CN103995575B - A kind of startup of server method and server - Google Patents

Abstract

The invention discloses a method for starting a server. The server includes: a power supply module PSU, a fan module, a system management control module SMC and a plurality of node modules. Each other module in the server; after the SMC startup is completed, judge whether the power supply module is working normally, if normal, then detect whether the fan module and each node module are in place, if the fan module is in place, then instruct the fan module to start the fan, Monitor the running status of the power supply module and fan module; the SMC calculates the power supply and cooling capacity of the server based on the status of the node module and the running status of the power module and fan module. For example, the power supply and cooling capacity of the server meets the operating requirements of the node module in place , it instructs the node module to start up. The invention can improve the startup reliability of the server. The invention also discloses a server.

Description

A method for starting a server and the server

technical field

The present invention relates to the technical field of computer servers, in particular to a method for starting a server and the server.

Background technique

High-end servers, that is, key application hosts, have dozens of times the processing power of ordinary servers, and are the core systems of lifeline industries such as finance, telecommunications, energy, and transportation. For example, a high-end server can be expanded to 32 processors with hundreds of boards.

The power-on operation of ordinary servers is only through simple buttons or remote control. Before power-on, the system status is not detected and judged. If a module or node is abnormal, the system may be damaged after power-on.

Compared with ordinary servers, high-end servers have higher requirements on reliability. Therefore, how to ensure the power-on reliability of high-end servers is a problem that needs to be solved.

Contents of the invention

The technical problem to be solved by the present invention is to provide a server starting method and a server, which can improve the starting reliability of the server.

In order to solve the above-mentioned technical problems, the present invention provides a method for starting a server, the server comprising: a power supply module PSU, a fan module, a system management control module SMC and a plurality of node modules, characterized in that the method includes:

The power module is powered on and outputs the power supply voltage to other modules in the server;

After the SMC startup is complete, judge whether the power module is working normally. If it is normal, check whether the fan module and each node module are in place. If the fan module is in place, instruct the fan module to start the fan and monitor the power module and fan module. operating status;

The SMC calculates the power supply and heat dissipation capabilities of the server based on the presence of the node modules and the running status of the power modules and fan modules. If the power supply and heat dissipation capabilities of the server meet the operating requirements of the node modules in place, the SMC instructs the node modules to start up.

Further, the method also includes the following features:

The node module includes baseboard management controller BMC and complex programmable logic device CPLD;

The SMC instructs the node modules to start up, including:

The SMC sends a power-on command to the BMC of each node module;

After receiving the start-up instruction, the BMC instructs the CPLD of the node module to perform a start-up operation.

Further, the method also includes the following features:

The node module includes any one or any several of the following modules: computing node module, IO node module, switching node module;

Among them, the computing node module is responsible for the calculation and processing of the server, the IO node module is responsible for the external connection of the server, and the switching node module is responsible for interconnecting the computing nodes in the server to form resource sharing partitions.

Further, the method also includes the following features:

When the node module is a computing node module, before the SMC instructs the node module to start up, it also includes:

The SMC receives the partition information set by the user;

The SMC sends master/slave node routing configuration information to the BMC of the corresponding computing node module according to the partition information;

After receiving the master/slave node routing configuration information, the BMC configures the route from the node module to the switching node module, and notifies the CPLD of the node module to perform the power-on operation of the master/slave node;

After receiving the notification from the BMC, the CPLD starts up according to the power-on sequence of the master/slave node.

Further, the method also includes the following features:

The BMC collects the voltage, temperature and chip register information of the node module, and reports it to the SMC.

In order to solve the above technical problems, the present invention also provides a server, including a power supply module PSU, a fan module, a system management control module SMC and a plurality of node modules, wherein:

The SMC connected to the PSU, fan module and multiple node modules is used to judge whether the power module is working normally after the startup is completed. If it is normal, it will detect whether the fan module and each node module are in place. If the fan module is in place, it will indicate The fan module starts the fan, monitors the running status of the power supply module and the fan module; calculates the power supply and cooling capacity of the server according to the presence of the node module and the running status of the power supply module and the fan module, such as the power supply and cooling capacity of the server. If the operation requirement of the bit node module is used, it instructs the node module to start up.

