Server and control method

Abstract

The application provides a server and a control method, which can be applied to the technical field of hardware. The server comprises a baseboard management controller, a backup management controller, a component and a selection circuit, wherein the selection circuit is connected with the baseboard management controller, the backup management controller and the component, the backup management controller is connected with the baseboard management controller, the baseboard management controller controls the selection circuit to establish a first bus link between the baseboard management controller and the component through a first enabling signal, and performs platform management on the component through the first bus link; and the backup management controller controls the selection circuit to establish a second bus link between the backup management controller and the component through a second enabling signal in response to detecting that the baseboard management controller fails, and takes over the platform management on the component through the second bus link.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hardware, and more particularly to a server and a control method. BACKGROUND

[0002] I2C (Inter-Integrated Circuit) is a kind of serial communication bus, with extremely simple hardware design and flexible multi-master multi-slave topology, becoming the core solution of low-speed communication inside the server, indispensable in the scene of state monitoring, configuration storage, sensor reading, etc. When communication connection fails, it will affect the normal work of the server. SUMMARY

[0003] In view of the above problems, the present application provides a server and a control method.

[0004] According to a first aspect of the present application, a server is provided, comprising: a baseboard management controller, a backup management controller, a component, and a selection circuit, wherein the selection circuit is connected with the baseboard management controller, the backup management controller and the component, the backup management controller is connected with the baseboard management controller, the selection circuit is used to establish a first bus link between the baseboard management controller and the component under the control of the baseboard management controller, and establish a second bus link between the backup management controller and the component under the control of the backup management controller; the baseboard management controller is used to control the selection circuit to establish the first bus link between the baseboard management controller and the component through a first enable signal, and perform platform management on the component through the first bus link; the backup management controller is used to, in response to detecting that the baseboard management controller fails, control the selection circuit to establish the second bus link between the backup management controller and the component through a second enable signal, and take over the platform management on the component through the second bus link.

[0005] A second aspect of the present application provides a control method of the above-mentioned server, the control method comprising: the baseboard management controller controls the selection circuit to establish a first bus link between the baseboard management controller and the component through a first enable signal, and performs platform management on the component through the first bus link; the backup management controller, in response to detecting that the baseboard management controller fails, controls the selection circuit to establish a second bus link between the backup management controller and the component through a second enable signal, and takes over the platform management on the component through the second bus link.

[0006] According to the embodiments of the present application, a dual-host management mode for the component is established by using the selection circuit. The selection circuit can be controlled by the baseboard management controller to communicatively connect the baseboard management controller with the component, so that the baseboard management controller can perform platform management on the component to health-manage the running state of the server. The selection circuit can also be controlled by the standby management controller to communicatively connect the standby management controller with the component, so that the standby management controller can take over the platform management on the component when detecting that the baseboard management controller has failed, and health-manage the running state of the server by the standby management controller in the case that the baseboard management controller has failed, so that the server can maintain a normal working state or can recover to a normal working state when the communication connection fails. BRIEF DESCRIPTION OF DRAWINGS

[0007] The above content of the present application and other purposes, features and advantages will be more apparent through the following description of the embodiments of the present application with reference to the accompanying drawings.

[0008] Figure 1 A schematic diagram of a server I2C topology in the related art is shown.

[0009] Figure 2 A schematic diagram of a server according to a first embodiment of the present application is shown.

[0010] Figure 3 A schematic diagram of a server according to a second embodiment of the present application is shown.

[0011] Figure 4 A schematic diagram of a server according to a third embodiment of the present application is shown.

[0012] Figure 5 A schematic diagram of a server according to a fourth embodiment of the present application is shown.

[0013] Figure 6 An operation flowchart of a control method of a server according to the embodiments of the present application is shown. DETAILED DESCRIPTION

[0014] Hereinafter, the embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary, and are not intended to limit the scope of the present application. In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to one skilled in the art that one or more embodiments can be practiced without these specific details. In addition, in the following description, descriptions of well-known structures and techniques have been omitted to avoid unnecessarily obscuring the concepts of the present application.

[0015] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting thereof. The use of the terms "including," "comprising," "having," and the like are meant to encompass the items listed thereafter, but do not preclude the addition of one or more other items.

[0016] All terms used herein, including technical and scientific terms, have the meaning commonly understood by one of ordinary skill in the art unless otherwise defined herein. It should be noted that the use of any terms herein should not be interpreted to limit the scope of the present disclosure, but rather the scope of the disclosure is defined by the claims.

[0017] In the case of using expressions similar to "at least one of A, B, and C, etc.", it is generally understood that the expression is intended to include one or more of the items listed (e.g., "a system having at least one of A, B, and C" should include, but not be limited to, a system having A alone, a system having B alone, a system having C alone, a system having both A and B, a system having both A and C, a system having both B and C, and / or a system having A, B, and C, etc.).

[0018] The I2C bus is the "sensing nerve" of the server hardware monitoring system, which can be used to ensure the stable operation of the server. If the I2C bus fails, the management card cannot obtain sensor data. For example, the management card can monitor the temperature, voltage, fan speed, etc. of the internal devices of the server through multiple I2C signal lines, and can also adjust the output of the power module and adjust the fan speed. In the case of I2C bus failure, the management card fails to access the I2C device, resulting in the loss of key indicators such as temperature and voltage, which may cause the server to overheat, slow down, or even crash. The management card cannot write a pulse width modulation value to adjust the fan speed through the I2C bus, resulting in uncontrolled fan speed, causing insufficient heat dissipation or excessive noise.

[0019] I2C bus devices can be divided into two categories, I2C host and I2C device. The I2C host is the master of the I2C bus, which is used to initiate communication, control data transmission timing, and determine whether to interact with the I2C device. The I2C device is the controlled party of the I2C bus, which can be identified by a unique address and can respond after receiving the address from the I2C host. There can be multiple I2C hosts and multiple I2C devices on one I2C link, but only one I2C host can initiate communication at the same time. Multiple I2C devices are distinguished by configuring different addresses, and multiple I2C devices can establish communication with the I2C host according to their respective addresses.

[0020] Figure 1 A schematic diagram of a server I2C topology in the related art is shown.

[0021] As Figure 1As shown, the management card can poll the I2C devices behind the I2C signal lines through the multiple sets of I2C signal lines. The management card can poll the multiple power supply units behind the first set of I2C signal lines through the first set of I2C signal lines, which can include power supply unit I and power supply unit II. The management card can poll the multiple fan controllers behind the second set of I2C signal lines through the second set of I2C signal lines, which can include fan controller I, fan controller II, and fan controller III. The management card can poll the multiple voltage regulators behind the third set of I2C signal lines through the third set of I2C signal lines, which can include voltage regulator I, voltage regulator II, voltage regulator III, and voltage regulator IV. The management card can poll the multiple temperature sensors behind the fourth set of I2C signal lines through the fourth set of I2C signal lines, which can include temperature sensor I and temperature sensor II. The management card can poll the multiple management hardware behind the fifth set of I2C signal lines through the fifth set of I2C signal lines, which can include management hardware I, management hardware II, and management hardware III.

[0022] In the case of failure of the I2C host, the function of the I2C link will fail, and the monitoring management system will crash. The server may be down, or may be running continuously at high temperature and cause failure.

[0023] In the case of failure of the I2C link, the I2C host cannot access the I2C device behind the I2C link, and part of the information is abnormal. For example, in the case of voltage abnormality, the I2C host cannot obtain voltage information due to failure of the I2C link, so it cannot make a judgment of voltage abnormality, and it is difficult to cut off the power supply of the I2C device in time, and the I2C device may be damaged due to continuous high voltage.

[0024] In the case of failure of the I2C device, the I2C host cannot access the I2C device. In some abnormal cases, the I2C host can send a reset signal to the I2C device through the I2C link, so that the I2C device can reestablish communication with the I2C host. For a single-host I2C bus, it is not possible to send a reset signal to the I2C device through the I2C link, and it is necessary to power off the server or other operations to restore normal operation.

