0067] Power-on command effectiveness control: When the first baseboard management controller detects that the BOX is in a power-off state, the power-on command issued to the machine head will not be effective. If power-on is required, it needs to be forced to power on via OEM (Original Equipment Manufacturer Command) command.

[0068] Power-off command effectiveness control: If the head is in the Power on state after a power-off command is issued from the tail, and the power-off command issued to the tail is ineffective, then if the BOX needs to be powered off, it needs to be forced to power off via OEM command.

[0069] Figure 3 This is a schematic diagram of the server structure provided in an embodiment of this application. The server includes the structure of the server power-on / off control system described in the above embodiment.

[0070] It should be noted that the foregoing explanation of the server power-on / off control system embodiment also applies to the server in this embodiment, and will not be repeated here.

[0071] The following will combine Figure 4 The server power-on / off control method, applied to the first baseboard management controller of the server power-on / off control system in the above embodiment, includes the following steps:

[0072] In step S401, the server's boot sequence is obtained when the server is powered on.

[0073] The power-on sequence includes starting with AC power-on of the head / tail unit, dominated by the first baseboard management controller, with the tail unit power-on time preceding the head unit power-on time. The entire process involves polling the BOX power-on status, controlling the head unit power-on action under different conditions, and logging the process. Ultimately, the head unit is powered on and normal or abnormal power-on logs are recorded according to the status, achieving orderly coordination between the head unit and the BOX components during the power-on phase.

[0074] In step S402, the head assembly and tail assembly are controlled to perform power-on actions according to the power-on sequence. During the control process, the power-on status of the tail assembly is polled, and the power-on actions of the head assembly and tail assembly are controlled according to the first polling result. The first polling result is the polling result of the power-on status of the tail assembly.

[0075] In this embodiment, controlling the head unit and tail unit to perform power-on actions according to the power-on sequence includes: sending a first polling command to a second baseboard management controller, the first polling command being used to poll the power-on status of the tail unit when the server is powered on; determining whether the tail unit has completed power-on based on the first polling result; if the tail unit has completed power-on, controlling the head unit to perform the power-on action; if the tail unit has not been powered on, sending a power-on command to the second baseboard management controller, the second baseboard management controller controlling the tail unit to perform the power-on action based on the power-on command, and controlling the head unit to perform the power-on action after a preset time after sending the power-on command.

[0076] Understandably, the power-on sequence of the server's head unit and tail unit is coordinated. The entire process uses "whether the tail unit has completed power-on" as the key judgment node: first, a first polling command is sent to the second baseboard management controller to confirm the power-on status of the tail unit; if the tail unit is powered on, the head unit is directly powered on; if the tail unit is not powered on, the second baseboard management controller first powers on the tail unit, and then the head unit is started after a preset time. This method avoids confusion in the power-on sequence of the head and tail units, ensuring the rationality of the power-on process; through the mechanism of "polling confirmation + preset time waiting," it ensures that the tail unit is ready first, and avoids the head unit from getting stuck in infinite waiting due to tail unit abnormalities, thus balancing the reliability and efficiency of power-on; the tail unit's status query and control are realized by the second baseboard management controller, without the need for additional hardware, simplifying the control logic and reducing costs.

[0077] The server power-on / off control method proposed in this application controls the power-on sequence of the front and rear components through preset timing rules and polling feedback. The polling mechanism can capture the power-on status of the rear components in real time, reducing the risk of power-on failure due to rear component abnormalities. Automated timing control and condition triggering reduce manual intervention, avoid system failures caused by misoperation, and improve operational stability.

[0078] The following specific example further illustrates the power-on timing control logic of a server power-on control system. Figure 5 As shown, the specific steps are as follows:

[0079] In step S501, the server AC is powered on, triggering the system startup process. After the BOX baseboard management controller starts up, it immediately performs the BOX power-on operation.

[0080] In step S502, the first substrate management controller starts polling the power-on status of the BOX and continuously monitors the power-on status of the BOX.