Further, the server also includes the following features:

The node module includes baseboard management controller BMC and complex programmable logic device CPLD;

Described BMC is used for receiving the start-up command that SMC sends, and instructs the CPLD of this node module to carry out start-up operation;

The CPLD is configured to perform power-on operation after receiving an instruction from the BMC.

Further, the server also includes the following features:

The node module includes any one or any several of the following modules: computing node module, IO node module, switching node module;

Among them, the computing node module is responsible for the calculation and processing of the server, the IO node module is responsible for the external connection of the server, and the switching node module is responsible for interconnecting the computing nodes in the server to form resource sharing partitions.

Further, the server also includes the following features:

The SMC is also used to receive the partition information set by the user before sending the boot command to the BMC of the computing node module, and send master/slave node routing configuration information to the BMC of the corresponding computing node module according to the partition information:

BMC is also used to configure the route from the node module to the switching node module after receiving the master/slave node routing configuration information, and notify the CPLD of the node to perform the power-on operation of the master/slave node;

The CPLD is also used to start up according to the power-on sequence of the master/slave node after receiving the notification from the BMC.

Further, the server also includes the following features:

It also includes a first LED indicator light, a second LED indicator light and a third LED indicator light, wherein:

The first LED indicator connected to the power module is used to indicate the working state of the power module;

The second LED indicator connected to the SMC is used to indicate the initialization process of the SMC;

The CPLD of each computing module is connected to a third LED indicator light, and the third LED indicator light is used to indicate that the node module is the master node or the slave node of the logical partition to which it belongs.

Compared with the prior art, the present invention provides a server startup method and server. The entire server adopts a unified power supply and unified heat dissipation architecture, which is centrally managed by the SMC. The startup process adopts hierarchical startup, which can prevent a certain module or node from being abnormal. Causing damage to the system can improve server startup reliability.

Description of drawings

FIG. 1 is a flowchart of a method for starting a server according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of a server according to an embodiment of the present invention.

FIG. 3 is a schematic diagram of a server structure of an application example of the present invention.

FIG. 4 is a schematic diagram of the SMC-BMC-CPLD three-level start-up management topology of the server of the application example of the present invention.

Detailed ways

In order to make the purpose, technical solution and advantages of the present invention more clear, the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined arbitrarily with each other.

The present invention proposes a highly reliable high-end server start-up method and server, whose features mainly include: the entire server adopts a unified power supply and unified heat dissipation architecture, a centralized management strategy by the SMC, and the start-up process adopts hierarchical start-up, and the start-up process There are strict detection and output instructions, which can prevent a certain module or node abnormality from causing damage to the system.

As shown in FIG. 1 , an embodiment of the present invention provides a method for starting a server. The server includes: a power supply module PSU, a fan module, a system management control module SMC, and a plurality of node modules. The method includes:

S10, the power module is powered on and outputs a power supply voltage to other modules in the server;

S20. After the SMC is started, it is judged whether the power supply module is working normally. If it is normal, then check whether the fan module and each node module are in place. If the fan module is in place, then instruct the fan module to start the fan, monitor the power supply module and The running status of the fan module;

S30. The SMC calculates the power supply and cooling capacity of the server according to the presence of the node module and the operating status of the power module and the fan module. If the power supply and cooling capacity of the server meet the operating requirements of the node module in place, the SMC instructs the node module to start.

The method may also include the following features:

Preferably, the node module includes a baseboard management controller (Baseboard Management Controller, BMC) and a complex programmable logic device (Complex Programmable Logic Device, CPLD).

Preferably, the SMC instructs the node module to start up, including:

The SMC sends a power-on command to the BMC of each node module;

After receiving the start-up instruction, the BMC instructs the CPLD of the node module to perform a start-up operation.

Preferably, the node module includes any one or more of the following modules: a computing node module, an IO node module, and a switching node module;

Among them, the computing node module is responsible for the calculation and processing of the server, the IO node module is responsible for the external connection of the server, and the switching node module is responsible for interconnecting the computing nodes in the server to form resource sharing partitions.

Preferably, when the node module is a computing node module, before the SMC instructs the node module to start up, it also includes:

The SMC receives the partition information set by the user;

The SMC sends master/slave node routing configuration information to the BMC of the corresponding computing node module according to the partition information;

After receiving the master/slave node routing configuration information, the BMC configures the route from the node module to the switching node module, and notifies the CPLD of the node module to perform the power-on operation of the master/slave node;

After receiving the notification from the BMC, the CPLD starts up according to the power-on sequence of the master/slave node

Wherein, the partition information refers to: master/slave computing node information constituting a logical partition, that is, which computing node in a logical partition is a master node and which computing nodes are slave nodes.

