A server management system

Abstract

The application discloses a server management system, and relates to the technical field of computers, which comprises a target management substrate, a signal repeater and a firmware storage, wherein the target management substrate is connected with the signal repeater and the firmware storage respectively, the signal repeater is further connected with the firmware storage, the server management system is connected with a host management board through the target management substrate, the firmware storage is used for storing repeater firmware, the signal repeater is used for transmitting firmware loading parameters to the target management substrate in the process of loading the repeater firmware, the target management substrate is used for receiving and responding to a state detection request for requesting to detect the running state of the signal repeater sent by the host management board, detecting the access state of the signal repeater according to the firmware loading parameters, accessing the running state of the signal repeater according to the access state, obtaining the target running state of the signal repeater, and transmitting the target running state to the host management board.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of computer, and particularly, relate to a server management system. BACKGROUND

[0002] In the related art, all devices (including signal relays) on a server management system are collectively managed by a host management board through a shared bus. This means that all management instructions and data must be transmitted through the only bus. This centralized management scheme faces multiple technical challenges: first, the shared bus is overloaded when facing multiple devices, resulting in low communication efficiency and signal problems; second, this management method is highly dependent on the host, and once the host system fails, the underlying devices cannot be effectively managed; the most core problem is that the host management board in the related art usually directly accesses the signal relays on the server management system, and at this time some signal relays may be in the firmware loading period, which is extremely easy to cause firmware loading failure or even device hanging.

[0003] For the technical problems such as high access failure rate of the server management system in the related art, no effective solutions have been proposed. SUMMARY

[0004] Embodiments of the present application provide a server management system to at least solve the technical problem of high access failure rate of the server management system in the related art.

[0005] According to an embodiment of the present application, a server management system is provided, comprising: a target management substrate, a signal relay and a firmware storage, the target management substrate is connected with the signal relay and the firmware storage respectively, the signal relay is further connected with the firmware storage, and the server management system is connected with a host management board through the target management substrate, wherein the firmware storage is configured to store a relay firmware.

[0006] The signal relay is configured to transmit a firmware loading parameter to the target management substrate in a process of loading the relay firmware, wherein the firmware loading parameter is configured to indicate a loading condition of the relay firmware.

[0007] The target management substrate is configured to receive and respond to a state detection request sent by the host management board, the state detection request is configured to request to detect a running state of the signal relay, to detect an access state of the signal relay according to the firmware loading parameter, to access the running state of the signal relay according to the access state, to obtain a target running state of the signal relay, and to transmit the target running state to the host management board, wherein the access state is configured to indicate whether the signal relay can be accessed.

[0008] A server management system is provided in the present application, a target management substrate as a local management core of the server management system, independent of a host management board, directly connected with a signal repeater and a firmware storage. During loading of the firmware by the signal repeater, it transmits firmware loading parameters indicating the loading situation to the target management substrate. The target management substrate uses the parameters to determine whether the signal repeater is currently allowed to be accessed, and decides whether to perform a state access operation according to the determination result. The server management system provided in the present application can solve the problem of high access failure rate of the server management system, because it fundamentally solves the core pain point in the related art: device hang-up caused by false access. By performing intelligent determination before access, i.e. confirming that the signal repeater is in an allowed access state according to the firmware loading parameters before access, the present solution perfectly avoids the conflict caused by access during the sensitive period of firmware loading, thereby greatly reducing the failure rate of access and improving the reliability and stability of the whole system. Therefore, the technical problem of high access failure rate of the server management system in the related art is solved, and the technical effect of reducing the access failure rate of the server management system is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0010] Figure 1 is a schematic diagram of a server management system according to an embodiment of the present application;

[0011] Figure 2 is a schematic diagram of a target management substrate according to an embodiment of the present application;

[0012] Figure 3 is a connection diagram of a substrate controller, a state controller and a signal repeater according to an embodiment of the present application;

[0013] Figure 4 is a schematic diagram of an isolator according to an embodiment of the present application;

[0014] Figure 5 is a schematic diagram of a device topology of a server management system according to an embodiment of the present application. DETAILED DESCRIPTION

[0015] With reference to the drawings and embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0016] It should be noted that, in the description of the present application, the terms "comprise", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment comprising a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. The terms "first", "second" and the like in the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence.