[0081] In step S503, it is determined whether the power-on status of the BOX can be obtained by the machine head. If it cannot be obtained after more than 15 minutes, the first baseboard management controller will force the machine head to power on and record the power-on abnormality log.

[0082] In step S504, if the BOX power-on status can be obtained by the chassis head, further determination is made: if the BOX is in a power-off state and the first chassis management controller has not issued a Power on command to the BOX, then the following operations are performed: after waiting for 60 seconds, the first chassis management controller controls the BOX to power on via the command transmission link: for example, ipmitool -I lanplus -H top_bmc_ip -U user -P Password -b 10 -t 50 chassis power on; then, the first chassis management controller returns to the "polling BOX power-on status" process (corresponding to logic S302) and continuously monitors the BOX status; if the BOX is in a power-off state and the first chassis management controller has issued a Power on command to the BOX, then it is determined whether more than 15 minutes have passed since the server AC was powered on (corresponding to logic S303). If less than 15 minutes have passed, the process returns to the "chassis head BMC" process. Poll the BOX power-on status (corresponding to S302 logic), continue waiting for the BOX to complete power-on. If more than 15 minutes have passed, execute "force power-on of the machine head and record the abnormal log", skipping the abnormal state of the BOX and ensuring the machine head starts up. This avoids system stagnation and unresponsiveness, and records logs for fault analysis.

[0083] In step S505, when the power-on status of the BOX can be acquired and it is in the power-on state, a 60-second wait operation is performed, and the first baseboard management controller sends a power-on command to the machine head to reserve preparation time for stable system operation.

[0084] In step S506, the normal power-on process of the machine head is completed, the normal power-on log is recorded, and the machine head is ensured to start stably according to the timing rules.

[0085] The above operations are integrated into the firmware of the first baseboard management controller and run automatically without manual intervention, thus meeting the power-on timing requirements of the machine head and tail.

[0086] The following will combine Figure 6 The server power-on / off control method is described below. This method is applied to the second baseboard management controller of the server power-on / off control system in the above embodiment. The method includes the following steps:

[0087] In step S601, the server shutdown sequence is obtained when the server is shut down.

[0088] The shutdown sequence includes a complete logic that starts with the "head and tail shutdown process" and controls the shutdown of the head and tail components in an orderly manner through command issuance, status polling, and condition judgment. The power-down sequence is such that the tail component shutdown time is later than the head component shutdown time: First, a shutdown command is issued to the head and tail. After the head performs the shutdown action, the second baseboard management controller polls the power-on status of the head. If the head is still in the Poweron state, the tail shutdown command is not effective and the head status log is recorded. If the head is not in the Poweron state, the tail shutdown is completed and a normal log is recorded, ensuring that the shutdown actions of the head and tail are coordinated and the status is controllable.

[0089] In step S602, the head assembly and tail assembly are controlled to perform shutdown actions according to the shutdown timing sequence. During the control process, the power-down status of the head assembly is polled, and the shutdown action of the tail assembly is controlled according to the second polling result. The second polling result is the polling result of the power-down status of the head assembly.

[0090] In this embodiment, controlling the head unit and tail unit to perform a shutdown action according to the shutdown timing includes: sending a second polling command to a first baseboard management controller, the second polling command being used to poll the power-down status of the head unit when the server is shut down; determining whether the head unit has completed power-down based on the second polling result; if the head unit has completed power-down, controlling the tail unit to perform a shutdown action; if the head unit is not powered on, not responding to the shutdown command from the first baseboard management controller; the first baseboard management controller is further configured to, after not obtaining the first polling result, start timing from the time the first polling command is sent, and count whether the first polling result is obtained within a preset time period; if the first polling result is not obtained within the preset time period, the head unit performs a power-on action, wherein if the second baseboard management controller does not provide feedback, the first polling result has not been obtained.