[0025] In the structural design of part I2C bus, the redundant I2C host is used to assist in realizing the management function. However, the redundant I2C host has limited assisted management function. Generally, the redundant I2C host takes over the control function of fan speed when the I2C host fails, so that the fan can run at normal speed or high speed to prevent the related device from failing due to high temperature. In the case of failure of the I2C link between the redundant I2C host and the temperature sensor or fan controller, the redundant I2C host cannot realize the control function of fan speed.

[0026] Therefore, the embodiments of the present application propose a dual-host management scheme. When the I2C host fails, the standby host can take over the work. When the I2C link fails, both dual hosts can provide I2C signals. When the I2C device fails, both dual hosts can provide reset signals, so that the server can continue to work normally.

[0027] Figure 2 A schematic diagram of a server according to the first embodiment of the present application is shown.

[0028] As shown in Figure 2 , the server includes a baseboard management controller 1, a standby management controller 2, a component 3, and a selection circuit 4.

[0029] The selection circuit 4 is connected with the baseboard management controller 1, the standby management controller 2, and the component 3, and the standby management controller 2 is connected with the baseboard management controller 1. The selection circuit 4 can establish a first bus link between the baseboard management controller 1 and the component 3 under the control of the baseboard management controller 1, and establish a second bus link between the standby management controller 2 and the component 3 under the control of the standby management controller 2.

[0030] The baseboard management controller 1 and the standby management controller 2 can be independent hardware or fixed components, and can remotely monitor, manage, and maintain the server. The component 3 can be a key hardware component in the server, for example, a sensor component.

[0031] The baseboard management controller 1 can control the selection circuit 4 to establish the first bus link between the baseboard management controller 1 and the component 3 through a first enable signal, and perform platform management on the component 3 through the first bus link. The baseboard management controller 1 can perform platform management on the component 3 to realize the health management function of the server, such as temperature monitoring, voltage monitoring, hardware state detection, fan speed control, and asset information management.

[0032] The standby management controller 2 can control the selection circuit 4 to establish the second bus link between the standby management controller 2 and the component 3 through a second enable signal in response to detecting that the baseboard management controller 1 fails, and take over the platform management on the component 3 through the second bus link.

[0033] The selection circuit 4 is a logic switching circuit. The selection circuit 4 can be a multiplexer. The selection circuit 4 can also be a tri-state gate. The selection circuit 4 can also be a programmable logic device. The selection circuit 4 can determine the on-off of the communication or the switching of the path according to a logic condition or a control signal. For example, the selection circuit 4 can directly control the on-off of the communication or the switching of the path based on a logic level of "1" or "0". The selection circuit 4 can also trigger the on-off of the communication or the switching of the path based on a timing logic, i.e. the rising edge or the falling edge of a clock pulse. The selection circuit 4 can also perform a forced switching of the path based on a trigger signal of an external event.

[0034] According to an embodiment of the present application, the selection circuit 4 is used to establish a dual-host management mode for the component 3. The selection circuit 4 can be controlled by the baseboard management controller 1 to communicatively connect the baseboard management controller 1 with the component 3, so that the baseboard management controller 1 can perform platform management on the component 3 to perform health management on the running state of the server. The selection circuit 4 can also be controlled by the backup management controller 2 to communicatively connect the backup management controller 2 with the component 3, so that the backup management controller 2 can take over the platform management on the component 3 when detecting a failure of the baseboard management controller 1. In the case of a failure of the baseboard management controller 1, the backup management controller 2 performs health management on the running state of the server, so that the server can maintain a normal working state or can recover to a normal working state when a communication connection fails.

[0035] Figure 3 A schematic diagram of a server according to a second embodiment of the present application is shown.

[0036] As shown in Figure 3 In the present embodiment, the baseboard management controller 1 can be connected with the selection circuit 4 through the first bus I2C_1 and the first enable signal line 1OE_L. The baseboard management controller 1 is configured to provide a first enable signal to the first enable signal line 1OE_L and transmit a first bus signal for platform management through the first bus I2C_1. The first enable signal line 1OE_L can be a wire carrying an "enable" logic level. The first enable signal provided by the baseboard management controller 1 to the first enable signal line 1OE_L can be active high or active low. The baseboard management controller 1 provides the first enable signal, which can control the running state of the device or circuit at the back end of the first enable signal line 1OE_L as needed.

[0037] The backup management controller 2 can be connected with the selection circuit 4 through a second bus I2C_2 and a second enable signal line 2OE_L, and the backup management controller 2 is configured to provide a second enable signal to the second enable signal line 2OE_L in response to detecting that the baseboard management controller 1 fails, and transmit a second bus signal for platform management through the second bus I2C_2. The second enable signal line 2OE_L can be a wire carrying an “enable” logic level. The second enable signal provided by the backup management controller 2 to the second enable signal line 2OE_L can be active high or active low. The backup management controller 2 provides the second enable signal in response to detecting that the baseboard management controller 1 fails, and can control the operating state of the device or circuit at the back end of the second enable signal line 2OE_L as needed.

[0038] The selection circuit 4 is connected with the component 3 through a third bus I2C_3, and the selection circuit 4 is configured to be able to respond to receiving the first enable signal from the first enable signal line 1OE_L to communicatively connect the first bus I2C_1 with the third bus I2C_3, so as to establish a first bus link between the baseboard management controller 1 and the component 3. The selection circuit 4 is further configured to be able to respond to receiving the second enable signal from the second enable signal line 2OE_L to communicatively connect the second bus I2C_2 with the third bus I2C_3, so as to establish a second bus link between the backup management controller 2 and the component 3.

[0039] In one example, in the case that the baseboard management controller 1 is normally operating, the baseboard management controller 1 can provide the first enable signal to the first enable signal line 1OE_L to control the selection circuit 4 to communicatively connect the first bus I2C_1 with the third bus I2C_3, so as to establish the first bus link. In the case that the baseboard management controller 1 fails, the baseboard management controller 1 cannot establish communication with the component 3, and needs to establish communication between the backup management controller 2 and the component 3. The backup management controller 2 provides the second enable signal to the second enable signal line 2OE_L in response to detecting that the baseboard management controller 1 fails, to control the selection circuit 4 to communicatively connect the second bus I2C_2 with the third bus I2C_3, so as to establish the second bus link.

[0040] In one example, the component 3 can include a plurality of temperature sensors and a fan controller. The first bus signal can include a collection signal of the plurality of temperature sensors by the baseboard management controller 1, and can also include a control signal of the fan controller by the baseboard management controller 1. The baseboard management controller 1 can transmit the first bus signal through the first bus link, and can poll the plurality of temperature sensors in the component 3 to periodically collect the temperature of a plurality of key positions in the server, such as the motherboard, the processor, the memory bank, the hard disk backplane, the power module, etc. In the case where the temperature exceeds a certain threshold, the baseboard management controller 1 can control the fan controller to adjust the fan speed according to the heat dissipation regulation strategy to perform heat dissipation processing. In the case where the temperature exceeds a certain threshold, if the baseboard management controller 1 fails, the baseboard management controller 1 cannot collect the temperature information, or in the case where the temperature information can be collected, the baseboard management controller 1 cannot make a judgment according to the heat dissipation regulation strategy; or the baseboard management controller 1 cannot adjust the fan speed through the fan controller to perform heat dissipation processing.

[0041] The standby management controller 2 can transmit the second bus signal through the second bus link in response to detecting that the baseboard management controller 1 fails. The second bus signal can include a collection signal of the plurality of temperature sensors by the standby management controller 2, and can also include a control signal of the fan controller by the standby management controller 2. The standby management controller 2 can take over the polling of the plurality of temperature sensors in the component 3, and can also control the fan controller to adjust the fan speed according to the heat dissipation regulation strategy to perform heat dissipation processing in the case where the temperature exceeds a certain threshold.