Among them, the CPLD of the computing node module controls the LED indicator to turn on and off according to the instructions of the BMC to indicate that the computing node module is a master node or a slave node; for example, if it is a master node, the LED is on, otherwise it is off.

Preferably, the BMC also collects the voltage, temperature and chip register information of the node module, and reports them to the SMC.

Preferably, the power supply module is powered on and outputs a supply voltage to each module, including:

The air switch connecting the three-phase mains and the power distribution unit (Power Distribution Unit, PDU) is closed, and the PDU is connected to the three-phase mains and converts the three-phase mains into single-phase AC output to the power module (Power SupplyUnit, PSU);

The PSU converts single-phase AC power into DC power and outputs it to the backplane;

Wherein, after the PSU works, the LED indicator light can be used to indicate whether the power supply module is working normally.

Wherein, the direct current output by the PSU includes: 5V and 12V direct current. Among them, 5V DC is provided to SMC and switches (for example, Ethernet switches), and 12V DC is provided to fan modules and node modules;

Among them, the 5V DC output from the PSU can control the SMC to be powered on through a switch;

Wherein, the SMC is connected to each node module through a switch (for example, an Ethernet switch);

Wherein, the server further includes a cabinet liquid crystal controller, and the cabinet liquid crystal controller is powered by the single-phase alternating current output by the PDU, and is connected to the SMC through an Ethernet switch. Users can log in to the SMC management interface through the cabinet LCD controller, and the cabinet LCD controller can display the information monitored by the SMC.

Wherein, after the SMC is powered on, it is also initialized, and the process of the SMC initialization can be indicated by the LED indicator, for example: the initialization is in progress and the initialization is completed.

As shown in Figure 2, an embodiment of the present invention provides a server, including a power supply module PSU, a fan module, a system management control module SMC and multiple node modules, wherein:

The SMC connected to the PSU, fan module and multiple node modules is used to judge whether the power module is working normally after the startup is completed. If it is normal, it will detect whether the fan module and each node module are in place. If the fan module is in place, it will indicate The fan module starts the fan, monitors the running status of the power supply module and the fan module; calculates the power supply and cooling capacity of the server according to the presence of the node module and the running status of the power supply module and the fan module, such as the power supply and cooling capacity of the server. If the operation requirement of the bit node module is used, it instructs the node module to start up.

The server may also include the following features:

Wherein, the power supply module is used to supply power to other modules in the server;

The fan module is used to dissipate heat for the server;

Preferably, the node module includes a baseboard management controller BMC and a complex programmable logic device CPLD;

Described BMC is used for receiving the start-up command that SMC sends, and instructs the CPLD of this node module to carry out start-up operation;

The CPLD is configured to perform power-on operation after receiving an instruction from the BMC.

Preferably, the node module includes any one or more of the following modules: a computing node module, an IO node module, and a switching node module;

Among them, the computing node module is responsible for the calculation and processing of the server, the IO node module is responsible for the external connection of the server, and the switching node module is responsible for interconnecting the computing nodes in the server to form resource sharing partitions.

Preferably, the SMC is also used to receive partition information set by the user before sending the boot command to the BMC of the computing node module, and send master/slave node routing configuration information to the BMC of the corresponding computing node module according to the partition information:

The BMC of the computing node module is also used to configure the route from the node module to the switching node after receiving the master/slave node routing configuration information, and notify the CPLD of the node module to perform the power-on operation of the master/slave node;

The CPLD of the computing node module is also used to start up according to the power-on sequence of the master/slave node after receiving the notification from the BMC.

Preferably, the BMC is also used to collect the voltage, temperature and chip register information of the node module, and report to the SMC.

Preferably, the server also includes a first LED indicator, a second LED indicator and a third LED indicator, wherein:

The first LED indicator connected to the power module is used to indicate the working state of the power module;

The second LED indicator connected to the SMC is used to indicate the initialization process of the SMC;

The CPLD of each computing module is connected to a third LED indicator light, and the third LED indicator light is used to indicate that the node module is the master node or the slave node of the logical partition to which it belongs.