[0091] Understandably, the second baseboard management controller sends a second polling command to the first baseboard management controller to poll the power-down status of the head unit, and then controls the tail unit to shut down based on the result: if the head unit is powered down, the tail unit is shut down; if the head unit is not powered down, the shutdown command is not responded to. Simultaneously, abnormal handling logic is added for the power-on phase: after the first baseboard management controller sends the first polling command, if it does not receive the first polling result within a preset time, it directly controls the head unit to perform the power-on action. This method, by "polling the head unit's power-down status" and adhering to the timing constraint of "tail unit shutdown after head unit shutdown," avoids hardware conflicts or data risks caused by disordered component shutdown order, ensuring the rationality of the shutdown process. The "timeout fallback power-on" logic in the power-on phase can effectively cope with abnormal scenarios such as second baseboard management controller failure and communication link interruption, preventing the head unit from entering an infinite wait due to lack of feedback, thus improving the reliability and fault tolerance of the server power-on.

[0092] The server power-on / off control method proposed in this application, through preset timing and real-time polling, follows the logic of "powering off the first server first and the last server last," avoiding data loss and hardware damage caused by disordered power-off order. Automated timing control and condition triggering reduce manual intervention, avoid system failures caused by misoperation, and improve operational stability.

[0093] The following specific example further illustrates the shutdown timing control logic of the server power-on / off control system. Figure 7 As shown, the specific steps are as follows:

[0094] In step S701, a shutdown command is issued to the head and tail of the machine.

[0095] In step S702, the machine head is immediately shut down.

[0096] In step S703, the second baseboard management controller polls the power-on status of the head unit. If the head unit is powered on, the tail unit shutdown command is not effective. If the BOX needs to be shut down, the BOX needs to be forced to shut down via OEM command and the head unit status log is recorded.

[0097] In step S704, the second baseboard management controller polls the power-on status of the machine head. If the power-on status of the machine head is Power off, the tail section is shut down and a normal shutdown log is recorded.

[0098] Therefore, this embodiment of the application can use preset timing and real-time polling to follow the logic of "shutting down the first machine first and the last machine last," avoiding problems such as data loss and hardware damage caused by disordered shutdown order. Automated timing control and condition triggering reduce manual intervention, avoid system failures caused by misoperation, and improve operational stability.

[0099] Therefore, since the BOX acts as a slave or extension device of the head unit, it must be powered on before it can be recognized by the head unit. When shutting down, the head unit must be shut down first, and then the BOX. Otherwise, data loss or hardware damage to the head unit may occur. Thus, the power-on / off sequence is crucial to system stability and security. However, related technologies lack effective linkage control mechanisms, making it difficult to guarantee correct timing and prone to problems due to improper operation.

[0100] This application solves the problem of power-on and power-off timing control for composite models (including front and rear structures) in related technologies. Through precise process design, it avoids the problem of mutual interference between the power-on and power-off timing of the front and rear of the GPU BOX server model, provides convenience for production operation and maintenance, avoids the risk of power-on and power-off failures caused by illegal operation or abnormal state, and ensures the stability and reliability of the server startup and shutdown process.

[0101] Figure 8 This is a schematic diagram of the structure of a server power-on / off control device provided in an embodiment of this application. This device is applied to the first baseboard management controller of the aforementioned server power-on / off control system, such as... Figure 8 As shown, the server power-on / off control device 80 includes: a first acquisition module 801 and a power-on module 802.

[0102] The first acquisition module 801 is used to acquire the server's power-on sequence when the server is powered on; the power-on module 802 is used to control the head assembly and tail assembly to perform power-on actions according to the power-on sequence. During the control process, the power-on status of the tail assembly is polled, and the power-on actions of the head assembly and tail assembly are controlled according to the first polling result. The first polling result is the polling result of the power-on status of the tail assembly.