[0017] The technical terms and concepts involved in the present application are as follows:

[0018] OAM (OAM Open Accelerator Module), open accelerator module;

[0019] UBB (Universal Baseboard), a board card that can carry a GPU;

[0020] Retimer: signal re-timer, i.e. signal repeater in the present application, used for high-speed digital link, reshaping and regenerating signals to ensure data transmission.

[0021] FPGA (Field-Programmable Gate Array): field-programmable gate array, a logic chip that can be programmed by users.

[0022] In order for those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0023] In the present embodiment, a server management system is provided, Figure 1 is a schematic diagram of a server management system according to an embodiment of the present application, as Figure 1 shown, the server management system comprises a target management baseboard, a signal repeater and a firmware storage, the target management baseboard is connected with the signal repeater and the firmware storage respectively, the signal repeater is further connected with the firmware storage, and the server management system is connected with a host management board through the target management baseboard, wherein the firmware storage is used to store repeater firmware.

[0024] The signal relayer is configured to transmit a firmware loading parameter to the target management board during loading of the relayer firmware, wherein the firmware loading parameter is used to indicate a loading status of the relayer firmware.

[0025] The target management board is configured to receive and respond to a state detection request sent by the host management board, wherein the state detection request is used to request detection of a running state of the signal relayer, to detect an access state of the signal relayer according to the firmware loading parameter, to access the running state of the signal relayer according to the access state, to obtain a target running state of the signal relayer, and to transmit the target running state to the host management board, wherein the access state is used to indicate whether the signal relayer can be accessed.

[0026] The server management system proposed in the present application has a target management board as a local management core of the server management system, which is independent of the host management board and directly connected with the signal relayer and the firmware storage. During loading of the firmware by the signal relayer, the signal relayer transmits a firmware loading parameter indicating a loading status to the target management board. The target management board uses the parameter to determine whether the signal relayer is currently accessible, and decides whether to perform a state access operation according to the determination result. The server management system proposed in the present application can solve the problem of a high access failure rate of the server management system, because it fundamentally solves the core pain point in the related art: device hang-up caused by false access. By performing intelligent determination before access, i.e., confirming that the signal relayer is in an accessible state according to the firmware loading parameter before access, the present application perfectly avoids the conflict caused by access during the sensitive period of firmware loading, thereby greatly reducing the failure rate of access and improving the reliability and stability of the entire system. Therefore, the present application can solve the technical problem of a high access failure rate of the server management system in the related art, and achieve the technical effect of reducing the access failure rate of the server management system.

[0027] Optionally, in this embodiment, the server management system is equipped with a signal repeater and a firmware storage. The firmware storage pre-stores the operation program of the repeater. The server management system is also configured with a target management baseboard, which has a separate connection with the signal repeater and the firmware storage. When a host management board (i.e., the management board on the server host) needs to obtain the operation status of the signal repeater, it sends a status detection request to the target management baseboard on the baseboard. The signal repeater automatically loads the operation program from the firmware storage when it is powered on. During the loading process, the signal repeater continuously sends a firmware loading parameter to the target management baseboard. The parameter is a binary value, for example, "10", indicating that the firmware is being loaded. When the loading is completed, the parameter becomes "11". After receiving the request from the host management board, the target management baseboard first reads the firmware loading parameter to determine whether the signal repeater is currently in an accessible access state. If the parameter is "11", it determines that the signal repeater is in an allowed access state; if the parameter is "10", it determines that it is in a prohibited access state. Only when it is determined to be in the allowed access state, the target management baseboard will perform a status access operation on the signal repeater, such as reading its temperature or error information, and finally aggregate the information into a target operation status and return it to the host management board. If it is in the prohibited access state, the target management baseboard will wait for the firmware loading to be completed.