[0042] According to the embodiments of the present application, the baseboard management controller 1 controls the selection circuit 4 to establish the first bus link by using the first enable signal line 1OE_L, so that the baseboard management controller 1 can perform platform management on the component 3, and the baseboard management controller 1 can perform main control in the case where the baseboard management controller 1 normally operates. The standby management controller 2 controls the selection circuit 4 to establish the second bus link by using the second enable signal line 2OE_L, so that the standby management controller 2 can take over the control in the case where the baseboard management controller 1 fails, and the reliability of the dual-host management mode can be improved.

[0043] As shown in FIG. 1, Figure 3 In one embodiment, the first bus I2C_1 can be an integrated circuit bus including a serial clock line (SCL) and a serial data line (SDA). The first bus I2C_1 includes two cables, i.e., a clock line I2C_1_SCL of the first bus I2C_1 and a data line I2C_1_SDA of the first bus I2C_1.

[0044] The second bus I2C_2 can be an integrated circuit bus, and the second bus I2C_2 includes two cables, namely a clock line I2C_2_SCL of the second bus I2C_2 and a data line I2C_2_SDA of the second bus I2C_2.

[0045] The third bus I2C_3 can be an integrated circuit bus, and the third bus I2C_3 includes two cables, namely a clock line I2C_3_SCL of the third bus I2C_3 and a data line I2C_3_SDA of the third bus I2C_3.

[0046] The communication connection between the first bus I2C_1 and the third bus I2C_3 includes that the clock line I2C_1_SCL of the first bus I2C_1 is in communication connection with the clock line I2C_3_SCL of the third bus I2C_3, and the data line I2C_1_SDA of the first bus I2C_1 is in communication connection with the data line I2C_3_SDA of the third bus I2C_3.

[0047] The communication connection between the second bus I2C_2 and the third bus I2C_3 includes that the clock line I2C_2_SCL of the second bus I2C_2 is in communication connection with the clock line I2C_3_SCL of the third bus I2C_3, and the data line I2C_2_SDA of the second bus I2C_2 is in communication connection with the data line I2C_3_SDA of the third bus I2C_3.

[0048] As shown in Figure 3 The substrate management controller 1 can also be connected with the selection circuit 4 through the first reset signal line Rt1, and the substrate management controller 1 is further configured to send a first reset signal for resetting the component 3 to the first reset signal line Rt1 in response to detecting the first bus link abnormality.

[0049] The standby management controller 2 can also be connected with the selection circuit 4 through the second reset signal line Rt2, and the standby management controller 2 is further configured to send a second reset signal for resetting the component 3 to the second reset signal line Rt2 in response to detecting the second bus link abnormality.

[0050] The selection circuit 4 can also be connected with the component 3 through the third reset signal line Rt3, and the selection circuit 4 is further configured to, in response to receiving the first enable signal from the first enable signal line 1OE_L, communicatively connect the first reset signal line Rt1 with the third reset signal line Rt3; and in response to receiving the second enable signal from the second enable signal line 2OE_L, communicatively connect the second reset signal line Rt2 with the third reset signal line Rt3.

[0051] For the part of hardware devices in component 3, there is no self hardware reset signal, and the hardware reset signal can only be transmitted through the data line I2C_1_SDA of the first bus I2C_1 in the first bus link and the data line I2C_3_SDA of the third bus I2C_3. In the case of abnormality of the first bus link, the baseboard management controller 1 cannot transmit the hardware reset signal to the part of hardware devices in component 3 through the data line I2C_1_SDA of the first bus I2C_1 in the first bus link and the data line I2C_3_SDA of the third bus I2C_3.

[0052] The case of abnormality of the first bus link can be that one or more of the first bus signals on the clock line I2C_1_SCL of the first bus I2C_1, the data line I2C_1_SDA of the first bus I2C_1, the clock line I2C_3_SCL of the third bus I2C_3 and the data line I2C_3_SDA of the third bus I2C_3 in the first bus link are long-term at low level or high level in the case that the baseboard management controller 1 or component 3 does not correctly execute the protocol logic, and the baseboard management controller 1 and component 3 cannot communicate. In one example, the read-only memory in component 3 is accidentally powered off during data transmission, causing the data line I2C_1_SDA of the first bus I2C_1 and the data line I2C_3_SDA of the third bus I2C_3 to be continuously clamped at low level, and the node in the first bus link loses arbitration right, causing the first bus link communication to stop.

[0053] The case of abnormality of the first bus link can also be that external noise instantaneously pulls one or more of the first bus signals on the clock line I2C_1_SCL of the first bus I2C_1, the data line I2C_1_SDA of the first bus I2C_1, the clock line I2C_3_SCL of the third bus I2C_3 and the data line I2C_3_SDA of the third bus I2C_3 in the first bus link to an undeserved high level or low level, and the baseboard management controller 1 or component 3 erroneously judges the unexpected jump as a start signal or an end signal, causing the communication protocol to fail.

[0054] The case of abnormality of the first bus link can also be that the arbitration of the baseboard management controller 1 fails, causing the data line I2C_1_SDA of the first bus I2C_1 and the data line I2C_3_SDA of the third bus I2C_3 in the first bus link to be occupied all the time.

[0055] In response to detecting an anomaly in the first bus link, the baseboard management controller 1 controls the selection circuit 4 to communicatively connect the first reset signal line Rt1 and the third reset signal line Rt3. The baseboard management controller 1 can send a first reset signal to reset component 3 via the first reset signal line Rt1 and the third reset signal line Rt3, thereby re-establishing the first bus link without requiring the server to stop operating or undergo power-off maintenance. The first reset signal can be used to reset the hardware devices of component 3 or to reset the software register information of component 3.

[0056] Similarly, the situation regarding the second bus link failure is similar to that of the first bus link failure. In response to detecting a second bus link failure, the backup management controller 2 connects the second reset signal line Rt2 and the third reset signal line Rt3 via the control selection circuit 4. The backup management controller 2 can then send a second reset signal to reset component 3 through these two lines, re-establishing the second bus link without requiring the server to stop operating or undergo power-off maintenance. The second reset signal can be used to reset the hardware of component 3 or to reset the software register information of component 3.

[0057] According to embodiments of this application, in the event of a first bus link failure, the baseboard management controller 1 can re-establish the first bus link via a separate reset link. In the event of a second bus link failure, the backup management controller 2 can re-establish the second bus link via a separate reset link. This allows the system to be isolated from the internal logic of either the first or second bus link, forcibly restoring either the first or second bus link from an unknown state to a secure state, thereby improving server reliability.

[0058] like Figure 3 As shown, the backup management controller 2 is connected to the baseboard management controller 1 via the first detection signal line Ht. The backup management controller 2 is configured to receive status signals from the baseboard management controller 1 via the first detection signal line Ht. These status signals can be power supply signals, clock signals, or self-test signals from the baseboard management controller 1. The backup management controller 2 can determine whether the baseboard management controller 1 has malfunctioned based on the status signals. In response to detecting a malfunction in the baseboard management controller 1, the backup management controller 2 controls the selection circuit 4 to connect the backup management controller 2 to the component 3 and takes over the platform management of the component 3.

[0059] According to the embodiment of the present application, the backup management controller 2 judges whether to take over the platform management of the component 3 according to the running state of the baseboard management controller 1, takes over the platform management of the component 3 in time when the baseboard management controller 1 fails, can avoid server downtime, and can also timely eliminate the baseboard management controller 1 from the platform management of the component 3, preventing the baseboard management controller 1 from outputting an error signal due to failure and affecting the normal operation of the server.

[0060] According to the embodiment of the present application, the state signal is a heartbeat signal periodically sent by the baseboard management controller 1. The heartbeat signal is a regular pulse signal and can be used to represent the working state of the baseboard management controller 1. The heartbeat signal can judge whether the node and process are normal in real time, and can also run independently without affecting the main process of the baseboard management controller 1.

[0061] As shown in Figure 3 The backup management controller 2 is also connected with the baseboard management controller 1 through a fourth reset signal line Rt4. The backup management controller 2 is also configured to send a fourth reset signal to the fourth reset signal line Rt4 in response to detecting that the baseboard management controller 1 fails.