[0103] Optionally, step 802 of the power-on module is used to: send a first polling command to the second baseboard management controller, the first polling command being used to poll the power-on status of the tail component when the server is powered on; determine whether the tail component has completed power-on based on the first polling result; if the tail component has completed power-on, control the head component to perform the power-on action; if the tail component has not been powered on, send a power-on command to the second baseboard management controller, the second baseboard management controller controlling the tail component to perform the power-on action based on the power-on command, and controlling the head component to perform the power-on action after a preset time after sending the power-on command.

[0104] Figure 9 Another embodiment of this application provides a schematic diagram of a server power-on / off control device. This device is applied to the first baseboard management controller of the aforementioned server power-on / off control system, such as... Figure 4 As shown, the server power-on / off control device 90 includes: a second acquisition module 901 and a power-off module 902.

[0105] The second acquisition module 901 is used to acquire the shutdown sequence of the server when the server is shut down; the shutdown module 902 is used to control the head assembly and the tail assembly to perform shutdown actions according to the shutdown sequence. During the control process, the power-down status of the head assembly is polled, and the shutdown action of the tail assembly is controlled according to the second polling result. The second polling result is the polling result of the power-down status of the head assembly.

[0106] Optionally, the second acquisition module 901 is further configured to: send a second polling command to the first baseboard management controller, the second polling command being used to poll the power-down status of the head assembly when the server is powered off; determine whether the head assembly has completed power-down based on the second polling result; if the head assembly has completed power-down, control the tail assembly to perform a power-off action; if the head assembly is not powered on, not respond to the power-off command from the first baseboard management controller; the first baseboard management controller is further configured to, after failing to acquire the first polling result, start timing from the time the first polling command is sent, and count whether the first polling result is acquired within a preset time period; if the first polling result is not acquired within the preset time period, the head assembly performs a power-on action, wherein if the second baseboard management controller does not provide feedback, the first polling result has not been acquired.

[0107] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0108] This application also provides an electronic device. Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0109] The memory 1001, the processor 1002, and the computer program stored on the memory 1001 and capable of running on the processor 1002.

[0110] When the processor 1002 executes the program, it implements the key transmission method provided in the above embodiments.

[0111] Furthermore, electronic devices also include:

[0112] Communication interface 1003 is used for communication between memory 1001 and processor 1002.

[0113] The memory 1001 is used to store computer programs that can run on the processor 1002.

[0114] The memory 1001 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0115] If the memory 1001, processor 1002, and communication interface 1003 are implemented independently, then the communication interface 1003, memory 1001, and processor 1002 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 10 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0116] Optionally, in a specific implementation, if the memory 1001, processor 1002, and communication interface 1003 are integrated on a single chip, then the memory 1001, processor 1002, and communication interface 1003 can communicate with each other through an internal interface.

[0117] The processor 1002 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of this application.

[0118] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the embodiments of the thin volume resource management method in the storage system described above when it is run.

[0119] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0120] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in the above-described server power-on / off control method embodiments.

[0121] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps in the above-described embodiments of the server power-on / off control method in the storage system.

[0122] Those skilled in the art will further 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, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. 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 application.

[0123] The foregoing has provided a detailed description of a server power-on / off control system, server, and method provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only intended to aid in understanding the method and core ideas of this application. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A server power on / off control system, characterized in that, include: A head assembly, the head assembly including a first baseboard management controller; A tail assembly, the tail assembly including a second baseboard management controller, wherein the first baseboard management controller and the second baseboard management controller communicate with each other; The first baseboard management controller acquires the power-on sequence of the server when the server is powered on, and controls the head assembly and the tail assembly to perform power-on actions according to the power-on sequence. During the control process, the power-on status of the tail assembly is polled, and the power-on actions of the head assembly and the tail assembly are controlled according to the first polling result. The first polling result is the polling result of the power-on status of the tail assembly. The second baseboard management controller acquires the shutdown sequence of the server when the server is powered off, and controls the head assembly and the tail assembly to perform shutdown actions according to the shutdown sequence. During the control process, the power-down state of the head assembly is polled, and the shutdown action of the tail assembly is controlled according to the second polling result. The second polling result is the polling result of the power-down state of the head assembly.