[0028] In the formula, the target management baseboard is a separate management chip on the server management system, responsible for localized management. The signal repeater is a chip for enhancing high-speed signals. The firmware storage is a chip that stores the operation program of the signal repeater. The host management board (also known as HOST management board (the HOST BMC (Host Baseboard Management Controller, central management unit of the entire server) is deployed on the HOST management board)) is the central management unit of the server. The firmware loading parameter is a signal sent by the signal repeater to indicate the progress or status of firmware loading. The access state is the state determined by the target management baseboard based on the firmware loading parameter, indicating whether the signal repeater is currently allowed to be accessed. The status access operation is the management access performed by the target management baseboard on the signal repeater.

[0029] Optionally, in this embodiment, the target management baseboard can be but is not limited to a UBB baseboard.

[0030] In the formula, the UBB baseboard is a baseboard that can be equipped with a GPU module, and forms a complete GPU platform through the equipped GPU to realize direct connection of the GPU acceleration module, provide a channel for high-speed data transmission and exchange, and help realize multi-GPU collaboration to cope with computationally intensive tasks in the AI field such as image recognition and machine learning.

[0031] Optionally, in some embodiments, the UBB substrate can also provide power management, heat dissipation and other support for modules such as GPU, to ensure stable operation of the related modules.

[0032] Optionally, in some embodiments, the UBB substrate can also support multiple GPUs (such as 8 OAM modules working in various wiring and interconnection topologies).

[0033] Through the present embodiment, the following can be achieved: (1) hierarchical management, reducing the burden of the host bus: the management work of the bottom layer is distributed to the target management substrate, so that the host management board does not need to communicate directly with the bottom layer device, improving the management efficiency and scalability of the entire system. (2) Intelligent access, avoiding hanging: by judging the access state before access, the problem of device hanging caused by misaccess is fundamentally solved, greatly reducing the access failure rate. (3) Fault isolation: even if the host system fails, the target management substrate can still manage the signal repeater independently, realizing the isolation of the fault domain and facilitating fault troubleshooting.

[0034] As an optional solution, Figure 2 is a schematic diagram of a target management substrate according to an embodiment of the present application, as Figure 2 shown, the target management substrate includes a substrate controller and a state controller, the substrate controller and the state controller are connected through a first communication link, and the state controller is also connected with the signal repeater through a second communication link, wherein the signal repeater is used to send the firmware loading parameters to the state controller through the second communication link in the process of loading the repeater firmware.

[0035] The state controller is used to transmit the firmware loading parameters to the substrate controller through the first communication link after receiving the firmware loading parameters.

[0036] The substrate controller is used to detect the access state of the signal repeater according to the received firmware loading parameters.

[0037] Optionally, in the present embodiment, Figure 3 is a connection diagram of a substrate controller, a state controller and a signal repeater according to an embodiment of the present application, as Figure 3As shown, on the server management system, the target management baseboard (which can be but is not limited to a UBB baseboard) is not a single chip, but is cooperated by a baseboard controller (for example, a UBB BMC (UBB Baseboard Management Controller)) and a state controller (for example, a CPLD), and the baseboard controller is deployed on the UUB management board. The baseboard controller and the state controller are connected through a first communication link (for example, an I2C (Inter-Integrated Circuit, a kind of serial communication bus) / I3C (Improved Inter-Integrated Circuit, an improved version of I2C, a kind of serial communication protocol) bus). The signal relay (RETIMER) is connected with the state controller through a second communication link (for example, a GPIO (General-purpose input / output, a kind of programmable digital pin) signal line). When the signal relay starts to load the firmware from the firmware storage, it will send a high-level signal to the state controller through its GPIO pin as a firmware loading parameter, and this signal indicates that the firmware is being loaded. After the state controller receives this signal, it will immediately convert this information into a digital signal through the first communication link and transmit it to the baseboard controller. After the baseboard controller receives this information, it will immediately update its internal database to indicate that the signal relay is loading firmware, so as to make subsequent access state judgment. This layered communication mode ensures that the firmware loading state can be reliably transmitted from the signal relay to the baseboard controller without passing through the I2C bus which may be blocked or incompatible.

[0038] Among them, the baseboard controller is the part of the target management baseboard responsible for high-level management logic. The state controller is the part of the target management baseboard responsible for low-level signal conversion and coordination. The first communication link is the communication channel between the baseboard controller and the state controller (such as I2C). The second communication link is the communication channel between the signal relay and the state controller (such as GPIO), which is usually used to transmit simple status signals.