[0062] In one example, the backup management controller 2 sends a fourth reset signal to the baseboard management controller 1 through the fourth reset signal line Rt4 in response to an abnormality of the heartbeat signal periodically sent by the baseboard management controller 1. The baseboard management controller 1 is also configured to perform a reset according to the fourth reset signal, and re-take over the platform management of the component 3 in the case of failure recovery after the reset.

[0063] In another example, the backup management controller 2 sends a fourth reset signal to the baseboard management controller 1 through the fourth reset signal line Rt4 in response to an abnormality of the heartbeat signal periodically sent by the baseboard management controller 1. The baseboard management controller 1 cannot perform a reset according to the fourth reset signal, or the baseboard management controller 1 performs a reset according to the fourth reset signal but cannot recover to normal operation, and the backup management controller 2 can continue to take over the platform management of the component 3 while issuing a failure indication to request administrator intervention.

[0064] According to the embodiment of the present application, by setting the fourth reset signal line Rt4, the backup management controller 2 sends a fourth reset signal to the baseboard management controller 1 in response to detecting that the baseboard management controller 1 fails, to reset the baseboard management controller 1, which can eliminate the failure factors of the baseboard management controller 1, reduce manual intervention, improve repair efficiency, and also issue a failure indication after the baseboard management controller 1 fails to reset, avoiding long-term downtime of the baseboard management controller 1.

[0065] As shown in Figure 3As shown, the standby management controller 2 is also connected with the baseboard management controller 1 through the fourth bus I2C_4, and information interaction between the standby management controller 2 and the baseboard management controller 1 can be realized. The standby management controller 2 and the baseboard management controller 1 are further configured to, in the case that neither the baseboard management controller 1 nor the standby management controller 2 has a fault, periodically alternate taking over the platform management of the component 3 and transmitting detection information of the component 3 generated in the platform management process to each other through the fourth bus I2C_4.

[0066] In one example, the baseboard management controller 1 can determine whether the standby management controller 2 is in a normal working state, and the standby management controller 2 can determine whether the baseboard management controller 1 is in a normal working state. In the case that both the baseboard management controller 1 and the standby management controller 2 are in a normal working state, a preset time length can be set by the processor of the server, and the platform management of the component 3 by the baseboard management controller 1 can be periodically switched to the platform management of the component 3 by the standby management controller 2, and information interaction and mutual transmission of detection information can be performed. For example, the processor of the server can send a clock signal to the baseboard management controller 1 and the standby management controller 2. When the clock signal is in a high level state, the baseboard management controller 1 takes over the platform management of the component 3. When the clock signal is in a low level state, the standby management controller 2 takes over the platform management of the component 3. The preset time length can be one minute, the preset time length can be two minutes, the preset time length can be three minutes, and the preset time length can be set according to actual conditions.

[0067] In one example, in the case that the baseboard management controller 1 determines that the standby management controller 2 is in an abnormal working state, the server processor can send a signal to prohibit the baseboard management controller 1 and the standby management controller 2 from periodically alternating taking over the platform management of the component 3, and the baseboard management controller 1 alone can take over the platform management of the component 3 until the standby management controller 2 returns to a normal working state, and then the periodically alternating taking over of the platform management of the component 3 is resumed.

[0068] In one example, in the case that the standby management controller 2 determines that the baseboard management controller 1 is in an abnormal working state, the server processor can prohibit the baseboard management controller 1 and the standby management controller 2 from periodically alternating taking over the platform management of the component 3, and the standby management controller 2 alone can take over the platform management of the component 3 until the baseboard management controller 1 returns to a normal working state, and then the periodically alternating taking over of the platform management of the component 3 is resumed.

[0069] According to the embodiments of the present application, in the case that neither the baseboard management controller 1 nor the standby management controller 2 has a fault, the periodically alternating taking over of the platform management of the component 3 can timely troubleshoot the self-abnormalities of the baseboard management controller 1 and the standby management controller 2.

[0070] According to an embodiment of the present application, the platform management includes obtaining target data from the component 3, comparing the target data with reference information, sending control information to the component 3 according to the comparison result, and the detection information is the comparison result of the target data and the reference information.

[0071] In the server design, the platform management of the component 3 includes obtaining target data from the component 3. The baseboard management controller 1 can communicate and collect data from the devices of the component 3 according to the polling mechanism through the first bus link. The baseboard management controller 1 can send requests to each device of the component 3 on the first bus link in a fixed time sequence and period, and can obtain the target data. The baseboard management controller 1 can compare the target data with the reference information, and send control information to the component 3 through the first bus link according to the comparison result. The baseboard management controller 1 can execute corresponding management measures or issue fault alarm information according to the detection information.

[0072] In the case of failure of the baseboard management controller 1 or abnormality of the first bus link, it may cause that the target data cannot be obtained or the wrong target data is obtained, which affects the logical judgment of the baseboard management controller 1. The standby management controller 2 can communicate and collect data from the devices of the component 3 according to the polling mechanism through the second bus link. The standby management controller 2 can send requests to each device of the component 3 on the second bus link in a fixed time sequence and period, and can obtain the target data. The standby management controller 2 can compare the target data with the reference information, and send control information to the component 3 through the second bus link according to the comparison result. The standby management controller 2 can execute corresponding management measures or issue fault alarm information according to the detection information. The target data can include test data of multiple sensors, state information of multiple hardware, asset information, etc. The polling period can be 2 seconds, 3 seconds, 4 seconds, and the polling period can be adjusted according to the design of the server.

[0073] In one example, the baseboard management controller 1 can poll the multiple power supply units in the component 3 through the first bus link with a period of 2 seconds, send a request to each power supply unit, and obtain the voltage of each power supply unit. The baseboard management controller 1 compares the voltage of each power supply unit with a preset voltage threshold, the baseboard management controller 1 judges that the voltage in the target data exceeds the preset voltage threshold, the baseboard management controller 1 records the event log, and issues an alarm, the baseboard management controller 1 sends control information to the power supply unit whose voltage exceeds the preset voltage threshold through the first bus link, which can control the power supply unit whose voltage exceeds the preset voltage threshold to be set to the off mode, and switch the standby power supply unit to ensure that the load does not power off and prevent damage to the server.

[0074] In case of failure of the baseboard management controller 1 or abnormality of the first bus link, it can cause that the voltage of each power supply unit cannot be acquired or an error voltage is acquired, which affects the logical judgment of the baseboard management controller 1 and causes damage of the server. The standby management controller 2 can periodically poll the plurality of power supply units in the component 3 through the second bus link with a period of 2 seconds, send a request to each power supply unit, and acquire the voltage of each power supply unit. The standby management controller 2 compares the voltage of each power supply unit with a preset voltage threshold, judges that the voltage in the target data exceeds the preset voltage threshold, records the event log, and issues an alarm. The standby management controller 2 sends control information to the power supply unit with the voltage exceeding the preset voltage threshold through the second bus link, and can control the power supply unit with the voltage exceeding the preset voltage threshold to be set to an off mode, and switch the standby power supply unit, to ensure that the load does not power off and prevent damage to the server.

[0075] According to the embodiments of the present application, the baseboard management controller 1 and the standby management controller 2 can execute corresponding management measures or issue fault alarm information according to the detection information generated in the execution of the platform management process, to prevent damage to the server.

[0076] In one example, the baseboard management controller 1 and the standby management controller 2 are further configured to judge whether there is a difference between the detection information generated in the execution of the platform management process and the detection information received from the other party, and in response to the difference existing for K consecutive times, issue an alarm information, where K is a preset integer.

[0077] For example, the baseboard management controller 1 and the standby management controller 2 exchange detection information through the fourth bus I2C_4, the standby management controller 2 judges that there is a difference between the detection information of the mainboard temperature generated by the standby management controller 2 and the detection information of the mainboard temperature received from the baseboard management controller 1, and the standby management controller 2 issues an alarm information in response to the difference existing for 3 consecutive times, requesting the administrator to intervene to determine whether the standby management controller 2 or the temperature sensor of the mainboard is abnormal.