Preferably, the server also includes:

The cabinet liquid crystal controller connected with the SMC is used to provide the user with an interface for managing the SMC and display the information monitored by the SMC.

Preferably, the server also includes:

The switch is used to connect the SMC, the cabinet liquid crystal controller, and the BMC of the node module into a startup management network through Ethernet.

Application example

As shown in Fig. 3, the present invention proposes a hierarchical start-up management mode of SMC-BMC-CPLD, that is, all operation instructions are issued by the topmost SMC, and the entire start-up process is strictly supervised by the SMC. The SMC transmits instructions to the BMC in each node through the Ethernet switch, and the BMC communicates with the CPLD through the control bus. Among them, in addition to issuing all instructions, SMC also directly manages the power module PSU and fan module in a centralized and unified manner; BMC is mainly responsible for the temperature, voltage, and register information collection of main chips in each node; CPLD controls the power-on sequence, shutdown, and reset of each node operation etc.

A highly reliable high-end server, its composition is shown in Figure 3, including: (1) air switch, (2) PDU (PowerDistribution Unit, power distribution unit), (3) PSU (Power Supply Unit, power supply module) , (4) fan module, (5) SMC (System Management Controller), (6) switch, (7) computing node, (8) IO node, (9) switching node, (10) cabinet LCD control device.

(1) Air switch: connect the three-phase mains and the PDU. When the air switch is closed, the PDU is connected to the three-phase mains. The reason for using three-phase mains power is that high-end servers have many modules and nodes and high power consumption, so high-power three-phase power supply is required.

(2) PDU: Connect the air switch and the PSU. The PDU is responsible for converting the three-phase mains power into a single-phase AC power suitable for input to the PSU.

(3) PSU: Connect the PDU and the backplane. The input and output ends of all the PSUs are connected in parallel to provide centralized power supply, and are uniformly monitored and managed by the SMC. The PSUs are connected to the SMC through the control bus. After the PSU is powered on, it can output 5V DC (P5V_STBY) and 12V DC (P12V_STBY). P5V_STBY is provided to the SMC and switches through the backplane, and P12V_STBY is provided to the fan modules, computing nodes, IO nodes, and switching nodes through the backplane.

(4) Fan module: The entire server system adopts a centralized heat dissipation design, which is uniformly monitored and managed by the SMC. The fan module is responsible for the heat dissipation of the entire server. The fan module is connected to the SMC through the control bus.

(5) SMC: The topmost management system of the entire server is connected to the client host, and all commands during the startup process are issued by the SMC. In addition to issuing all commands, the SMC also directly monitors and manages the PSU and fan modules in a unified manner. In addition, the SMC is connected to the cabinet liquid crystal controller and the BMC of the node (computing node, IO node, switching node) through the switch. As a top-level management system, SMC builds a hierarchical start-up management network of SMC, BMC, and CPLD.

(6) Switch: connect the SMC, the cabinet controller, and the BMC in the nodes (computing nodes, IO nodes, switching nodes) through Ethernet to form a startup management network.

(7) Computing node: The node responsible for computing in the server, including BMC and CPLD.

(8) IO node: The node responsible for external IO interconnection in the server, including BMC and CPLD.

(9) Switching node: The server is responsible for interconnecting the CPUs in the computing nodes to form a resource sharing partition, including BMC and CPLD.

Among them, BMC: computing nodes, IO nodes, and switching nodes have their own BMCs, and their main function is to collect information such as voltage, temperature, and main chip registers in each node and report them to the SMC. The BMC in the computing node also has the function of setting whether the node is a master node or a slave node, and configuring a route between the computing node and the switching node.

Among them, CPLD: CPLD is connected to BMC through control bus (control bus can be SMBus bus (System Management Bus, system management bus)), and cooperates with BMC to complete the management of main board, such as the sequence control of starting, shutting down and resetting of nodes. The CPLD in the computing node can indicate whether the node is a master node or a slave node through LEDs, and decide whether to execute the power-on sequence of the master node or the power-on sequence of the slave node. (The power-on sequence of the two is different)

(10) Cabinet LCD controller: connected to the SMC through an Ethernet switch, the cabinet LCD controller can log in to the management interface of the SMC and display the information monitored by the SMC. The cabinet LCD controller is powered by the single-phase AC output from the PDU.

As shown in Figure 4, the startup process of this highly reliable high-end server is described in detail below, including the following steps:

(1) The "air switch" is connected to the three-phase mains and the PDU module. When the air switch is closed, the PDU is connected to the three-phase mains.