[0039] Through this embodiment, the following can be achieved: (1) Enhance communication reliability: solve the problem that some signal relays do not support I2C access during firmware loading, and ensure that firmware loading state information can be reliably transmitted by using a simpler GPIO signal as a second communication link. (2) Division of labor: the responsibilities of the baseboard controller and the state controller are clear, one is responsible for high-level management logic, and the other is responsible for low-level signal coordination, which improves the efficiency and stability of the whole management system.

[0040] As an optional solution, the baseboard controller is configured to:

[0041] detecting a loading progress of the relay firmware according to the firmware loading parameter;

[0042] in a case where the loading progress is greater than or equal to a preset loading progress, determining an access state of the signal relay as an allowed access state, wherein the allowed access state is used to indicate that the signal relay is allowed to be accessed;

[0043] in a case where the loading progress is less than the preset loading progress, determining the access state of the signal relay as a forbidden access state, wherein the forbidden access state is used to indicate that the signal relay is forbidden to be accessed.

[0044] Optionally, in the embodiment, after the substrate controller receives the firmware loading parameter from the state controller, the substrate controller starts to execute the judgment logic. It is assumed that the substrate controller presets a firmware loading progress threshold, for example, 95%. The "firmware loading parameter" can be a percentage value or a state code indicating the completion degree. When the substrate controller detects that the firmware loading parameter is 80% (less than 95%), it marks the access state of the signal relay as the forbidden access state. If the firmware loading parameter is received again, and the loading progress is displayed as 98% (greater than 95%), the substrate controller immediately updates the access state to the allowed access state. This logic enables the substrate controller to accurately grasp the access permission of each signal relay.

[0045] The loading progress is the completion degree of the firmware loading. The preset loading progress is a judgment threshold set in the substrate controller. The allowed access state is a state in which the signal relay can be safely accessed. The forbidden access state is a state in which the signal relay should not be accessed.

[0046] The embodiment can achieve the following: (1) fine access control: fine access control based on the firmware loading progress is achieved, and access to the signal relay that is performing a critical operation is avoided, further reducing the failure rate. (2) access opportunity judgment: the substrate controller can intelligently determine the best access opportunity, ensuring the successful execution of the management operation.

[0047] As an optional solution, the substrate controller is configured to:

[0048] in a case where the access state is the allowed access state, accessing the running state of the signal relay, wherein the allowed access state is used to indicate that the signal relay is allowed to be accessed;

[0049] In the case that the access state is the forbidden access state, a forbidden access duration of the signal repeater is acquired; and after the forbidden access duration, an operation state of the signal repeater is accessed, wherein the allowed access state is used to indicate that the signal repeater is allowed to be accessed, and the forbidden access duration is a duration in which the signal repeater still needs to be in the forbidden access state.

[0050] Optionally, in the embodiment, it is assumed that the host management board sends a state detection request to the baseboard controller. The baseboard controller first detects its internal database and finds that the access state of a certain signal repeater is the forbidden access state. Instead of immediately performing the state access operation, the baseboard controller calculates a forbidden access duration, for example, 2 seconds. The baseboard controller waits for 2 seconds and then attempts to access the signal repeater again after 2 seconds. If the access state of the signal repeater is updated to the allowed access state at this time, the baseboard controller normally performs the state access operation and collects the operation state. If the baseboard controller finds that another signal repeater is currently in the allowed access state, it will not wait and immediately perform the state access operation. This process ensures that even if an obstacle is encountered when initially attempting to access, the system can ultimately complete the task through intelligent waiting mechanism.

[0051] The forbidden access duration refers to a duration in which the signal repeater needs to continue to be in the forbidden access state. The duration is used to indicate how long the baseboard controller needs to wait before attempting to access again.

[0052] Through the embodiment, the following can be achieved: (1) The robustness of the management system is improved. Even if a device is temporarily inaccessible, the system can intelligently wait and ultimately complete the task, rather than immediately reporting an error, thereby improving the fault tolerance and stability of the system. (2) The task completion rate is ensured. It is ensured that the state detection request issued by the host management board can be ultimately executed, thereby improving the overall task completion rate.