[0078] According to the embodiments of the present application, in the case that neither the baseboard management controller 1 nor the standby management controller 2 fails, the platform management of the component 3 is periodically alternately taken over, and the detection information of both parties is monitored, so that the information abnormality of the first bus link and the component 3 and the information abnormality of the second bus link and the component 3 can be timely investigated, to avoid causing more serious damage to the server.

[0079] According to the embodiments of the present application, the fourth bus I2C_4 can be an integrated circuit bus.

[0080] AsFigure 3 As shown, the backup management controller 2 is also connected with the baseboard management controller 1 through a second detection signal line GPIO. The second detection signal line GPIO can be a general purpose input output (GPIO).

[0081] When the baseboard management controller 1 and the backup management controller 2 abnormally communicate the detection information through the fourth bus I2C_4, the backup management controller 2 is further configured to send the state information of the backup management controller 2 to the baseboard management controller 1 through the second detection signal line GPIO. For example, the state information of the backup management controller 2 is 0, indicating that the backup management controller 2 is in a normal working state, and the state information of the backup management controller 2 is 1, indicating that the backup management controller 2 is in an abnormal working state.

[0082] When the baseboard management controller 1 and the backup management controller 2 abnormally communicate the detection information through the fourth bus I2C_4, the information can be communicated through the second detection signal line GPIO, and the fourth bus I2C_4C can continuously transmit a plurality of sets of data information.

[0083] In one example, the state information of the backup management controller 2 can be represented by the high level state or the low level state of the digital level pin of the second detection signal line GPIO. The state of the digital level pin of the second detection signal line GPIO can send the state information of the backup management controller 2 to the baseboard management controller 1 through the second detection signal line GPIO with zero delay, fully utilize the hardware resources, and timely feedback the state information of the backup management controller 2, so as to timely perform abnormal processing when the backup management controller 2 is in an abnormal working state.

[0084] According to the embodiment of the present application, the backup management controller 2 is realized by a programmable circuit module provided by the server.

[0085] According to the embodiment of the present application, the programmable circuit module of the server can be programmed and designed according to the requirements of the server of the embodiment of the present application, can provide the same number of the second bus I2C_2 as the first bus I2C_1 of the baseboard management controller 1, can provide the same number of the second enable signal line 2OE_L as the first enable signal line 1OE_L of the baseboard management controller 1, can provide the same number of the second reset signal line Rt2 as the first reset signal line Rt1 of the baseboard management controller 1, and can provide the same number of the second bus link as the first bus link of the baseboard management controller 1. The programmable circuit module of the server can dynamically reconfigure the internal logic unit, without replacing or disassembling the server device, can customize the functions of the programmable circuit module, can take over the platform management of the component 3, including functions such as processing test data of multiple sensors in parallel and reading part of the register, and can also respond to signals from the baseboard management controller 1 and signals from the component 3 in real time.

[0086] Figure 4 A schematic diagram of a server according to a third embodiment of the present application is shown.

[0087] As shown in Figure 4 , the server further includes a central processing unit connected with the baseboard management controller 1 through a fifth bus PCIe_5. The fifth bus PCIe_5 can be a Peripheral Component Interconnect Express (PCIe).

[0088] In one example, the central processing unit can be in communication connection with the baseboard management controller 1 through the fifth bus PCIe_5, and can actively read the health status of the baseboard management controller 1, including the firmware version, running state, and whether it is working normally. For example, the central processing unit can prevent the baseboard management controller 1 from running abnormally by checking whether the baseboard management controller 1 is online, so as to cause the platform management function of the baseboard management controller 1 to the component 3 to fail. For example, the central processing unit can query the hardware error record by reading the log of the baseboard management controller 1.

[0089] In one example, the central processing unit can be in communication connection with the baseboard management controller 1 through the fifth bus PCIe_5, and can send commands to the baseboard management controller 1 to adjust the management strategy. For example, the central processing unit can periodically control the alternate takeover of the platform management of the component 3 by the standby management controller 2 and the baseboard management controller 1 through the fifth bus PCIe_5 in the case that both the baseboard management controller 1 and the standby management controller 2 are in a normal working state.

[0090] In one example, the central processing unit can be communicatively connected with the baseboard management controller 1 through a fifth bus PCIe_5, and in response to an alarm information of the baseboard management controller 1 in a system crash, read the last error log of the baseboard management controller 1 to assist troubleshooting.

[0091] In one example, the central processing unit can be communicatively connected with the standby management controller 2 (a programmable circuit module provided by the server), and in response to an alarm information of the standby management controller 2 in a case where the baseboard management controller 1 fails, read the last error log of the standby management controller 2 to assist troubleshooting.

[0092] As shown in Figure 4 , the central processing unit is connected with the output end of the selection circuit 4 through a sixth bus I2C_6. The sixth bus I2C_6 can be an integrated circuit bus.

[0093] The baseboard management controller 1 can be communicatively connected with the central processing unit through a first bus link. The baseboard management controller 1 can obtain the core voltage, performance parameters, safe boot state, etc. of the central processing unit through the first bus link.

[0094] The standby management controller 2 can be communicatively connected with the central processing unit through a second bus link. The standby management controller 2 can obtain the core voltage, performance parameters, safe boot state, etc. of the central processing unit through the second bus link.

[0095] As shown in Figure 4 , the central processing unit is also connected with a peripheral component interconnect express bus switching circuit (PCIe switching circuit) through a seventh bus PCIe_7. The seventh bus PCIe_7 can be a peripheral component interconnect express bus.

[0096] As shown in Figure 4 , the baseboard management controller 1, the standby management controller 2, the selection circuit 4 and the central processing unit are located on the mainboard of the server, and the component 3 includes at least one of the components on the mainboard and the components outside the mainboard.

[0097] The components outside the mainboard can include functional board cards, such as GPU (Graphics Processing Unit, graphics processor) cases, hard disk backplanes, fan plates, and power supply plates. The GPU case can include devices such as GPUs, the hard disk backplane can include devices such as hard disks, the fan plate can include devices such as fans, and the power supply plate can include devices such as PSUs (Power Supply Unit, power supply unit).

[0098] As shown in Figure 4As shown, the server further comprises a cabinet and a middle backboard arranged in the cabinet, the middle backboard divides the cabinet into a first space and a second space, the mainboard is located in the first space, components outside the mainboard are located in the second space, and the selection circuit 4 is connected with the components outside the mainboard through the middle backboard. The middle backboard can serve as an interconnection hub of the server, support the interconnection of the mainboard and various functional boards, and provide standardized slots and standardized power supply for the components outside the mainboard.

[0099] As shown in the figure, Figure 4 The components 3 can include a multiplexing circuit and a plurality of hardware units, the input end of the multiplexing circuit is connected with the selection circuit, and the output end of the multiplexing circuit is connected with the plurality of hardware units. The hardware units include a heat sink, a power supply or management hardware.

[0100] The multiplexing circuit can be an analog multiplexer, a digital multiplexer, a time division multiplexing circuit, etc. The multiplexing circuit can enable the baseboard management controller 1 to access multiple hardware units at the same time, save the resources of the first bus link, and improve the utilization rate of the first bus link. The multiplexing circuit can enable the standby management controller 2 to access multiple hardware units at the same time, save the resources of the second bus link, and improve the utilization rate of the second bus link.

[0101] The multiplexing circuit can be arranged on the mainboard according to the position requirements of the plurality of hardware units. For example, the mainboard multiplexing circuit is arranged on the mainboard. The mainboard multiplexing circuit can be connected in communication with the third bus I2C_3 through the first bus I2C_1, so that the baseboard management controller 1 and the hardware unit I, the hardware unit II and the hardware unit III transmit the first bus signal. The mainboard multiplexing circuit can be connected in communication with the third bus I2C_3 through the second bus I2C_2, so that the standby management controller 2 and the hardware unit I, the hardware unit II and the hardware unit III transmit the second bus signal.