(2) The three-phase mains power is converted into single-phase AC power through the PDU.

(3) After the conversion of single-phase AC power is completed, first light "LED1", indicating that the conversion is completed.

(4) The single-phase alternating current also supplies power to the LCD control panel of the cabinet, and the LCD panel starts to power on.

(5) Each power supply unit PSU0, PSU1...PSUn in the power supply module PSU is connected to the PDU and the backplane, and the PSU outputs 5V DC (P5V_STBY) and 12V DC (P12V_STBY) after the single-phase AC is connected. Among them, P5V_STBY is provided to SMC and switches through the backplane, and P12V_STBY is provided to fan modules, computing nodes, IO nodes, and switching nodes through the backplane.

(6) Close "switch 1" and power on the SMC. After P5V_STBY is provided to the SMC and the switch, the two start initialization respectively. During the initialization process of the SMC, the indicator light "LED2" keeps flashing, indicating that the SMC is being initialized, and no operations can be performed at this time. After the SMC initialization is completed, "LED2" is always on, allowing the user to operate.

(7) After the SMC is started, read information such as voltage and temperature of all PSUs through "control bus 1" (control bus 1 can be SMBus (System Management Bus, system management bus)), and judge whether the power module is working normally. All PSUs are centrally and uniformly monitored and managed by the SMC.

(8) After P12V_STBY is output to the fan module, computing node, IO node, and switching node, the BMC and CPLD in each node start to power on and initialize until they are completed.

(9) The SMC detects the presence of fan modules, computing nodes, IO nodes, and switching nodes. The detection method is to judge the presence signals of the fan modules and nodes.

(10) The SMC turns on the system fan, and detects information such as the number and speed of the fan modules through the "control bus 2" (the control bus 2 can be the SMBus bus (System Management Bus, system management bus)) to determine whether the fan is abnormal, Whether the cooling capacity is sufficient.

Specifically, the SMC detects the status of the PSU in step (7), detects the presence of fan modules and nodes in step (9), and detects the operating status of the fans in step (10). Check whether the power supply of the PSU is sufficient and whether the cooling capacity of the fan module is sufficient.

If there is no problem after the SMC calculation, you can proceed to the next steps; if there is a problem or hidden danger, alarm through the SMC management page and the cabinet LCD controller, and prohibit the next operation.

(11) In addition, the cabinet LCD controller is connected to the SMC through an Ethernet switch, and the cabinet LCD controller can log in to the management interface of the SMC, and can obtain and display the above-mentioned PSU status information, each module or node presence information, fan module status information.

(12) The client host is connected to the SMC through Ethernet, and the user can log in to the management interface of the SMC to set the partition information: that is, set which computing nodes form a logical partition, which computing node is the master node, and which computing node or nodes are the slave node.

(13) After the partition information is set, the SMC transmits the partition information to the BMC in the computing node through the Ethernet.

(14) After receiving the partition information, the BMC starts to configure the routing function between the computing node and the switching node. Another function of the BMC is to collect information such as voltage, temperature, and main chip registers in the node before power-on and report it to the SMC.

(15) CPLD is connected to BMC through "control bus 3" (control bus 3 can be SMBus bus (System Management Bus, system management bus)), and after receiving the partition information passed by BMC, it will be indicated as the master node through "LED3" It is still a slave node. If it is a master node, LED3 will be on, otherwise it will not be on. In addition, the CPLD needs to decide whether to execute the power-on sequence of the master node or the power-on sequence of the slave node. (The master node is a Legacy device, and the slave node is a Non-Legacy device. Their power-on sequences are different)

(16) After the above steps are completed, log in to the SMC management interface through the client host or the cabinet LCD controller to start the partition just set.

(17) Executing the boot operation: the SMC transmits the boot command to the BMCs in all nodes via the Ethernet. The computing node is powered on by the CPLD, the IO node is powered on by the CPLD, and the switching node is powered on by the CPLD.

(18) After the start-up is completed, each BMC will continue to report the voltage, temperature and register information of the node after the start-up to the SMC.

In the server startup method and server provided by the above embodiments, the entire server adopts a unified power supply and unified heat dissipation architecture, and is managed centrally by the SMC. The startup process adopts hierarchical startup, and there are strict detection and output instructions during the startup process, which can Preventing a certain module or node abnormality from causing damage to the system can improve the startup reliability of the server.