[0053] As an optional solution, the baseboard controller is configured to:

[0054] According to the firmware loading parameter, a loading progress of the repeater firmware is detected;

[0055] According to a preset loading progress and the loading progress, a remaining loading progress is generated, wherein the remaining loading progress is a progress that the signal repeater still needs to load before reaching the preset loading progress;

[0056] According to the remaining loading progress and a loading speed, a remaining loading duration of the signal repeater is generated, wherein the loading speed is a speed at which the signal repeater loads the repeater firmware, and the remaining loading duration is a duration in which the signal repeater needs to load to reach the preset loading progress.

[0057] determining the remaining loading duration as the access-prohibited duration.

[0058] Optionally, in this embodiment, when the substrate controller detects that a certain signal repeater is in an access-prohibited state, it needs to calculate the access-prohibited duration. Assuming that the preset loading progress of the firmware is 95%, the currently detected firmware loading progress is 80%, and the loading speed of the firmware is 10% / second. The substrate controller first calculates the remaining loading progress: 95%-80%=15%. Then, it calculates the remaining loading duration according to this remaining progress and the loading speed: 15% ÷ 10% / second = 1.5 seconds. Finally, the substrate controller determines this 1.5-second remaining loading duration as the access-prohibited duration and starts waiting. This dynamic calculation ensures that the waiting time is optimal, neither too long to reduce efficiency nor too short to cause access failure.

[0059] wherein the remaining loading progress refers to the progress that the signal repeater needs to load before reaching the preset loading progress, the loading speed refers to the speed at which the signal repeater loads the firmware, and the remaining loading duration refers to the time that the signal repeater needs to reach the preset loading progress.

[0060] This embodiment can achieve the following: (1) Accurate control of waiting time: The system can dynamically adjust the waiting duration according to the actual situation of firmware loading, avoiding unnecessary long waiting time and improving management efficiency. (2) Optimization of resource utilization: The system can more reasonably arrange resources and tasks, and will not waste computing power during the waiting process, making the overall operation more efficient.

[0061] As an optional solution, the operation state of the signal repeater includes:

[0062] The substrate controller collects at least one of the following information of the signal repeater: running information, temperature information, and error information, wherein the running information is used to indicate the functional running condition of the signal repeater, the temperature information is used to indicate the running temperature of the signal repeater, and the error information is used to indicate the error condition of the signal repeater during operation.

[0063] According to at least one of the running information, the temperature information, and the error information, the operation state of the signal repeater is generated to obtain the target operation state.

[0064] Optionally, in this embodiment, the substrate controller will start a status access operation after obtaining the access right. This operation specifically includes: reading a series of data from the registers of the signal repeater through the I2C bus. For example, it can read the running information indicating whether the current function of the signal repeater is running normally; read the temperature information indicating its working temperature; and read any error information occurring during operation. The substrate controller integrates and analyzes these data, and then generates a comprehensive running status report. For example, the report may include: "function normal, temperature 55°C, no error". This final report is the target running status, which will be returned to the host management board.

[0065] Among them, the running information, the temperature information and the error information are three key data dimensions constituting the running status. The target running status is the final running report obtained by the substrate controller after comprehensively analyzing multiple information.

[0066] Through this embodiment, the following can be achieved: (1) Comprehensive status monitoring: It ensures that the host management board can obtain comprehensive and detailed running data about the signal repeater, not just a simple "normal" or "abnormal" judgment. (2) Efficient fault diagnosis: It provides detailed error information, making remote fault diagnosis possible without physical disassembly, saving a lot of time and maintenance cost.

[0067] As an optional solution, the target management substrate is further configured to:

[0068] perform a firmware adjustment operation on the repeater firmware stored in the firmware memory;

[0069] control the signal repeater to prohibit access to the firmware memory during execution of the firmware adjustment operation;

[0070] control the signal repeater to resume access to the firmware memory after the execution of the firmware adjustment operation is completed.