[0102] The multiplexing circuit can be arranged on each functional board according to the position requirements of the plurality of hardware units. For example, the multiplexing circuit T1 of the GPU device is arranged on the GPU cabinet, which is used to transmit the first bus signal between the baseboard management controller 1 and the GPU device. The multiplexing circuit T2 of the hard disk device is arranged on the hard disk backboard, which is used to transmit the first bus signal between the baseboard management controller 1 and the hard disk device. The multiplexing circuit T3 of the fan is arranged on the fan board, which is used to transmit the first bus signal between the baseboard management controller 1 and the fan device. The multiplexing circuit T4 of the PSU is arranged on the power supply board, which is used to transmit the first bus signal between the baseboard management controller 1 and the PSU device.

[0103] The selection circuit 4 is arranged on the mainboard of the server in the first space of the cabinet, which can reduce the number of signals of the middle backboard.

[0104] According to embodiments of this application, component 3 may include a heat sink, a power supply, or management hardware.

[0105] According to an embodiment of this application, the server includes N selection circuits and N components, wherein the first input terminal of the nth selection circuit is connected to the baseboard management controller, the second input terminal of the nth selection circuit is connected to the backup management controller, and the output terminal of the nth selection circuit is connected to the nth component, where N and n are both positive integers, and 1≤n≤N.

[0106] Figure 5 A schematic diagram of a server according to a fourth embodiment of this application is shown.

[0107] like Figure 5 As shown, in this embodiment, N=5, and the server may include 5 selection circuits and 5 components. The backup management controller may be a programmable circuit module.

[0108] The first selection circuit is connected to the baseboard management controller via the first first bus I2C1_1, to the programmable circuit module via the first second bus I2C1_2, and to the first component via the first third bus I2C1_3. The first component may include a first multiplexing circuit and multiple power supply units, which may include a first power supply unit PSU1, a second power supply unit PSU2, a third power supply unit PSU3, and a fourth power supply unit PSU4.

[0109] The second selection circuit is connected to the baseboard management controller via a second first bus I2C2_1, to the programmable circuit module via a second second bus I2C2_2, and to the second component via a second third bus I2C2_3. The second component may include a second multiplexing circuit and multiple cooling fans, which may include a first cooling fan F1, a second cooling fan F2, a third cooling fan F3, a fourth cooling fan F4, a fifth cooling fan F5, and a sixth cooling fan F6.

[0110] The third selection circuit is connected to the baseboard management controller via the third first bus I2C3_1, to the programmable circuit module via the third second bus I2C3_2, and to the third component via the third third bus I2C3_3. The third component may include a third multiplexing circuit and multiple voltage regulators, which may include a first voltage regulator VR1, a second voltage regulator VR2, a third voltage regulator VR3, a fourth voltage regulator VR4, a fifth voltage regulator VR5, a sixth voltage regulator VR6, a seventh voltage regulator VR7, and an eighth voltage regulator VR8.

[0111] The fourth selection circuit is connected with the baseboard management controller through a fourth first bus I2C4_1, connected with the programmable circuit module through a fourth second bus I2C4_2, and connected with a fourth component through a fourth third bus I2C4_3. The fourth component can include a fourth multiplexing circuit and a plurality of temperature sensors, and the plurality of temperature sensors can include a first temperature sensor TS1, a second temperature sensor TS2, a third temperature sensor TS3, and a fourth temperature sensor TS4.

[0112] The fifth selection circuit is connected with the baseboard management controller through a fifth first bus I2C5_1, connected with the programmable circuit module through a fifth second bus I2C5_2, and connected with a fifth component through a fifth third bus I2C5_3. The fifth component can include a fifth multiplexing circuit and a plurality of management hardware, such as a clock generator, a power monitor, a programmable unit, a peripheral component interconnect device, a read-only memory, and a hot plug module.

[0113] In the server of the embodiment of the present application, the baseboard management controller can take over the platform management of the component 3 through the first bus link, and can implement the following functions.

[0114] 1. Temperature monitoring. The baseboard management controller can poll the temperature sensors of a plurality of key positions such as the mainboard, the processor, the memory bar, the hard disk backboard, and the power module through the first bus link, and periodically refreshes the temperature of each area in real time. When the temperature of any node exceeds the preset threshold, the baseboard management controller will promptly issue an alarm and push it to the administrator, reminding the administrator to take measures to prevent the temperature from rising and damaging the hardware.

[0115] 2. Voltage monitoring. The baseboard management controller can poll the voltage sensors on the mainboard, the processor, the memory, and the power module through the first bus link, and periodically lock the accurate value of the key voltage. When the reading of the voltage sensor deviates from the preset tolerance, such as being too high or too low, the baseboard management controller will immediately record the abnormal time, the measured value, etc., and promptly issue an alarm and push it to the administrator, reminding the administrator to take measures to prevent the hardware from malfunctioning or being damaged due to voltage problems.

[0116] 3. Hardware state detection. The baseboard management controller can poll the state registers of devices such as the processor, the memory, the hard disk, and the network card through the first bus link, and periodically collects the state flags. When the running state of the device appears abnormal, for example, the fan speed is zero or the communication link is reduced, the baseboard management controller will immediately record the fault information such as the device code, the fault code, and the timestamp, and provide it to the administrator to help the administrator quickly locate and solve the problem.

[0117] 4. Fan speed control. The baseboard management controller continuously reads the data of the plurality of temperature sensors and power consumption chips through the first bus link to assess the thermal load of the key areas in real time. The baseboard management controller communicates with the fan controller through the I2C bus to automatically adjust the fan speed according to the thermal load of the system. When the temperature of the system rises, the baseboard management controller sends a command to the fan controller through the I2C bus to increase the fan speed, so that the heat can be quickly discharged from the case. When the temperature falls, the baseboard management controller controls the fan controller to reduce the speed through the I2C bus, which can reduce the wind noise and save energy, so that the system can maintain good heat dissipation effect under different loads.

[0118] 5. Asset information management. The baseboard management controller accesses the mainboard, processor, memory, hard disk, power supply and other components one by one through the first bus link, automatically reads the model, serial number, configuration information, etc., and stores them locally. Precise asset information also supports batch inventory, on-demand upgrade, fault tracing and budget forecasting, etc. The administrator can view the accurate configuration of the server in real time through the cloud management platform, which facilitates the administrator to manage and track the server assets uniformly.

[0119] In one example, the baseboard management controller can control the fourth selection circuit to connect the fourth first bus I2C4_1 and the fourth third bus I2C4_3 in communication, obtain the temperature of each area, and in the case that the temperature exceeds a certain threshold, the baseboard management controller can control the corresponding fan controller through the fourth first bus I2C4_1 and the fourth third bus I2C4_3 in communication according to the heat dissipation control strategy, to increase the fan speed to enhance the dissipation effect. When the temperature falls to the normal range, the baseboard management controller can control the corresponding fan controller through the fourth first bus I2C4_1 and the fourth third bus I2C4_3 in communication according to the heat dissipation control strategy, to reduce the fan speed to reduce the wind noise and save power.

[0120] The heat dissipation regulation strategy can be to regulate the fan speed to a corresponding gear according to different temperature thresholds. For example, the temperature thresholds can be set to 50℃, 70℃, and 90℃, respectively, and the fan speed can be set to four gears, for example, 30%, 50%, 70%, and 100%. When the temperature is below 50℃, the baseboard management controller can determine that it is low temperature, and regulate the fan speed to the 30% gear. When the temperature is in the range of 50℃-70℃, the baseboard management controller can determine that it is normal temperature, and regulate the fan speed to the 50% gear. When the temperature is in the range of 70℃-90℃, the baseboard management controller can determine that it is higher temperature, and regulate the fan speed to the 70% gear. When the temperature exceeds 90℃, the baseboard management controller can determine that it is too high temperature, and regulate the fan speed to the 100% gear. In the case that the temperature exceeds 90℃ and lasts for a certain period of time, the baseboard management controller will issue an alarm information, and at the same time, can also protect the equipment and server by reducing the frequency of the control processor and other operations.

[0121] In the server of the embodiments of the present application, the standby management controller can realize the above functions through the second bus link.