Those of ordinary skill in the art can understand that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a magnetic disk or an optical disk, and the like. Optionally, all or part of the steps in the above embodiments can also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the above embodiments can be implemented in the form of hardware, or can be implemented in the form of software function modules. The form is realized. The present invention is not limited to any specific combination of hardware and software.

It should be noted that the present invention can also have other various embodiments, without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and deformations according to the present invention, but these Corresponding changes and deformations should belong to the scope of protection of the appended claims of the present invention.

Claims (4)

1. a kind of startup of server method, the server include：Power module PSU, blower module, system administration control module SMC and multiple node modules, it is characterised in that this method includes：

Electric and output supplies voltage to each other modules in server on power module；

After SMC start completions, judge whether the power module work is normal, it is such as normal, then detect blower module and each node Whether module is in place, as blower module is in place, it indicates that the blower module starts fan, monitors power module and blower module Running status；

SMC is according to the power supply of the situation in place and power module of node module and the running status calculation server of blower module And heat-sinking capability, power supply and heat-sinking capability such as server meet the service requirement of node module in place, it indicates that node module Boot up；

Node module includes baseboard management controller BMC and complex programmable logic device (CPLD)；

SMC instruction node modules boot up, including：

SMC sends start-up command to the BMC of each node module；

After BMC receives the start-up command, indicate that the CPLD of this node module boots up operation；

BMC gathers the voltage, temperature and chip register information of this node module, and reports SMC；

The node module includes with any of lower module or appointed several：Calculate node module, I/O node module, exchange section Point module；

Wherein, the calculation process of calculate node module charge server, the external connection of I/O node module charge server, exchange Node module is responsible for that the calculate node in server interconnects to form the subregion of resource-sharing.

2. the method as described in claim 1, it is characterised in that：

When the node module is calculate node module, before SMC instruction node modules boot up, in addition to：

SMC receives the partition information that user is set；

SMC sends master/slave node routing configuration information according to the partition information to the BMC of corresponding calculate node module；

After BMC receives the master/slave node routing configuration information, the route that this node module arrives switching node module is configured, The CPLD of this node module is notified to perform the power on operation of master/slave node；

After CPLD receives BMC notice, booted up according to the electrifying timing sequence of master/slave node.

3. a kind of server, including power module PSU, blower module, system administration control module SMC and multiple node modules, Wherein：

The SMC being connected with PSU, blower module and multiple node modules, after start completion, whether judge power module work Normally, as normal, then detect blower module and whether each node module is in place, as blower module is in place, it indicates that the fan Module starts fan, monitors the running status of power module and blower module；According to the situation in place and power supply of node module The power supply of the running status calculation server of module and blower module and heat-sinking capability, power supply and heat-sinking capability such as server are expired The service requirement of foot node module in place, it indicates that node module boots up；

Node module includes baseboard management controller BMC and complex programmable logic device (CPLD)；

The BMC, for receiving the start-up command of SMC transmissions, indicate that the CPLD of this node module boots up operation；

The CPLD, power-on operation is carried out after the instruction for receiving BMC；

The node module includes with any of lower module or appointed several：Calculate node module, I/O node module, exchange section Point module；

Wherein, the calculation process of calculate node module charge server, the external connection of I/O node module charge server, exchange Node module is responsible for that the calculate node in server interconnects to form the subregion of resource-sharing；

Also include the first LED light, the second LED light and the 3rd LED light, wherein：Be connected with power module One LED light, for indicating the working condition of power module；

The second LED light being connected with SMC, for indicating SMC initialization process；

One the 3rd LED light of CPLD connections of each computing module, the 3rd LED light, for indicating this section Point module is the host node of affiliated logical partition or from node.

4. server as claimed in claim 3, it is characterised in that：

SMC, it is additionally operable to before sending start-up command to the BMC of calculate node module, the partition information that user is set is received, according to institute State partition information and send master/slave node routing configuration information to the BMC of corresponding calculate node module：

BMC, it is additionally operable to after receiving the master/slave node routing configuration information, configures this node module and arrive switching node module Route, notify this node CPLD perform master/slave node power on operation；

CPLD, booted up according to the electrifying timing sequence of master/slave node after the notice for being additionally operable to receive BMC.