[0071] Optionally, in this embodiment, the target management substrate can not only manage the signal repeater, but also has the ability to independently operate the firmware memory. When it is necessary to upgrade or downgrade the firmware of a signal repeater, the target management substrate can initiate a firmware adjustment operation to directly write new firmware data to the firmware memory without going through the signal repeater. During the execution of this firmware adjustment operation, the target management substrate will issue an instruction to the signal repeater to prohibit access to the firmware memory. When the firmware adjustment operation is completed, the target management substrate will lift the prohibition and restore the access right of the signal repeater to the firmware memory. This design ensures the reliability of firmware update, and even if the signal repeater itself fails, it can still be repaired remotely through this channel.

[0072] The firmware adjustment operation is a write, update, or the like of the firmware memory by the target management substrate. The control signal relay prohibits access to the firmware memory, which means that the access of the signal relay to the firmware memory is limited at the software level or the hardware level.

[0073] The present embodiment can achieve the following: (1) out-of-band firmware update: firmware update is achieved without relying on the state of the signal relay, and even if the device is abnormal, remote repair can be performed. (2) Ensure update safety: avoid damage to the firmware memory caused by the signal relay during the firmware update process, and ensure the success rate of the update operation.

[0074] As an optional solution, Figure 4 is a schematic diagram of an isolator according to an embodiment of the present application, as Figure 4 shown, an isolator is also deployed on the communication link between the signal relay and the firmware memory, and the target management substrate is connected to the isolator, wherein the isolator is used to control the communication link to be disconnected or conductive.

[0075] The target management substrate is used to control the isolator to disconnect the communication link between the signal relay and the firmware memory during the execution of the firmware adjustment operation, and control the isolator to conduct the communication link between the signal relay and the firmware memory after the execution of the firmware adjustment operation is completed.

[0076] Optionally, in the present embodiment, in order to more reliably achieve exclusive access to the firmware memory, the present solution deploys an isolator (for example, a 9306 chip) on the communication link between the signal relay and the firmware memory. This isolator is a controllable physical switch, which is controlled by the target management substrate. When the target management substrate executes the firmware adjustment operation, it sends a signal to the isolator to control it to disconnect the communication link between the signal relay and the firmware memory. After the firmware adjustment operation is completed, the target management substrate sends a signal to the isolator again to control it to conduct the communication link, and restore normal data transmission. This design guarantees that the signal relay cannot perform any operation on the firmware memory during the firmware adjustment, thereby completely eliminating the access conflict.

[0077] The isolator is a hardware chip, and its function is to disconnect or conduct the communication link under software control. Disconnecting or conducting the link means cutting off or restoring the physical connection between two devices through hardware control.

[0078] Through the embodiment, the following can be achieved: (1) physical layer isolation: the access conflict during the firmware update is fundamentally eliminated, the highest level of security is provided, and the success rate of the firmware update is ensured; and (2) no fear of software failure: the hardware isolation mechanism is not affected by software failure, and even if the internal software of the signal repeater has a problem, the firmware update cannot be disturbed.

[0079] As an optional solution, the target management substrate includes a substrate controller and a state controller, the substrate controller is connected with the state controller, and the state controller is further connected with the isolator, wherein the state controller is configured to send a link disconnection signal to the isolator during execution of the firmware adjustment operation, and send a link connection signal to the isolator after the execution of the firmware adjustment operation ends.

[0080] The isolator is configured to control the communication link to be disconnected when the link disconnection signal is received, and control the communication link to be connected when the link connection signal is received.

[0081] Optionally, in the embodiment, when the target management substrate controls the isolator, instead of direct communication, the state controller is used as an intermediate layer. The substrate controller (as part of the target management substrate) is connected with the state controller. When it is necessary to disconnect the link, the substrate controller sends an instruction to the state controller. After receiving the instruction, the state controller sends a link disconnection signal (for example, the control pin of the 9306 chip is controlled to be low) to the isolator. After receiving the signal, the isolator immediately disconnects the link. When it is necessary to connect the link, the substrate controller sends an instruction to the state controller again, and the state controller sends a link connection signal (for example, the control pin is controlled to be high) to the isolator to restore the communication. This process ensures that the logic of the underlying hardware control is concentrated in the state controller, so that the substrate controller can focus on high-level management tasks.

[0082] The link disconnection signal and the link connection signal are control signals sent by the state controller to the isolator, and are used to control the physical state of the isolator.