[0122] In one example, in response to detecting that the baseboard management controller fails, the programmable circuit module controls the fourth selection circuit to communicatively connect the fourth second bus I2C4_2 and the fourth third bus I2C4_3, obtains the temperature of each region, and in the case that the temperature exceeds a certain threshold, the programmable circuit module can regulate the corresponding fan controller according to the heat dissipation regulation strategy, the fourth second bus I2C4_2 and the fourth third bus I2C4_3 communicatively connected to increase the fan speed to enhance the scattering effect. When the temperature falls back to the normal range, the programmable circuit module can regulate the corresponding fan controller according to the heat dissipation regulation strategy, the fourth second bus I2C4_2 and the fourth third bus I2C4_3 communicatively connected to reduce the fan speed to reduce the wind noise and save power.

[0123] The present application also provides a control method of the above-mentioned server.

[0124] Figure 6 An operation flowchart of a control method of a server according to an embodiment of the present application is shown.

[0125] As shown in Figure 6 The control method includes operation S610 and operation S620.

[0126] In operation S610, the baseboard management controller controls the selection circuit to establish a first bus link between the baseboard management controller and the component through the first enable signal, and performs platform management on the component through the first bus link.

[0127] In operation S620, the backup management controller, in response to detecting that the baseboard management controller fails, controls the selection circuit to establish a second bus link between the backup management controller and the component through the second enable signal, and takes over the platform management of the component through the second bus link.

[0128] According to embodiments of the present application, the baseboard management controller controls the selection circuit to communicatively connect the baseboard management controller and the component, and performs platform management on the component to health manage the running state of the server. In the case of failure of the baseboard management controller, the backup management controller can control the selection circuit to communicatively connect the backup management controller and the component, take over the platform management of the component, and health manage the running state of the server by the backup management controller, so that the server can maintain a normal working state or can recover to a normal working state in the case of communication connection failure.

[0129] According to embodiments of the present application, the baseboard management controller provides a first enable signal to the selection circuit through a first enable signal line, and the selection circuit communicatively connects the first bus and the third bus in response to the first enable signal to establish a first bus link between the baseboard management controller and the component.

[0130] According to embodiments of the present application, the baseboard management controller controls the selection circuit to establish the first bus link by using the first enable signal line, so that the baseboard management controller can perform platform management on the component, and the baseboard management controller can perform master control in the case of normal operation of the baseboard management controller.

[0131] According to embodiments of the present application, the backup management controller receives a state signal from the baseboard management controller through a first detection signal line, judges whether the baseboard management controller fails according to the state signal, and provides a second enable signal to the selection circuit through a second enable signal line in response to detecting that the baseboard management controller fails. The selection circuit communicatively connects the second bus and the third bus in response to the second enable signal to establish a second bus link between the backup management controller and the component.

[0132] According to embodiments of the present application, the backup management controller controls the selection circuit to establish the second bus link by using the second enable signal line, so that the backup management controller can take over the control in the case of failure of the baseboard management controller, and the reliability of the dual-host management mode can be improved.

[0133] According to an embodiment of the present application, the control method further comprises: in response to detecting that the first bus link is abnormal, the baseboard management controller sends a first reset signal for resetting the components to the selection circuit through a first reset signal line, and the selection circuit, in response to the first enable signal, communicatively connects the first reset signal line with a third reset signal line; and in response to detecting that the second bus link is abnormal, the backup management controller sends a second reset signal for resetting the components to the selection circuit through a second reset signal line, and the selection circuit, in response to the second enable signal, communicatively connects the second reset signal line with the third reset signal line.

[0134] According to an embodiment of the present application, in the case of the first bus link being abnormal, the baseboard management controller can reestablish the first bus link through a separate reset link. In the case of the second bus link being abnormal, the backup management controller can reestablish the second bus link through a separate reset link, can be stripped from the internal logic of the first bus link or the second bus link, forcibly restore the safe state from the unknown state of the first bus link or the second bus link, and improve the reliability of the server.

[0135] According to an embodiment of the present application, the control method further comprises: in response to detecting that the baseboard management controller fails, the backup management controller sends a fourth reset signal to a fourth reset signal line; and the baseboard management controller performs a reset according to the fourth reset signal, and in the case of failure recovery after the reset, resumes the platform management of the components.

[0136] According to an embodiment of the present application, the backup management controller determines whether to take over the platform management of the components according to the running state of the baseboard management controller, takes over the platform management of the components in a timely manner when the baseboard management controller fails, can avoid server downtime, and can also timely exclude the baseboard management controller from the platform management of the components, preventing the baseboard management controller from outputting an error signal due to failure and affecting the normal operation of the server.

[0137] According to an embodiment of the present application, the control method further comprises: in the case that neither the backup management controller nor the baseboard management controller fails, the backup management controller and the baseboard management controller periodically and alternately take over the platform management of the components, and transmit detection information of the components generated in the platform management process to each other through a fourth bus.

[0138] According to an embodiment of the present application, by periodically and alternately taking over the platform management of the components in the case that neither the baseboard management controller nor the backup management controller fails, the baseboard management controller and the backup management controller can be timely investigated for their own abnormalities.

[0139] According to the embodiments of the present application, the platform management comprises obtaining target data from the component, comparing the target data with the reference information, sending control information to the component according to the comparison result, and detecting information is the comparison result of the target data and the reference information.

[0140] According to the embodiments of the present application, the substrate management controller and the standby management controller can execute corresponding management measures or send fault alarm information according to the detection information generated in the execution of the platform management process, so as to prevent the damage of the server.

[0141] According to the embodiments of the present application, the control method further comprises: the substrate management controller and the standby management controller judging whether the detection information generated in the execution of the platform management process is different from the detection information received from the other party, and sending alarm information in response to the difference for K times, wherein K is a preset integer.

[0142] According to the embodiments of the present application, by periodically alternating the takeover of the platform management of the component in the absence of the fault of both the substrate management controller and the standby management controller, and monitoring the detection information of both, the information abnormality of the first bus link and the component, and the information abnormality of the second bus link and the component can be timely investigated, so as to avoid causing more serious damage to the server.

[0143] According to the embodiments of the present application, the control method further comprises: the standby management controller sending the state of the standby management controller to the substrate management controller through the second detection signal line.

[0144] According to the embodiments of the present application, the state information of the standby management controller can be represented by the high level state or the low level state of the digital level pin of the second detection signal line. The state of the digital level pin of the second detection signal line can send the state information of the standby management controller to the substrate management controller through the second detection signal line with zero delay, fully utilize the hardware resources, timely feedback the state information of the standby management controller, and timely perform abnormal processing when the standby management controller is in an abnormal working state.

[0145] Those skilled in the art can understand that the features described in various embodiments of the present application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present application. In particular, the features described in various embodiments of the present application can be combined and / or combined in various ways without departing from the spirit and teachings of the present application. All these combinations and / or combinations fall within the scope of the present application.

[0146] The embodiments of the application have been described. However, these embodiments are merely for illustration and are not intended to limit the scope of the application. Although each embodiment is described above separately, this does not mean that the measures in each embodiment cannot be used advantageously in combination. Various alternatives and modifications to the embodiments described herein will be apparent to those skilled in the art in view of the foregoing description. Such alternatives and modifications are intended to fall within the scope of the application.