2. The server power on / off control system according to claim 1, characterized in that, The power-on sequence is that the tail assembly is powered on earlier than the head assembly, and the power-off sequence is that the tail assembly is powered off later than the head assembly.

3. The server power on / off control system according to claim 1, characterized in that, At least one command transmission link is provided between the head assembly and the tail assembly. The first baseboard management and the second baseboard management controller send target commands through the command transmission link. The target commands include at least one of a first polling command, a power-on command, a second polling naming command, and a power-off command. The first polling command is used to poll the power-on status of the tail assembly when the server is powered on, and the second polling command is used to poll the power-off status of the head assembly when the server is powered off.

4. The server power on / off control system according to claim 3, characterized in that, The head assembly includes at least one timing sub-component, and the first baseboard management controller is connected to the timing sub-component. The tail assembly includes at least one switching sub-component, and the second baseboard management controller is connected to the switching sub-component. The command transmission link connects the timing sub-component and the switching sub-component at both ends.

5. The server power on / off control system according to claim 4, characterized in that, An electrical interconnect link is provided between the timing subcomponent and the switching subcomponent, and the first baseboard management controller and the second baseboard management controller communicate with each other through the electrical interconnect link.

6. A server, characterized in that, Includes the server power on / off control system as described in any one of claims 1-5.

7. A server power-on / off control method, characterized in that, The method is applied to the first baseboard management controller of the server power-on / off control system according to any one of claims 1-5, wherein the method includes the following steps: Obtain the server's boot sequence when the server is powered on; The head assembly and the tail assembly are controlled to perform power-on actions according to the power-on sequence. During the control process, the power-on status of the tail assembly is polled, and the power-on actions of the head assembly and the tail assembly are controlled according to the first polling result. The first polling result is the polling result of the power-on status of the tail assembly.

8. The server power-on / off control method according to claim 7, characterized in that, The step of controlling the head assembly and the tail assembly to perform the power-on action according to the power-on sequence includes: Send a first polling command to the second baseboard management controller. The first polling command is used to poll the power-on status of the tail assembly when the server is powered on. Determine whether the tail assembly has been powered on based on the results of the first round of polling; If the tail assembly is powered on, then the head assembly is controlled to perform the power-on action; If the tail assembly is not powered on, a power-on command is sent to the second baseboard management controller. The second baseboard management controller controls the tail assembly to perform a power-on action based on the power-on command. After a preset time after sending the power-on command, the head assembly is controlled to perform the power-on action.

9. A server power-on / off control method, characterized in that, The method is applied to the second baseboard management controller of the server power-on / off control system according to any one of claims 1-5, wherein the method includes the following steps: Obtain the shutdown sequence of the server when it is shut down; The head assembly and the tail assembly are controlled to perform shutdown actions according to the shutdown timing sequence. During the control process, the power-down status of the head assembly is polled, and the shutdown action of the tail assembly is controlled according to the second polling result. The second polling result is the polling result of the power-down status of the head assembly.

10. The server power-on / off control method according to claim 9, characterized in that, The step of controlling the head assembly and the tail assembly to perform a shutdown action according to the shutdown timing includes: A second polling command is sent to the first baseboard management controller. The second polling command is used to poll the power-down status of the head assembly when the server is powered off. Determine whether the head unit has been powered down based on the results of the second round of queries; If the head assembly is powered down, then the tail assembly is controlled to perform the shutdown action; If the head assembly is not powered on, it will not respond to the power-off command from the first baseboard management controller; The first baseboard management controller is further configured to, after failing to obtain the first polling result, start timing from the time the first polling instruction is sent, and count whether the first polling result is obtained within a preset time period. If the first polling result is not obtained within the preset time period, the head unit performs a power-on action. If the second baseboard management controller does not provide feedback, the first polling result has not been obtained.

Images