[0083] Through the embodiment, the following can be achieved: (1) separation of duties and simplified design: the underlying hardware control (performed by the state controller) and the high-level management logic (performed by the substrate controller) are separated, and the design of the substrate controller is simplified; and (2) improved system stability: it is ensured that the underlying physical control logic is stable and controllable, and is not easily affected by changes in the high-level management logic.

[0084] As an optional solution, Figure 5 is a schematic diagram of a device topology of a server management system according to an embodiment of the application, like Figure 5As shown, the signal repeaters in the server management system are multiple (Retimer0 to Retimer7 (part of which is not shown in the figure)), Retimer0 to Retimer7 correspond to a firmware memory (EEPROM (Electrically Erasable Programmable Read-Only Memory, used to store the firmware of Retimer.)) respectively, each Retimer and EEPROM are connected through an isolator (9306 (9306 I2C busswitch, an I2C bus switch used for physical isolation in this scheme.)) and a state controller (which can be but not limited to a CPLD (Complex Programmable Logic Device, used to control the logic of the underlying hardware in this scheme), or the I2C Switch shown in the figure), each of the signal repeaters is connected with the target management substrate, wherein the target management substrate is used to access multiple reference signal repeaters in the multiple signal repeaters in parallel, the access state of which is the access allowed state, to obtain the reference signal repeater and the reference running state with a corresponding relationship; the reference signal repeater and the reference running state with a corresponding relationship are transmitted to the host management board, wherein the access allowed state is used to indicate that the signal repeater allows to be accessed.

[0085] Optionally, in this embodiment, in a large server management system with multiple signal repeaters, the target management substrate will manage all signal repeaters at the same time. When the host management board requests to obtain the state of all signal repeaters, the target management substrate will first identify which signal repeaters are currently in the access allowed state. Then, it will send a state detection request to these "access allowed" signal repeaters in parallel. For example, if there are 8 signal repeaters, 5 of which are in the access allowed state, the target management substrate will initiate access to these 5 signal repeaters at the same time, rather than polling one by one. After all the requests are responded, it will transmit the reference signal repeater and the reference running state with a corresponding relationship (for example, Retimer 3 is normal, Retimer 5 has high temperature) to the host management board.

[0086] Wherein, parallel access means that the target management substrate initiates access operation to multiple signal repeaters at the same time. The reference signal repeater is the signal repeater that is currently in the access allowed state and is accessed. The reference running state is the running state of these accessed signal repeaters.

[0087] Through the embodiment, the following can be achieved: (1) large-scale parallel management: parallel management of multiple signal repeaters is achieved, the polling time is greatly shortened, and the management efficiency of the large-scale server management system is significantly improved; (2) fast response: the host management board can obtain the running states of all the signal repeaters more quickly, and the overall response speed of the system is improved.

[0088] The server management system provided in the application is described in detail. The principles and implementation modes of the application are described by applying specific examples. The above description of the embodiments is only applicable to help understand the method of the application and its core idea. It should be pointed out that, for those skilled in the art, without departing from the principles of the application, the application can be improved and modified in several ways, and these improvements and modifications also fall within the protection scope of the claims of the application.

Claims

1. A server management system, characterized by, include: The system comprises a target management board, a signal repeater, and a firmware memory. The target management board is connected to both the signal repeater and the firmware memory. The signal repeater is also connected to the firmware memory. The server management system is connected to the host management board via the target management board. The firmware memory is used to store the repeater firmware; The signal repeater is used to transmit firmware loading parameters to the target management board during the loading of the repeater firmware, wherein the firmware loading parameters are used to indicate the loading status of the repeater firmware. The target management board is configured to receive and respond to a status detection request sent by the host management board for requesting detection of the operating status of the signal repeater, detect the access status of the signal repeater according to the firmware loading parameters, and access the operating status of the signal repeater according to the access status to obtain the target operating status of the signal repeater; and transmit the target operating status to the host management board, wherein the access status is used to indicate whether the signal repeater can be accessed; The target management board is further configured to: perform a firmware adjustment operation on the repeater firmware stored in the firmware memory; control the signal repeater to prohibit access to the firmware memory during the execution of the firmware adjustment operation; and control the signal repeater to restore access to the firmware memory after the execution of the firmware adjustment operation is completed.