Claims

1. A server, characterized in that, The server includes: a baseboard management controller, a backup management controller, components, and a selection circuit. The backup management controller is implemented by a programmable circuit module integrated into the server. The backup management controller is connected to the baseboard management controller. The baseboard management controller is connected to the selection circuit via a first bus, a first enable signal line, and a first reset signal line. The baseboard management controller is configured to provide a first enable signal to the first enable signal line and transmit a first bus signal for platform management via the first bus. In response to detecting a first bus link abnormality, the controller sends a first reset signal to the first reset signal line to reset the component. The backup management controller is connected to the selection circuit via a second bus, a second enable signal line, and a second reset signal line. The backup management controller is configured to provide a second enable signal to the second enable signal line and transmit a second bus signal for platform management via the second bus in response to detecting a fault in the baseboard management controller; and to send a second reset signal to the second reset signal line for resetting the component in response to detecting an abnormality in the second bus link. The selection circuit is connected to the component via a third bus and a third reset signal line. The selection circuit is configured to, in response to receiving a first enable signal from a first enable signal line, communicate between the first bus and the third bus to establish a first bus link between the board management controller and the component, such that the board management controller performs platform management on the component via the first bus link, and communicate between the first reset signal line and the third reset signal line; and to, in response to receiving a second enable signal from a second enable signal line, communicate between the second reset signal line and the third reset signal line, and communicate between the second bus and the third reset signal line to establish a second bus link between the standby management controller and the component, such that the standby management controller takes over platform management of the component via the second bus link, and communicate between the second reset signal line and the third reset signal line. The first bus, second bus, and third bus each include clock lines and data lines. The first bus link anomaly includes at least one of the following: The data lines of the first bus and the third bus are continuously clamped at a low level; The first bus signal on one or more of the clock line, data line, clock line, and data line of the first bus is pulled to an undesirable high or low level; and The data lines of the first bus and the third bus are continuously occupied; The second bus link failure includes at least one of the following: The data lines of the second bus and the third bus are continuously clamped at a low level; The second bus signals on one or more of the clock line, data line, clock line, and data line of the second bus are pulled to an undesirable high or low level; and The data lines of the second and third buses are continuously occupied.

2. The server according to claim 1, characterized in that, The first bus, the second bus, and the third bus are integrated circuit buses.

3. The server according to claim 1, characterized in that, The backup management controller is connected to the baseboard management controller via the first detection signal line. The backup management controller is configured to receive a status signal from the baseboard management controller via the first detection signal line, determine whether the baseboard management controller has failed based on the status signal, and in response to detecting a failure in the baseboard management controller, the control selection circuit connects the backup management controller to the component for communication and takes over the platform management of the component.

4. The server according to claim 3, characterized in that, The status signal is a heartbeat signal periodically sent by the baseboard management controller.

5. The server according to claim 3, characterized in that, The backup management controller is also connected to the baseboard management controller via a fourth reset signal line, and the backup management controller is also configured to send a fourth reset signal to the fourth reset signal line in response to detecting a fault in the baseboard management controller; The baseboard management controller is also configured to perform a reset based on a fourth reset signal and, if the fault is recovered after the reset, to take over platform management of the components again.

6. The server according to claim 3, characterized in that, The backup management controller is also connected to the baseboard management controller via a fourth bus. The backup management controller and the baseboard management controller are configured to periodically take over the platform management of the components when neither the baseboard management controller nor the backup management controller is fault-free, and to exchange the detection information of the components generated during the platform management process via the fourth bus.

7. The server according to claim 6, characterized in that, The platform management includes obtaining target data from components, comparing the target data with reference information, and sending control information to components based on the comparison result. The detection information is the comparison result between the target data and the reference information.

8. The server according to claim 6, characterized in that, The baseboard management controller and the backup management controller are also configured to: determine whether there is a difference between the detection information generated by themselves during the execution of platform management and the detection information received from the other party, and issue an alarm message in response to K consecutive differences, where K is a preset integer.

9. The server according to claim 6, characterized in that, The fourth bus is an integrated circuit bus.

10. The server according to claim 3, characterized in that, The backup management controller is also connected to the substrate management controller via a second detection signal line, and the backup management controller is also configured to send the status information of the backup management controller to the substrate management controller via the second detection signal line.

11. The server according to claim 1, characterized in that, The server also includes a central processing unit, which is connected to the baseboard management controller via a fifth bus and to the output of the selection circuit via a sixth bus. The fifth bus is a high-speed bus for interconnecting peripheral components, and the sixth bus is an integrated circuit bus.

12. The server according to claim 11, characterized in that, The central processing unit is also connected to the peripheral component interconnect high-speed bus switching circuit via the seventh bus, which is the peripheral component interconnect high-speed bus.

13. The server according to claim 11, characterized in that, The baseboard management controller, backup management controller, selection circuit, and central processing unit are all located on the server's motherboard, and the components include at least one of the components on the motherboard and the components outside the motherboard.

14. The server according to claim 13, characterized in that, The server also includes a rack and a backplane installed in the rack. The backplane divides the rack into a first space and a second space. The motherboard is located in the first space, and the components outside the motherboard are located in the second space. The selection circuit is connected to the components outside the motherboard through the backplane.

15. The server according to any one of claims 1 to 14, characterized in that, The component includes a multiplexing circuit and multiple hardware units. The input of the multiplexing circuit is connected to a selection circuit, and the output of the multiplexing circuit is connected to multiple hardware units. The hardware units include heat sinks, power supplies, or management hardware.

16. The server according to any one of claims 1 to 14, characterized in that, The components include heat sinks, power supplies, or management hardware.

17. The server according to any one of claims 1 to 14, characterized in that, The server includes N selection circuits and N components, wherein the first input terminal of the nth selection circuit is connected to the baseboard management controller, the second input terminal of the nth selection circuit is connected to the backup management controller, and the output terminal of the nth selection circuit is connected to the nth component. N and n are both positive integers, and 1≤n≤N.

18. A method for controlling a server as described in any one of claims 1 to 14, characterized in that, The control method includes: The substrate management controller establishes a first bus link between the substrate management controller and the component through a first enable signal control selection circuit, and performs platform management on the component through the first bus link. In response to detecting a fault in the baseboard management controller, the backup management controller establishes a second bus link between the backup management controller and the component through a second enable signal control selection circuit, and takes over the platform management of the component through the second bus link.

19. The control method according to claim 18, characterized in that, The substrate management controller establishes a first bus link between the substrate management controller and the component by controlling the selection circuit through a first enable signal, including: The substrate management controller provides a first enable signal to the selection circuit via a first enable signal line. In response to the first enable signal, the selection circuit establishes a first bus link between the first bus and the third bus to establish a first bus link between the substrate management controller and the component.

20. The control method according to claim 18, characterized in that, In response to detecting a fault in the baseboard management controller, the backup management controller controls the selection circuit to establish a second bus link between the backup management controller and the component via a second enable signal, including: The backup management controller receives a status signal from the baseboard management controller via the first detection signal line, determines whether the baseboard management controller has failed based on the status signal, and provides a second enable signal to the selection circuit via the second enable signal line in response to detecting a failure in the baseboard management controller. The selection circuit responds to the second enable signal to communicate between the second bus and the third bus to establish a second bus link between the standby management controller and the components.

21. The control method according to claim 18, characterized in that, The control method further includes: In response to detecting a first bus link abnormality, the baseboard management controller sends a first reset signal to the selection circuit via a first reset signal line for resetting the component. In response to the first enable signal, the selection circuit connects the first reset signal line to the third reset signal line. In response to detecting a second bus link anomaly, the backup management controller sends a second reset signal to the selection circuit via the second reset signal line to reset the component. In response to the second enable signal, the selection circuit connects the second reset signal line to the third reset signal line.

22. The control method according to claim 18, characterized in that, The control method further includes: In response to detecting a fault in the baseboard management controller, the backup management controller sends a fourth reset signal to the fourth reset signal line; The baseboard management controller performs a reset based on the fourth reset signal, and resumes platform management of the components if the fault is recovered after the reset.

23. The control method according to claim 18, characterized in that, The control method further includes: when both the backup management controller and the baseboard management controller are fault-free, the backup management controller and the baseboard management controller periodically take over the platform management of the components, and exchange the detection information of the components generated during the platform management process through the fourth bus.

24. The control method according to claim 23, characterized in that, The platform management includes obtaining target data from components, comparing the target data with reference information, and sending control information to components based on the comparison result. The detection information is the comparison result between the target data and the reference information.

25. The control method according to claim 23, characterized in that, The control method further includes: the baseboard management controller and the backup management controller determining whether there is a difference between the detection information generated by themselves during the execution of platform management and the detection information received from the other party, and issuing an alarm message in response to K consecutive differences, where K is a preset integer.

26. The control method according to claim 18, characterized in that, The control method further includes: the backup management controller sending the status of the backup management controller to the baseboard management controller via a second detection signal line.

Images