2. The server management system according to claim 1, characterized in that, The target management board includes a board controller and a status controller, which are connected via a first communication link. The status controller is also connected to the signal repeater via a second communication link. The signal repeater is used to send the firmware loading parameters to the status controller via the second communication link during the loading of the repeater firmware. The status controller is used to transmit the firmware loading parameters to the baseboard controller via the first communication link after receiving the firmware loading parameters. The baseboard controller is used to detect the access status of the signal repeater based on the received firmware loading parameters.

3. The server management system according to claim 2, characterized in that, The baseboard controller is configured as follows: The loading progress of the repeater firmware is detected based on the firmware loading parameters. If the loading progress is detected to be greater than or equal to the preset loading progress, the access status of the signal repeater is determined to be an access-allowed status, wherein the access-allowed status is used to indicate that the signal repeater is allowed to be accessed; If the loading progress is detected to be less than the preset loading progress, the access status of the signal repeater is determined to be a prohibited access status, wherein the prohibited access status is used to indicate that the signal repeater is prohibited from being accessed.

4. The server management system according to claim 3, characterized in that, The baseboard controller is configured as follows: When the access state is an allowed access state, access the operating state of the signal repeater, wherein the allowed access state is used to indicate that the signal repeater is allowed to be accessed; When the access status is a prohibited access status, the prohibited access duration of the signal repeater is obtained; and after the prohibited access duration, the operating status of the signal repeater is accessed, wherein the allowed access status is used to indicate that the signal repeater is allowed to be accessed, and the prohibited access duration is the duration for which the signal repeater needs to remain in the prohibited access status.

5. The server management system according to claim 4, characterized in that, The baseboard controller is configured as follows: The loading progress of the repeater firmware is detected based on the firmware loading parameters. The remaining loading progress is generated based on the preset loading progress and the loading progress, wherein the remaining loading progress is the progress that the signal repeater still needs to load before reaching the preset loading progress; The remaining loading time of the signal repeater is generated based on the remaining loading progress and loading speed, wherein the loading speed is the speed at which the signal repeater loads the repeater firmware, and the remaining loading time is the time that the signal repeater still needs to load to reach the preset loading progress; The remaining loading time is determined as the access-restricted time.

6. The server management system according to claim 4, characterized in that, The access to the operating status of the signal repeater includes: The baseboard controller collects at least one of the following information from the signal repeater: operating information, temperature information, and error information, wherein the operating information is used to indicate the functional operation status of the signal repeater, the temperature information is used to indicate the operating temperature of the signal repeater, and the error information is used to indicate the error status of the signal repeater during operation. Based on the operational information, at least one of the temperature information and the error information is used to generate the operational status of the signal repeater, thereby obtaining the target operational status.

7. The server management system according to claim 1, characterized in that, An isolator is also deployed on the communication link between the signal repeater and the firmware memory, and the target management board is connected to the isolator. The isolator is used to control the disconnection or connection of the communication link it is connected to; The target management board is used to control the isolator to disconnect the communication link between the signal repeater and the firmware memory during the execution of the firmware adjustment operation; and to control the isolator to connect the communication link between the signal repeater and the firmware memory after the execution of the firmware adjustment operation is completed.

8. The server management system according to claim 7, characterized in that, The target management board includes a board controller and a status controller. The board controller is connected to the status controller, and the status controller is also connected to the isolator. The state controller is configured to send a link disconnect signal to the isolator during the execution of the firmware adjustment operation, and send a link reconnect signal to the isolator after the execution of the firmware adjustment operation is completed. The isolator is used to disconnect the communication link when it receives the link disconnection signal, and to connect the communication link when it receives the link connection signal.

9. The server management system according to claim 1, characterized in that, The server management system contains multiple signal repeaters, each of which is connected to the target management baseboard. The target management board is used to access multiple reference signal repeaters in parallel, where the access status is "allowed access," to obtain a corresponding relationship between the reference signal repeaters and a reference operating status; the corresponding relationship between the reference signal repeaters and the reference operating status is transmitted to the host management board, wherein the "allowed access" status is used to indicate that the signal repeater is allowed to be accessed.

Images